Yet Another View - by wwenze

Fiio E5 and E11 - PSP Edition (Work in progress)

2012-01-26T11:12:00.185+08:00

(Pardon the pictures taken under insufficient lighting. The figurines cannot take a head-on light from the onboard flash, and my slave flash is missing somewhere in Portland, Oregon.)

So, the X10 is a pain to use without an amp with low output impedance. Actually iPhone 4/4S has a decently low output impedance and its performance can put some of those external amps and DACs (Fiios included) to shame. But it costs SGD$500 (16GB) with a plan (the one I'm planning to get), does not have a 16:9 screen, and cannot play Pokemon Online. So screw it, Android FTW.

With my recontract still a while away, I am relying on my PSP for video duty on my daily commute. The amp's performance is quite acceptable, but with a high output impedance it messes with the X10's sound. I need an amp, small enough and just decent enough. Fiio E5 comes to mind.

Why E5 and not E6 or E11? Well, it's slightly cheaper than the E6 (USD$20 before shipping now, actually it once hit $15 before shipping but I didn't buy it then although I was tempted to, ため息。), but more importantly, according to NwAvGuy, E6 and E11 both have a nasty power supply noise (does it count as power supply if the switching power supply is inside the amp chip itself?) at high magnitude and frequency. From personal experience, ultrasonic noise can show up as whine or hiss in the audible frequencies depending on your setup, which probably explains why some people complained of hiss with certain earphones. And with X10 having a sensitivity of 110dB/mW, this is not a chance I'm going to take.

But then, in a sudden twist of events, I became the owner of an E11, hours after I received the E5 in the mail.

The measuring

There are already many reviews on both amplifiers, so I won't be talking about them in general. What I will be talking about, is how it boosts the performance of my PSP. Or reduce it.

Ok, because this is very situational, I'll have to control more test conditions.

PSP volume: 26/30, resulting in -25.3dB at Xonar DX's input when unloaded
Amp overall gain: 1 (whatever is the op-amp's gain times the attenuation by the volume control)
For Fiio E11: Gain switch is set to low

The results

Important: Measurement results are applicable to this specific scenario only!

Note: 20Hz measurement results removed, discussion on results removed, remaining results are for preview only and will be replaced with new ones, eventually.

Yea, because I'm going to do some changes to the measuring setup, for better accuracy. And it takes time for the parts to arrive and me to fix it up. So here are the remaining results for a glimpse of their relative performance:

Unloaded	PSP	PSP+E5	PSP+E11
THD@1kHz	0.0048%	0.0043%	0.0049%
THD+N@1kHz	0.076%(A)	0.070%(A)	0.077%(A)
THD@10kHz	0.0053%	0.0043%	0.0051%
THD+N@10kHz	0.088%(A)	0.083%(A)	0.088%(A)

Loaded w/33Ω	PSP	PSP+E5	PSP+E11
THD@1kHz	0.016%	0.0050%	0.0051%
THD+N@1kHz	0.079%(A)	0.070%(A)	0.077%(A)
THD@10kHz	0.10%	0.015%	0.0070%
THD+N@10kHz	0.13%(A)	0.085%(A)	0.088%(A)

I can still do things that do not require the parts flying over from China right now, although they will not be systematic and more of a "as I find it".

E5 input clipping

If you have read NwAvGuy's review of E5, the clipping threshold was measured to be 760mVrms. Well, clipping is one thing, but distortion will also be rising as the limit is approached.

0.0070% of E5 max, Xonar @ 70% vs 0.0075% of E5 gain 1, Xonar @ 50%.

While distortion is indeed higher with higher input levels, the difference is minimal. So it'll be safe as long as input level is within or below this range. Which is the case for majority of portable devices @ max volume. So you'd probably want to operate your player @ near max volume, so you will get less quantization error and noise. But not high enough to make the player clip though.

To be continued...

It's funny because it's true

2012-01-25T15:40:00.000+08:00

Random thought: Rephrasing

2012-01-23T08:07:00.000+08:00

You have a device running on 115V.

When you use it with a step-down transformer, you worry about the negative impact of the step-down transformer.

But when you use it with an isolation transformer that also does 230V -> 115V, you start hearing the improvements.

Note that some step-down transformers are autotransformers which do not offer isolation. However the step-down transformer inside your consumer equipment is almost always isolated.

PSP (fat) RMAA measurements

2012-01-16T21:09:00.007+08:00

This is a bit late (okay, a lot late), but here are the RMAA results for the first PSP.

PSP max volume unloaded
PSP max volume 33Ω

(Ignore the crosstalk measurements because they are wrong!)

Unloaded

33Ω

Because I want to see if I'm going to get an improvement with what I'm going to do.

Oh, and the output impedance is about 15.8Ω. Kinda high for some earphones. My X10 sounds like this:

TMPGEnc Video Mastering Works 5

2012-01-15T09:11:00.003+08:00

Ok, so I don't know when to call it a review and when to call it a quick look, so screw the concept, everything will be called reviews regardless of how good or rubbish they are. So I've joined the masses of creators of bad reviews.

Continuing from the previous review, I have decided to make x264 (but not Mediacoder!) the winner instead of MainConcept encoder used in TMPGEnc XPress.

Because I just tried out the new (ok... almost one-year-old) TMPGEnc Video Mastering Works 5. Despite the first impression that the webpage gives, it is really a very good program that is improved over TMPGEnc 4 Xpress.

Tha main thing of concern here is the removal of MainConcept H.264 encoder and replacing it with x264 (yes x264), CUDA encoder and Intel Media SDK (I presume is Quick Sync).

Screenshots - Mediacoder vs TMPGEnc 4.0 Xpress vs TMPGEnc Video Mastering Works 5

So when I tested it, I got a few seconds longer than 100 seconds. The sharpening is also there, confirming that it is applied by TMPGEnc (and not MainConcept encoder). But picture quality has noticeably improved over TMPGEnc 4's encode, making TMPGEnc Video Mastering Works 5 the clear winner in terms of image quality. Speed is a bit of an issue though. Consider this - when the encoding speed is supposedly limited by the decoding speed, TVMW5 took around 50% more time to encode compared to T4X. If we are dealing with HD encodes, the differences may get very big.

Funny thing is, TMPGEnc has always been slower than generic x264 encoders, but when TMPGEnc switched to x264 it got even slower. So the slowdown may not be due to the encoder but something else, like a lot of post-processing. Oh well, it works, so I'm not complaining.

CUDA encoder, speed is a few seconds shorter than 100 seconds. However the output video is unusably glitched and trashy.

Freemake vs Mediacoder vs TMPGEnc 4.0 Xpress - PSP Edition

2012-01-14T15:55:00.006+08:00

The rules are different from the previous review. Because the reason I wanted to do this in the first place is to see if I should use another encoder instead of TMPGEnc 4.0 Xpress to encode videos for my PSP. So this is a real-life problem with real-life limitations.

As for why I did not include TMPGEnc in the previous article and only post this article now, because it refused to read MKV again. Installing Shark007 did not help, but CCCP did.

Also, the original article which was supposed to be a quick look became an 8-way slugfest with 14 encodes instead as I spotted sources of errors and went to correct them. I swore to limit the number of encodes to five for all my future reviews, and I almost forgot about this already as I wrote this portion; I had to force myself to stop at five after realizing things are becoming messy.

Test setup

Input: 1920x1080 @ 23.976fps, 1m31s, 17.3Mbps H.264 (High@L5.1) (CABAC / 7 Ref Frames)
Output: One-pass, resolution 480x272, frame rate unchanged

Freemake: "Auto" was selected, resulted in a setting of 13MB, and a resulting filesize of 11.7MB.
Mediacoder: I trial-and-errored with this one, and with a bit of math settled on 62 quality with resultant filesize of 11.5MB on my third try.
TMPGEnc: The settings I have been using for PSP (quality-based, 24-25-27 for I-P-B frames respectively) resulted in a filesize of 11.2MB.

In short, I was very lucky. Or maybe not, because...

Where I screwed up

Freemake encoded at Main@L2.1, CABAC, 1 Ref Frame(s). I chose MP4 output and as mentioned in the previous article, settings are limited. But then I found a PSP output button which resulted in Baseline@2.1, 1 Ref Frame(s).
Freemake also did another encode with DXVA turned on.

Mediacoder encoded Baseline@L3.2, 2 Ref Frames, 1 B-Frame
Mediacoder also encoded another file at Baseline@L3.2, 3 Ref Frames, 3 B-Frames, which resulted in around the same encoding speed.

TMPGEnc encoded Main@L3.2, CABAC, 3 Ref Frames, 3 B-Frames

As you can see, this is very messy. I dunno why I encoded @ Baseline for Mediacoder, I have no excuse for that. But PSP definitely supports Main profile and CABAC, so there is little reason for Freemake to choose Baseline@2.1 for the PSP. So I'll ignore that encode and use the Main@L2.1 one.

Another screwup - I think CUDA acceleration was on in Freemake in the previous article as well as in the first encode for this article. However when I did another encode with CUDA off, the speed was slightly faster, like in the case for Mediacoder. Considering they both use x264, Freemake probably got the same image quality with and without CUDA turned on as was the case with Mediacoder. So no big harm done.

I also screwed up a little with aspect ratios - TMPGEnc and Freemake added black bars to make the video stay 16:9 (480x270 on a 480x272 screen), but I didn't do the same with Mediacoder. But in my defense, these were the default settings which I didn't touch. Also the impact is minimal.

For some reason Freemake decided to put the black bars only on the bottom, now this is not my fault.

Also, not as important, but I just wanted to say - TMPGEnc supports CUDA acceleration for filters only, so although I have a GTX 560 and CUDA is turned on, GPU was 0% all the way because no filters were used.

Encoding time comparison

You may be wondering, how come Freemake is vastly slower compared to Mediacoder this time round. Even TMPGEnc, which is slow as a m**********, is faster than Freemake.

This screenshot has the answer.

It is so easy to encode at 480x272, it takes more CPU power to decode say a second of the 1080p H.264 than to encode one at 480x272. So decoding speed becomes a more important factor at determining the encoding speed.

Because Mediacoder and TMPGEnc both used DirectShow to read the files, they decoded using LAV, or that is what I think because LAV is the only decoder capable of decoding Hi10P that is installed on my system. And LAV is a fast decoder. In comparison, Freemake used something else - probably built-in or Microsoft, I don't really know, but I know it is not LAV (at least not the same one) because it failed to read a Hi10P video properly.

This also means that DXVA decoding is going to help a lot with encoding speed, evident in the speed of Freemake with DXVA enabled. And most of the time we (as in consumers) convert a downloaded H.264 video to a smaller file with lower quality (why would you do the opposite?), so fast decoding is going to be important.

The great thing about using DirectShow to decode is that you can use any decoder and any filter with that encoder, and any settings that affect DirectShow players, like enhancements and subtitles, will also show up with that encoder.

Encoding quality comparison

Screenshots

Freemake looks better with DXVA on, interesting.

Mediacoder looks the same with or without the additional Ref and B-Frames.

Freemake is undoubtedly the worst out of the three.

TMPGEnc did some sharpening (I swear I didn't touch any setting). It was very nicely executed, because when everything is blur like **** at such low resolutions the extra sharpness comes in handy. Funnily I don't recall seeing this sharpening when downscaling is not involved, so it must have been added by TMPGEnc during downscaling, which is very smart. I mean, look at the screenshots, this is the first time I can safely say that there is no reason to not apply this sharpening.

However, this means TMPGEnc and Mediacoder are apple and orange.

Conclusion - which encoder is right for me?
(Because it isn't always about you)

Things learnt today:

- An encoder (or encoder settings) that is good in scenario A may not be the best in scenario B
- Decoding speed (and acceleration) is very important when encoding at low resolutions and bitrates
- DirectShow is important

So Freemake, despite poorer quality, was slower due to decoding speed, but after DXVA was enabled it was very fast. But still, the image quality (which improved with DXVA enabled?) has room for improvement. But more importantly, without Hi10P support it is going to be hard for me to use it, with more and more encodes going Hi10P these days.

Mediacoder... nothing much to say about it. TMPGEnc's extra sharpening has benefits that greatly outweigh the costs. Looks like I'm sticking to TMPGEnc.

Update 15/01/12 - Is TMPGEnc Xpress really the winner? Does the extra sharpening really make it better over Mediacoder using x264? What if one can get the best of both worlds? Find out here.

2012-01-08T08:34:00.001+08:00

[Review] Eight-Way H.264 Encoders Shootout: Arcsoft MediaConverter, Badaboom, Cyberlink MediaEspresso, Freemake, Handbrake, MediaCoder, SUPER, and TEncoder

2012-01-02T18:19:00.002+08:00

Update 14/01/12 - Is what is said in this review the absolute truth? Find out at Freemake vs Mediacoder vs TMPGEnc - PSP edition.

Table of Contents
Reason for doing this
Before we start - Frontends for x264 and encoding settings
GUI and feature comparison
Test setup
- Why two pass?
Encoding time comparison
Encoding quality comparison
- Thinning the herd
- Round one
- Round two
Conclusion - Which encoder is right for you?

Reason for doing this

So I got an i3-2100 into the family. So I now have GPUs capable of APP, CUDA, and Quick Sync. It was natural for me to do this.

You see, the issue with all the GPU-accelerated encoding is that they can only be done with a few specific software, and what the major hardware reviewing websites do is just compare GPU vs CPU in the same encoder. Back in the days of reviewing MPEG1 encoders I already learnt that different encoders' quality and speed can vary greatly, and usually newcomers perform significantly worse than tried-and-tested solutions. So sure, GPU may be much better than CPU with this particular encoder, but it is possible that another CPU-only encoder is still the better choice, particularly when compression ratio is important.

Before we start - Frontends for x264 and encoding settings

x264 is a free command-line H.264 encoder, and many video converter software use it to do the conversion. Thing is, x264 is just a video converter, so it needs another program to do the file reading, resizing and stuff. Plus most people cannot live without a GUI. That's where all these different software using x264 come in.

Software known to be using x264:
Handbrake
MediaCoder
SUPER
TEncoder
And many more

However, even though all these programs are using the same x264 to encode, they differ in their default settings and the settings available to the user, so features and encoding quality can be vastly different (as we shall soon see).

Well, I will only be testing a few out of the many available, because they are popular/used to be popular. No point testing too many of the same thing.

I will be testing all encoders, both x264-based and not, at as default settings as possible, to simulate what a novice will get after just installing the program.

GUI and feature comparison

First, the CPU-only programs:

Freemake Video Converter

Rather bare settings but enough to get the job done - we get to choose the bitrate, get to choose between one or two-pass, get to limit the file size (but I managed to control the file size via bitrate anyway, so question mark? Maybe it's more useful for one-pass.). No info on how the one-pass is done - I didn't find it out.

Handbrake

Handbrake is the second-best program in the list in terms of the settings you are allowed to tweak, we get two-pass encoding, reference frames B-frames motion estimation deblocking blah... stuff you find in a proper encoder. I would've liked a "maximum bitrate" setting, although the H.264 encoding profiles and levels take care of this somewhat, but as I was typing this I realize this setting isn't available either. But well, there is a ton of other settings, plus the command-line is also available for you to enter the parameters that are not on the menu.

SUPER

The only important setting we get is bitrate. No two-pass on this guy - only one-pass is available, and since there is only a single bitrate setting, it is either A) Constant bitrate, which sucks for quality or B) Constant quality, which means the bitrate kbps is "for reference only", in that the file size can pretty much be anything, which sucks if you are encoding for a portable device. Ironically, we actually get profiles and levels settings.

TEncoder

Somewhere between idiot-proof and geek-only, we get bitrate, two-pass, a bit of extra stuff, that's it. But it also has the option to enter command-line stuff, so it can be as powerful as how much your brain can remember the settings.

Now, the GPU-accelerated programs:

Arcsoft MediaConverter (Tested: AMD APP, Intel Quick Sync and CPU-only)

Arcsoft's MediaConverter supports AMD APP, nVidia CUDA, and Intel Quick Sync. Minimal settings, but some H.264 encoding options are available. Has a 5120kbps lower limit for bitrate.

Badaboom (Tested: nVidia CUDA and CPU-only)

We get to choose between Baseline and Main profiles - High is not available. Bitrate control... either VBR or constant quality... which pretty much mean the same thing because this thing only does one-pass and it doesn't even bother to tell you what bitrate it will try to get.

Cyberlink MediaEspresso (Tested: Intel Quick Sync)

So... we are stuck with 6Mbps minimum for bitrate. I'm seeing a trend here. (Also, remember AVIVO?) Btw, that's all you get for encoder settings.

Ok, so there is also choice for speed or quality, and whether you want to use acceleration. Both the HD 5670 and integrated graphics were installed and being used at the same time, but only Quick Sync got detected. Wierd.

I only tested Quick Sync, because when I unchecked "Enable hardware encoding":

Because I tested the trial version. Just so you know, all software tested were legit - either free or free to try.

MediaCoder

The only free GPU-accelerated encoder being tested is also the most promising-looking encoder being tested. On first glance it looks like a pretty ok interface, kind of limited but good enough to do a good job. It even has max/min bitrate and buffer size options. And then you press the Advanced button.

This is good stuff. If you want to be knowledgeable in video encoding, then look up what all the settings mean. The list goes on, and there is also a place to enter command-line parameters, but would you really need it...?

Test setup

Input: 1920x1080 @ 23.976fps, 1m31s, 17.3Mbps H.264 (High@L5.1) (CABAC / 7 Ref Frames)
Output: 4Mbps*, two-pass if possible, resolution and frame rate unchanged, other parameters as per encoder default

*Arcsoft MediaConverter and Cyberlink MediaEspresso encoded at 6Mbps, because 4Mbps was not available

Computer hardware:

For Arcsoft MediaConverter and Cyberlink MediaEspresso -
Core i3-2100
ASRock H61M-HVS
HD 5670 and HD3000
4GB ram

For all the rest -

Phenom X3 720
GTX 560
2GB ram

Why two-pass?

We encode to get more visual data out of less disk space. A good encoder compresses more data into the same file size. So to see an encoder's performance, we fix the file size and see the image quality.

Problem is, two-pass is one of the only two methods (not counting more passes) which allows the resultant file size to be close to what you set. The other method is constant bitrate (a one-pass method), which is totally suck and hence stupid and hence not used.

Which is why I have always used two-pass encodes to compare video encoders. Problem is, most of the new GPU-accelerated encoders only support one-pass. I have to trial-and-error to get a similar bitrate by playing with the quality-based encoding settings, and it is not easy.

Encoding time comparison

*Tested on the other system
^Two-pass encoding

Okay, so CUDA did increase speed significantly in Badaboom. But in MediaCoder, CUDA actually decelerated the encoding slightly. There seems to be some problem in Arcsoft because all three encodes took the same time. According to AnandTech's review we're supposed to see a significant speed boost with Quick Sync here.

According to this review here, Arcsoft's Quick Sync is supposed to be much faster than Badaboom. So if Quick Sync was faulty in my case, then the expected time should have been very short.

Generally, the encoders that are advertised to have GPU-acceleration all encoded much faster than those using x264 and other CPU-only encoding, even if they only used CPU. But note that the CPU-only encoders mostly encoded in two-pass, so the speed difference is not as big as what the graph may seem to be showing. In fact Freemake's one-pass took only 100 seconds.

So, do the GPU-accelerated programs encode faster because of GPU acceleration or because they choose speed over quality? Maybe both. That is what this review is going to find out.

Encoding quality comparison

Screenshots - all
Screenshots - sorted (see "Thinning the herd")

I ended up having 7 x 15 = 105 pictures, which is too many even if I display the links in a table (this blog's width is too small!), so I have them zipped into a single file. Anyway I won't be able to analyze all (14 choose 2) = 91 different encoder pairs possible, so I will be doing my comparisons like this -

1) Encoders will be categorized into different speed classes
2) The best (subjective) encoder in that particular speed class is picked
3) The best of each speed class are compared

The speed classes are -

Fast: Arcsoft, Badaboom CUDA, Cyberlink
Moderate: Badaboom CPU, Freemake, Handbrake FFMpeg, SUPER
Slow: Handbrake x264, Mediacoder, TEncoder

Note that speed class of an encode does not depend on whether it used GPU-acceleration or two-pass. Although the slow ones are all two-pass, which is kind of unfair because encoding time is almost doubled just for better filesize estimation and a possible slight increase in image quality. I will consider this when making my conclusion.

Thinning the herd

Cyberlink MediaEspresso's fast preset (the default one) actually resulted in a bitrate of ~4Mbps, even though 6Mbps was selected. The quality preset resulted in 6Mbps as it should. Because Arcsoft MediaConverter also encoded in 6Mbps, and Cyberlink's fast preset is on the verge of garbage, the encode by Cyberlink's fast preset will not be considered in the comparison.

Quick comparison also reveals that Badaboom's GPU-accelerated output is the garbage even when compared to Cyberlink's fast preset, despite having a similar bitrate and encoding speed. Hence it will also not be considered for comparison.

Arcsoft's APP and CPU-only settings resulted in identical pictures. And same encoding speed. Maybe the exact same things were done. Then again, Quick Sync resulted in identical speed but different picture. Nonetheless the APP-accelerated output will be ignored.

In the moderate speed category, Badaboom CPU is crap, and MPEG-4 SP (FFMpeg) was included for fun, so there are only two real candidates. And one of them encoded two-pass. One can probably already see where this is going.

Round one

Frame A

Let's start off with an easy one. All encoders (that are still being compared) produced decent to excellent results, Arcsoft's Quick Sync produced the best result in the fast category, Freemake takes a win from SUPER, and everybody in the slow category performed similarly.

Frame B

Poor bitrate allocation strikes hard in this frame. Many encodes are below decent, but the encoders capable of two-pass encoding still have good output. This is why you should do two-pass encoding, or choose quality-based encoding with tight quantization and flexible bitrate and frame type control.

Ranking of the fast encoders is subjective, but I would say that Cyberlink is the best. Freemake wins SUPER again, and the slow category has differences but hard to rank - you be the judge.

Frame C

Another easy frame which also tests how the encoder chooses what information to omit. Outside of the display screen and pavement, all encoders perform similarly and excellently. But the encoders perform differently when encoding the screen - most soften the lines on the screen to various extents, the worst being Arcsoft's Quick Sync. I would say that the best of each category are Arcsoft's CPU, Freemake and TEncoder.

Frame D

Checking for color banding and macroblocking. Around half of the outputs ended up as casualties, there are visible differences between the slow encoders, but even so it is still hard to subjectively judge which is better. Arcsoft CPU had the least artifacts in the fast class, Freemake in the moderate class.

Frame E

No particular reason for choosing this frame, but it turned out that none of the encoders managed to preserve the lines on the display screen properly. Only the two-pass encodes managed to preserve part of those lines. Again it is hard to tell the differences between the slow encoders, but I'd say that Arcsoft CPU is the winner of the fast ones here. Freemake comes out ahead of SUPER again.

Frame F

Cyberlink is the clear winner followed by Arcsoft Quick Sync then CPU-only in the fast class. In the moderate class, Freemake wins yet again. (I know, it gets boring.) Really hard to choose for the slow class.

Frame G

This scene is very easy to encode, contrary to my expectations. Arcsoft Quick Sync just looks that slightly better in the fast class, Freemake is noticeably better than SUPER, and all the slow ones look exactly the same.

Verdict for round one

Quick Sync gave both better and worse quality compared to CPU in Arcsoft; Quick Sync encodes the items that are the center of attraction more clearly but blurs the not-so-obvious. Well all encoders do that, they just differ in how much. I would say Quick Sync will be subjectively more pleasing, and after comparing the moving pictures side-by-side, I confirmed this to be the case. Despite CPU-only scoring so well with still frames.

But, if the still frames of Arcsoft Quick Sync and Cyberlink Quick Sync are compared to each other, Cyberlink's seems like a more balanced and all-rounded improvement over Arcsoft's. (Stupid subjective comparisons.) After comparing the moving pictures, I reluctantly gave Arcsoft's Quick Sync the win. Although really, you can't be too bad with any of the three.

In the moderate speed class, no fight here, Freemake wins, 'nuff said.

For the slow encoders, 4Mbps is too much bitrate to do a subjective comparison with, because all encodes looked the same and so similar to the source. Meanwhile some GPU-accelerated encoders are producing garbage with 4Mbps. This is why I always say that some encoders are garbage.

Now, if only I can get AVISynth to work on my Windows 7 installation, I would be using MSU Video Quality Measurement Tool to get some objective numbers. Maybe next time.

Since the slow encoders all performed so similarly, I will be picking the winner based on speed and features, and MediaCoder is the winner. It deserves the win. And should be your first choice for many things.

Round two - Arcsoft (Quick Sync) vs Freemake vs Mediacoder

Arcsoft's output is acceptably decent, although there is room for improvement. In 2 or 3 of the 7 frames, it looks indistinguishable from the slower encoders. But the point of throwing more bitrate and using better encoders is to minimize the artifacts that only appear in the more difficult scenes, the presence of which can spoil the experience.

Mediacoder is just that slightly better than Freemake, in certain scenes, and if you're lucky. Doesn't change that fact that it still is better.

Conclusion - which encoder is right for you?

Freemake gets special commendation for high-end quality at a reasonably fast speed, and has just enough features to get the basic things done. And it is free.
However, something needs to be mentioned - Freemake's one-pass did not yield decent quality. Remember when I said above that two-pass is just for better filesize estimation? Well this is only true if the rules of bitrate allocation (and other blahs) are the same. Apparently, it wasn't. But then its one-pass was more than twice as fast as the two-pass which gave a hint.
Also, I cannot give Freemake a complete victory (I'm ignoring one-pass performance), because the lack of profile and level and some etc settings can result in the encode being unable to be played on some players. So, not good enough.

Best encoder in terms of quality and features is MediaCoder no doubt. I'll have to try its one-pass someday. But it is slow, as x264 is.

Sadly, no GPU-accelerated fast encoders managed to match the slow CPU ones' quality. If there are visible differences at 4Mbps, then at 2 or 1Mbps it could become garbage already. But Arcsoft encodes so fast with acceptable quality even in CPU-only mode, it is worth considering if you need the speed - just make sure the quality stays acceptable. If Quick Sync increases speed even further, than just wow.

I'm actually tempted to give Arcsoft best speed/quality ratio if the Quick Sync worked properly, but there are different encoders for different jobs. Personally I would choose MediaCoder, because I don't need speed and I need quality and reliability (of playback of encoded videos by the players).

What high source output impedance meets uneven load impedance sounds like

2011-12-27T09:00:00.001+08:00

This has been sitting on my desktop for a while so I uploaded it before deleting it.

I did do this with an MS-1 last time, but MS-1's impedance is fairly flat so there isn't much to be heard. This one is more extreme.

Klipsch X10 Xonar DX recorded vs original (original from 01m29s onwards).mp3 - MediaFire

A quick look at three laptop power adapters

2011-12-24T09:29:00.001+08:00

So due to some circumstances, I have three laptop adapters, and a lapdesktop to power. Might as well see which is the most efficient.

Test setup:
Acer Aspire 4520 "lapdesktop" powered by AMD Athlon 64 X2 TK-53 @ 1.7GHz
OCCT Linpack used for CPU stressing/loading

Lite-On

This guy came bundled with one of the Acer laptops (or one ex-laptop) in the house. Acer, like many others, has always been using Lite-On adapters without relabelling them. Pity I couldn't find the other one, as it would be fun to see if they perform similarly.

Compaq

This thing came with an AMD K6 laptop (400 or 450MHz?), so it is at least ten years old.

eBay

...or should I call it "Replacement AC Adapter"? Standard generic China junk, I bought it on eBay so I'll call it that. It is lighter compared to the other two, and other laptop power adapters in general.

Comparison

Idle
Lite-On: 16.8W
Compaq: 16.1W
eBay: 16.7W

Load
Lite-On: 43.4W
Compaq: 42.5W
eBay: 43.2W

Surprisingly, the Compaq was the most efficient out of the three, despite it being ten-year-old technology. The numbers may not look that significant, but we're looking at 1-2% better efficiency during load and even higher when idling.

The eBay did just slightly better than the Lite-On at its job.

Conclusion

Not much to conclude really, because your power adapters, new and old, are going to be different from mine.

The eBay was slightly more efficient than the Lite-On, so should we buy it? Answer: No. I just blew it, not by Linpack which it survived, but by shutting down the lapdesktop, plugging out the power connector on the lapdesktop, and plugging it in again. Pop, smell, and the circuit breaker tripped from over-current.

Xonar DX output impedance

2011-12-13T23:45:00.005+08:00

= ~86Ω

Well you don't have to read the rest of my blabbering, but if you want to,

Well I needed to find this, so might as well give it its own post so people can find it on Google.

So I measured this guy, -11.1 to -11.2dB with a 33ohm resistor load. Well that's the problem with my current measurement method and the use of a logarithmic scale. The output impedance is in the range of ~86Ω, and highly resistive in nature, at least that's what the FFTs say.

No load (or whatever the load of Xonar DX's input is)

33Ω resistor

Well you can see that the frequency-dependent portion starts at below 20Hz.

Hmm why 86Ω? 82Ω resistor + 4Ω from everything else? Only by looking at the card can I know for sure, and I don't feel like doing that.

But well, with such an output impedance (resistance), it isn't going to drive phones without screwing up the frequency response. (This sentence have something to do with another post coming up.)

Diminishing returns

2011-12-07T17:32:00.001+08:00

"As with all audiophile products, diminishing return rule indeed..." - quoted from somewhere

But, there is another camp that believes the opposite - the higher you go, the more something affects.

How about the third camp -

It is possible to have both decreasing and increasing returns.

Because placebo effect doesn't care if you think the returns are increasing or decreasing - that is how you interpret the improvements... that don't exist.

[Review] Zalman Shenzhen Ice Dragon II (冰龙II) CH-90A-06 Heatsink

2011-11-26T09:07:00.157+08:00

Why I bought this a.k.a. the part you can skip unless you want to see me talk about myself, which is what blogs are apparently for

I'm going to build an LGA 1155 system soon, and the SG-02F that is going to be the enclosure does not have much clearance for a tall heatsink (I still have the Havik btw). And being the silent PC advocate that I am, I cannot stand using the stock heatsinks provided by Intel (and AMD). So I needed another aftermarket cooler, and one that is low enough.

But all the heatsink makers are crazy about gigantic tower coolers, and with few low-profile, top-blow-to-bottom designs available, their prices are expensive for the performance they deliver.

So I looked around, and found an interesting choice.

And decided to buy it, the reason why you can read below.

Zalman's "flower" coolers a.k.a. the part you can skip unless you care about the background of this thing

Zalman Shenzhen is apparently a legit company - the China division of Zalman which caters to the Chinese market with low-price products (the prices of "normal" Zalman products are rather high). Various "flower" heatsinks are being produced and sold, and the Ice Dragon II (冰龙II) CH-90A-06 is one of them.

The first thing to notice when looking at this thing is the resemblance to the old CNPS-7000.

And it is no wonder that Zalman Shenzhen is still continuing to produce heatsinks based on this design. This "flower" design is so efficient, that CNPS-7000 was one of the best coolers in its day and being also one of the smallest (62mm height) at the same time, and this flower design was continued with newer products until Zalman decided to switch back to the more conventional methods because, somehow, the newer products did not achieve the same type of exceptional performance the original had over its competitors while also costing a lot. Perhaps this design does not scale well.

But the roundness of the design is not the only thing that is unique to this heatsink - radial coolers have been in use since the Pentium 1 days, Thermaltake had their Orb series of coolers, and Intel has been using radial coolers for their stock heatsinks (again) since halfway through the Pentium 4 era, another testament to the efficiency of such a design.

Pentium heatsink

With the fan

What Zalman does differently from the others is the construction - many slices of metals (with the shape of the fins) are pressed together at the middle, then each individual piece bent and spread out into the shape of the flower heatsink. This construction method is also used for Zalman's heatsinks prior to CNPS-7000. Zalman holds the patent to this.

You can see how the difference pieces are pressed together in this above picture.

This construction has many advantages - thin fins are possible (the fins on the Intel stock heatsinks and Thermaltake Orbs are thick in comparison), which allows for more fins and bigger surface area. Each individual fins are in direct contact with the CPU. Plus it is probably much cheaper to manufacture a radial heatsink this way, evident from the price of this heatsink.

Unfortunately this also has its share of disadvantages - the fins are thin and thus heat transfer is an issue, and the parts furthest away from the CPU are receiving the most airflow. It is difficult to implement the use of heatpipes (the flower coolers with heatpipes have a different construction), so cooling efficiency with newer high-power CPUs becomes a possible area of concern. Also the fins are flimsy.

But all in all, the CNPS-7000 was a high-performing small-size cooler. A thermal resistance of ~0.30°C/W at low is nothing to laugh at even today.

So when I saw a descendant of this heatsink, sold at a cheap price of SGD$12 after shipping, why not?

Zalman Shenzhen Ice Dragon II (冰龙II) CH-90A-06 a.k.a. the part you can skip unless... heck... just skip it

This is not the only "flower" heatsink available on eBay - there are three from Zalman Shenzhen alone - Ice Dragon (冰龙), Ice Dragon II (冰龙II), and Polar Bear (北极熊). Then there are the clones from PC Cooler (which Zalman has successfully sued before), and a few from unknown brands. I actually took fancy to a model from an unknown brand, which looks well-built and is even cheaper, but I bought the Ice Dragon II (冰龙II) in the end because I believe in Zalman, and it is better for me to review a newer product that won't be disappearing for a while. Well this is one of the reasons I chose Ice Dragon II over Ice Dragon - because it is newer and probably is the replacement, but Ice Dragon II also has a higher fin count than Ice Dragon, although Ice Dragon's fins look more solidly built. So it was a hard decision for me. Polar Bear was not in my consideration due to a smaller and faster fan.

So after two weeks of waiting, here it is.

Oo nice dragon.

At this point I went and ordered the slave flash from eBay. This camera's flash sucks balls.

I mentioned about the fins being flimsy. Here you can see the fins at the top right were bent during transit.

Another angle.

Just as easy as they can be bent, I bent back the fins.

Not possible to totally unbend them, but at least they look uniform now.

Zalman-branded thermal paste, which I will probably not use, and leave it to dry up and expire. And the mounting pins.

The retention module. The heatsink is designed to work with the stock AMD bracket, so this probably works as a 1155/775 to AM2/3/FM1 adapter.

Instructions, in Chinese, of course. Warranty at the left. The fan cable actually has a sticker that says "warranty void if torn" and some kind of a serial number. Does the expensive case fans come with warranty? I don't know since I don't buy them.

More instructions.

Alright, lets take a look at the heatsink. Uses a clip like in the old days of Athlon XP, the old days when I swear a lot during installation because it is hard. There's a reason why nobody uses this anymore - even the stock AMD heatsinks' clips has changed. 3-pin fan, but I don't care about 4-pin or not.

The base is fairly smooth for a SGD$12 product.

Boy this is small. And I have a small hand, so this might be even smaller than what you think.

When looked at this way, a difference from CNPS-7000 becomes visible - the top part of the heatsink where the fins are not curving is significantly shorter than the CNPS-7000. But this is kinda expected, as at a height of 49mm (!) it is seriously shorter than the CNPS-7000, and is seriously short. But the depth of the fan is around the same, so the area with effective airflow is also around the same too.

...

So my English fails me, I don't know how many of you reading actually understood the above paragraph. But heck about that. The fin count of Ice Dragon II is 52 per semicircular half, while that of CNPS-7000 should be 65~70 if I'm not wrong. Ice Dragon has 40+. Lower fin count = lower fin density = lower area but less resistance, with low wind speeds this can go either way. Lets test this sucka.

Test setup a.k.a. the part you can skip unless you actually want to know if the results are legit

CPU: Intel Core i3-2100 @ stock
Thermal paste: Anabond
Motherboard: ASRock H61M-HVS
Software: OCCT Linpack

This is the heatsink I'm going to test the Ice Dragon II against: i3-2100 stock cooler, with a 0.60A fan.

Intel's recent stock coolers have reduced in size. Well yeah, that's good. Efficient use of surface area. If it's too tall the wind won't reach the bottom anyway.

As you can see, the Ice Dragon II is about as tall as the stock heatsink, which as we all know is quite short. Low-profile coolers, both definitely are.

Test results a.k.a. the part you can skip unless you buy things based on their performance, and not because everybody else is buying

So I started measuring the idle and load temperatures of the Intel stock cooler at two different fan speeds (~1200rpm and ~2200rpm).

Then I installed the Ice Dragon II.

The measured fan speed at full is ~2400rpm, which is higher than the rated 2200rpm though not by a lot, but closer to the 2500rpm of the Ice Dragon and what they should have called it (the speed I mean). And boy is it loud for a 2400rpm.

The measured temperatures are...

...

I switched off the com and reapplied the analbond.

Tested again.

Same.

Switched back to the stock cooler. Tested again. The stock cooler performed as it should.

I stopped the whole testing immediately.

You see, the Ice Dragon II performed so bad that even at full speed its temperatures were worse than the stock cooler's at low (~1200rpm). There is no point in finding precisely how much it sucks if it sucks more than the stock cooler.

How the thermal paste spread out suggests insufficient mounting pressure, but there seems to be no easy way to improve on this.

Verdict a.k.a. the part you can skip unless you want to read the summary of the whole thing

This thing sucks. Period. I do not know if this is representative of the other "flower" coolers, but if you see one using the same mounting mechanism, and made by Zalman Shenzhen in particular, proceed with caution.

I wonder if Mayoi would like to have this flower for her shell.

Extra! a.k.a. the part you can skip unless you reached this review after seeing my WTS post

I want to see if people actually do their own research when they are buying something they don't know about and reach this page. If you did, say hi.

Remember, never trust the words of a salesman.

My sentiments too!

2011-11-24T10:09:00.001+08:00

Too many cars in Singapore

2011-11-11T23:06:00.005+08:00

I mean, look at how crowded this carpark is!

Just to prove that I didn't take a picture of some deserted building in a tech park or business district at night, click for the bigger version and look at the red light on the building in the background.

Moving closer...

The words on the building are "THE FRONTIER".

Yupz, Jurong Point on the right (which is beside Boon Lay MRT).

While this picture was taken on deck 5A, by the third floor there was already many unoccupied parking lots. I just took the photo on the 5th floor for better impact and for the red words on the building to be also visible. And technically, there is only one car in the picture, the other 3 are goods vehicles, cat C COE.

And no I didn't hunt for this particular carpark, I just randomly picked one, and one that is close to Jurong West Central, one of the most overcrowded points in Singapore. If I went to Punggol, I'd be seeing the same emptiness by deck 3. At night when everybody is at home. I mean, this isn't a new discovery, many people know that less than half of many MCPs (multi-storey carparks) are unused, because people can't afford the goddamn car. And the percentage of people owning a car can only drop as COE growth rate is lower than population growth rate (0.5% vs 2.1%). So I can bet that the number of cars owned by humble, honest people living in HDB flats is going to go down over time, not increase.

So... the number of parking lots in the MCPs is overkill by at least 100%. Well, better to be safe than sorry since it doesn't cost that much to add a few more storeys. Meanwhile in some other places the number of cars owned by residents is more than the number of parking lots.

Geez...

It's a black world

2011-11-09T19:17:00.000+08:00

Black laptop bottom, black laptop sleeve, black phone, black PSP, black calculator, black camera pouch with a black camera inside, black wallet, black mouse, black keyboard, black pencil case, black cables, black power adapter, black speakers, black USB hub, black monitor base (the monitor itself is dark purple, well mostly black), even the memory cards at the bottom right are black.

Remember the time when everything was beige? Well now's the black age.

Why the basic CMoy's simple resister divider virtual ground is not a good idea

2011-11-02T18:17:00.009+08:00

I was thinking about why the simple resister divider virtual ground in a simple CMoy is inferior to active solutions, say, an active rail-splitter or a buffered resister divider. The usual answer would be "drifting of the virtual ground voltage as the amplifier pushes current into the load", or (over)simply said, "uneven splitting".

But, how much?

I started off thinking about how the load is connected in a simple resister divider virtual ground (SRDVG) CMoy.

I did not include the input circuitry to keep things clean. The values of the resistances are too large to have any significant current flow.

Then using small-signal analysis model, power supply rails are converted to ground:

It is clear that the two 4.7kΩ resistors in parallel is too high a resistance compared to the load. Most of the voltage from the amplifier will appear across the 4.7kΩ resistors (effectively 2.35kΩ)

So a capacitor is placed in parallel to the 4.7kΩ resistors to reduce the impedance to AC signals.

The capacitor appears as a short-circuit to AC, so no AC current flows across the 4.7kΩ resistors, so they can be removed in the AC model.

Wait...

The circuit on the right is the "capacitor in series with speaker load omg" which has somehow become a taboo. (As in, anything with output capacitor is immediately labeled as "suck".) Actually, the left circuit is worse than the right, because the finite impedance of the capacitor results in crosstalk when the other channel is connected to the same capacitor.

I still need to do my initial objective - to see by how much this is bad.

One thing good about using capacitors, (assuming ideal) is that for a fixed load the performance does not depend on volume, while active virtual grounds can handle a certain amount of current, above which unexpected things happen.

On the other hand, the circuit's performance becomes frequency-dependent, requiring huge capacitor values to avoid high impedance at lower frequencies and noticeable bass roll-off.

With a 33Ω load, in order to have a cutoff frequency of 20Hz, 241µF is needed. There's the second channel too, so 482µF, lets round it down to 470µF.

Actually, this isn't as bad as I thought. With higher-impedance loads, the capacitance required would be even less.

The impedance of a 235µF capacitor @ 20Hz is -j33.86Ω. Well, it should be around there, since |Vtotal| = √2*Re(Vtotal). Or in English, just know that in an RC high-pass-filter or low-pass-filter, the cutoff frequency is the frequency at which the capacitor's reactance, or 1/2πfC = the resistor's resistance R.

But, compare this with the 1-digit Ω or less impedance that active devices provide, 33.86Ω is a whole lot. Well... "whole lot" enough to reduce the voltage across the speaker load by √2 yea?

Lets see what happens at 100Hz -

Capacitor reactance = 6.77Ω
Resulting attenuation = -0.18dB
Phase shift = -11.6º
Oh yea there's still phase shift to worry about... how much phase shift is bad is debatable, I'll just leave the numbers here.

In comparison, lets try an active virtual ground with 10mA for each channel. Less attenuation issue, less crosstalk issue, less phase shift, but 0.33V max with a 33Ω load. Maximum allowable voltage increases as load impedance increases though. If a buffer capable of 100mA (for each channel) is used, then max voltage is 3.3V with 33Ω and 15V with 150Ω. That's more than ever needed.

As you can see, both capacitor-coupled output and active virtual ground methods have their own pluses and minuses. If cost and complexity permits I'd go for high-current active virtual ground, but very high current solutions can be expensive to have sometimes.

Back to the topic in the post title, how is the simple resistor divider virtual ground (SRDVG)(inclusive of capacitor C1) solution? A big fat no as I see it. While most circuits have their pros and cons vs others, the SRDVG is simply outclassed by the typical method of using DC-blocking capacitors. Unless you really have no space or budget for two additional 470uF capacitors (lets overkill a bit, shall we?)

Moral of story: Just because it doesn't seem to be "in series with the output" doesn't mean it isn't.

I need to reconsider the practice of installing cables without using my eyes

2011-10-31T21:56:00.002+08:00

This ability is acquirable through practice, from installing computers and televisions many times. With little working room available, being able to connect the cables after a glimpse of the connectors' positions is very useful.

But at the same time, this practice can be dangerous.

So I was connecting my speakers to one of my amplifiers. I saw the black connector of the cable was already in. So I took the red one, inserted it, then turned on the amp.

Nothing.

Well not totally nothing, the other speaker is working fine.

Reinserted the connectors on the amp side. Still nothing.

Looked at the connectors on the speaker side in order to reinsert them.

Both the red and black connectors plugged into the black terminals.

It can happen because my speakers are bi-wireable (or bi-ampable for people who don't acknowledge the legitimacy of that concept), and the terminals are arranged in a neat row:

I wasn't worried, because the short-circuit protection of this amplifier is tried and tested (not the first time I accidentally shorted it, which was 50% due to my own fault).

But try that on another amp, and who knows. I wouldn't try that on circuits designed by d.i.y.ers, for example.

This happened because I did not look at things when I am touching them. I did look at the black connector, but s*** happens, and at least if I had looked, I would have noticed the mistake before powering up.

While the "eyes-free" method is (relatively?) safe with computers and AV equipment, this is because of the unspoken rule of not using the same connector for two different things. As in two different types of signaling, voltage and etc. (Different USB devices use the same USB connection, and different-colored analogue RCAs, that's fine.) Or at the very least, if the same connector has to be used, make it in such a way that things don't explode when you plug them into the wrong hole.

Because, even if they are looking, people will plug any thing into any hole that fits. And push any switch that they can. I wonder what that 230V switch does.

This is one of the reasons why the USB connector is square on the equipment side (because connecting two computers via a normal USB cable will cause things to fry), the LPT cable is different on the computer vs the printer, why the 15-pin midi port is different from 15-pin VGA D-sub, why some connectors must be male and why some female. Apart from the obvious money from licensing. Well yea.

Well, audio is a place where such rules of safety (there are others too) cannot be properly followed, due to too many different requirements, the need for compatibility and the lack of customized standards.

(Speakon connector, something that solves multiple safety issues with regards to speaker connections.)

However, many places are still using bare wires. Carries anything from turntable output to 230V live AC. Ok, maybe live AC is exaggerating, but speaker cables capable of a few amps is possible.

And those exposed connectors, omg. I don't care if it is isolated (from the mains) or not, 100V is still 100V and it still burns you and stuff.

It would take some effort to child-proof this thing. But then again, the way people place their equipment on stands and platforms, they are probably not intending to have babies.

Anyway, my point is that getting high-current wires with opposite polarities to share the same type of connectors which are placed next to each other is a very bad idea, even if the connectors are color-coded.

Credits: All pictures are from the internet.

Random thought: Watching Ikoku Meiro no Croisée...

2011-10-30T18:08:00.000+08:00

HOW CAN ANYONE HATE CHEESE???

I mean, I hate it for killing my arteries and for being overly-expensive, but if I were immortal and rich, I would add cheese onto everything.

Cheese between bread + ketchup makes for a quick and nice snack.

Oh, ketchup, the wonder condiment. The brands available in Southeast Asia are mostly nasty though. Including Maggi. Maggi ketchup sucks. Sinsin and Kimball are also nice, but I find myself going for Heinz most of the time for both tomato and chili sauces.

Sinsin's garlic chili (the one McDonald's uses) is also a winner and also easily found, but don't buy the bottle version if you don't use it often. Something happens when it is left for a while.

How to avoid buying fake or overpriced products

2011-10-28T15:41:00.002+08:00

The audiophile market is filled with unscrupulous sellers, people who take something worth very little and sell it for a lot more than it should be.

This happens in some industries (not just audio) because it is very difficult to verify the product's performance. You can't tell if that that hair-growth formula really works, if that toothpaste really whitens, if that weight-loss pill is not poison, or if this exercise machine is not actually a... wtf.

A seller can simply take a mediocre component, relabel it as something that costs more, and sell it to unsuspecting buyers who are too (insert adjective here) to verify its authenticity. Or he can put a high price tag on a worthless design and sell it to unsuspecting buyers who are too (insert adjective here) to verify its performance. Or, the seller claims he "made improvements" to the original design and sells it for twice as much.

Fortunately, there is a way around this.

Sellers usually sell more than one product. Compare the prices of the other products with alternatives from other manufacturers.

For example, a seller in China is selling OPA627s at $10 each.

You don't know if it is fake or really cheap.

But, this seller also sells NE5532, at $2 each.

Shipping is definitely less than $1, because I have bought things @ $1 inclusive of shipping.

So even after subtracting shipping, it still costs more than $1.

Which is more expensive than what element14 sells it at.

If the China seller sells the NE5532 more expensively than a big trustable company, do you think he would sell a real OPA627 at a price much lower than that of the trusted company?

If one example is not enough (sometimes it is indeed not accurate), look for other products. If he sells all the other products overpriced, except for those that are widely counterfeited, then the signs are clear.

The same trick can be used for non-counterfeit goods, but products that you think might be overpriced.

For example, the seller of an amplifier board also sells power supplies. The PSUs may be just some capacitors on a PCB, or an SMPSes from another company.

He sells the amplifier board at $100 each.

He sells the PSU at $100 each

And you know you can get the equivalent somewhere else for much less.

Since he prices the amplifier board at the same price as the PSU, he is trying to say that the amplifier and PSU have the same value, which is $100. But you know that the PSU is worth much less, so if the amplifier is worth the same, then it is also worth much lower than $100.

Don't tell me "design cost". Whatever the cost, parts or design, is all grouped under cost. And the cost of a product, is, say, $200. You can build it yourself for $200 (inclusive of design and labor costs), or you can buy it from someone else for $200, it still costs $200. Then whatever price you sell it at is up to you - that's your profit.

Cost seldom decides the price anyway. When there are many alternatives, usually the case for a small-time seller, the value of the product decides the price, and available alternatives decide its value.

Which brings us one full circle back to the original issue - value is hard to judge in audio.

So sellers price their products at what they value them to be.

That's why have to check if they value $50 things at $100. If they do, then other things they value at $100 may also be worth $50.

Amplifier output power vesus load impedance

2011-10-11T22:06:00.002+08:00

Amplifiers have their own rated output powers. Typically consumer products only show one set of numbers, for example, 30W @ 8 ohm.

This often leads to confusion when speakers of impedances other than 8 ohm are used. Statements like: "4 (or 8) ohms gives more power", and "2 (or 16) ohms will spoil the amplifier".

The correct answer would be: it depends.

Lets look at what an amplifier does. Assume an ideal amplifier. No limits of any sort. Output voltage is input voltage multiplied by gain. Connect a speaker, one end to the output and the other end to ground. Voltage between output and ground forces current to flow across the speaker, delivering power to the speaker.

But real devices have limits. A real amplifier has voltage and current limits.

Voltage: the maximum output voltage of an amplifier is equal to the power supply voltage minus a few volts, depending on the amplifier
If an amplifier attempts to output a voltage ( = input voltage + gain) that is bigger than the limit, it simply cannot.

Current: semiconductors can pass a particular amount of current until it reaches a point at which amount of current does not increase further.

The amplifier is capable of outputting any combination of voltage and current (as long as they are within limits).

Power equals to voltage times current, so to get maximum power from the amplifier, operate the amplifier at maximum current and maximum voltage.

However, current depends on voltage and impedance. (In speaker amplifier, voltage and load impedance are independent variables, their values set by the user.)

If the load impedance is low, current draw at a specific output voltage will be higher compared to with a higher impedance. Current limit will be reached earlier than voltage limit as output voltage/volume is increased.

If load impedance is high, then voltage limit is reached before the load draws a large current.

Too low or too high load impedance both result in max output power allowed with that particular load lower than the absolute max capable by that amplifier.

To get the most power out of the amplifier, choose a load such that current limit meets voltage limit. Or (impedance) = (voltage) / (current).

So in a sense, an amplifier is optimized for maximum output with a certain load. Note that this is often mistakenly called "impedance matching" which means another thing.

Usually the speaker amp manufacturer would optimize it for somewhere between 4 and 8 ohms. Because the output power drops proportionally with the change in load impedance outside of the optimal value, e.g. an amp optimized for 10W @ 8 ohms only pushes 5W @ 4 ohms, being fully limited by current, while an amp optimized for 10W @ 4 ohms only pushes 5W into 8 ohm, limited by voltage. If the optimal value is somewhere in between, then both 4 ohm and 8 ohm loads are off from the optimal value by not too much, and powers into 8 and 4 ohms will not be too much different. (An amp optimized for 10W @ 5.567 ohm would deliver 7W into both 4 ohm and 8 ohm)

And when power into 4 and 8 ohm are different, manufacturers show the higher one.

Understanding sarcasm

2011-10-10T11:27:00.001+08:00

"Understanding the subtlety of this usage requires second-order interpretation of the speaker's intentions. This sophisticated understanding can be lacking in some people with certain forms of brain damage, dementia and autism, (although not always)^[11] and this perception has been located by MRI in the right parahippocampal gyrus.^[12]^[13]"

Sarcasm - Wikipedia, the free encyclopedia

Oh, k......?

ASUS Xonar Essence One Review

2011-10-07T00:23:00.016+08:00

ASUS

ASUS, a brand that a few years ago people wouldn't have associated with audio. And then ASUS decided that Creative was mistreating computer users' ears with their X-Fi and launched the Xonar series of cards. So started the 'X-Fi for gaming, Xonar for music' saying.

I confess, at first I thought it was just some marketing nonsense and users' self-satisfaction. But after getting the Xonar DX, I realized that this series of cards is a very cheap way to enter audiophilia.

I haven't heard the STX, but if it is worth its price over the DX, it should be very good. It measures very well too. How well? Put it this way - most audiophile/"audiophile" DACs don't go near the STX in measurements. In fact most of those have problem matching DX.

But still, despite good sound and awesome measurements, when people have SGD$700 to spend, they don't want to buy a ~SGD$250 sound card, which some (or is it many) also believe to be inferior to external DACs for whatever reasons, like computers being noisy (despite SNR measurements proving otherwise). But at the same time, they don't know what to buy. Ask the internet and you get "satisfied users" recommending some new product from China that nobody really knows anything about and gets discontinued after a year or two. Below $300 there are many choices including big-time brand names, at about a thousand there are DACs meant for pro-audio, but between this two price points, products that one can trust are few and far in between.

So when the Xonar Essence One was announced, I was excited. STX already measures so well, how much better can they do with an external DAC. Being an external DAC, it also satisfies sound card haters. And it also provides a solid choice for people with $700 to spend, considering ASUS's previous products.

And when I was offered a unit for a trial, I was very excited.

Xonar Essence One

Normally I like to start with a measurement. But not this time. Because, this is a Xonar. My measuring card is a Xonar, the cheapest one. This Xonar Essence One is the most expensive Xonar. Poorer performer cannot be used to measure better performer.

Plus, like the STX, this guy comes with an Audio Precision Test Report. How many companies have the guts to release their test report. (Some don't even own an analyzer)

Other reviews of the Xonars mentioned the report, but none mentioned its importance and significance, and I'm going to talk about it.

First, the publishing of such measurements is not a common practice in audio or even most consumer products. A lot of the time the numbers on the product specs are taken directly from the datasheets of the components that the products use, or simply out of posteriors of marketing departments.

Even for pro-audio and component (as in just the single chip itself for example) manufacturers who need to publish real (or, more accurate) specs because their products' performances can be easily verified, detailed specs such as the results on the spectrum analyzer are optional. Because not saying is not lying.

Second, in order to accurately measure an equipment's performance, the measuring equipment has to be better than the device under test (I'll call it DUT here). For example, the DUT has a noise level of -90dB. The measuring equipment has a noise level of -80dB. The measuring equipment will measure -80dB even though the DUT performs better than that.

For poorer-performing audio gear, a cheap sound card is enough. But a high-performing equipment needs to be measured with an audio analyzer, an equipment that is designed to test other devices, so that the measuring device does not bottleneck the DUT's performance. That is what Audio Precision is, or to be precise, SYS-2722 from Audio Precision. It really measures your equipment.

So when a manufacturer dares to publish the product's AP test report, he is very confident of the product's measured performance. Which is a rare and exceptional thing in audiophilia, where poor-measuring products are a dime-a-dozen.

Measurements

Update (13/10/11) - After reading some measurements from here, my doubts were partly confirmed.

The measurements of the One from the link above are not good at all, but one particular observation that coincided with what I had, was the hum at 100Hz.

http://www.hdfever.fr/wp-content/uploads/2011/10/Niveau-de-bruit-Xonar-Asus-One.png

Essence One has visible 100Hz and its 2nd, 3rd, 4th and higher harmonics (200Hz, 300Hz, 400Hz etc.) at rather high amplitudes. Hence I was suspicious at the noise level graph of the AP test report below. However at the same time I couldn't really trust my Xonar DX, being a sound card in a computer I am not sure how the ground is generated and connected and how that has an effect on measurements, especially power-related ones like this 100Hz is.

Actually, the rest of the noise level measurement that I got was also similar to the one in the above link. Even if the output is hypothetically distortion-free, add that wierd-looking noise floor and you get a nasty-looking THD graph. I'm not sure if I measured additional 2kHz though (I think I did, but I'm not sure).

If one person has unexpected results, it is better to not jump to conclusions, so I treated my measurements as an anomaly and did not post them in case people do not read the fine print (which always happens). But if two people have similarly unexpected results, then the device-under-test and/or the commonalities in test methodology is more likely to be the cause.

But at the same time, ASUS's AP measurements for the STX are not that far off from what others have measured. So I do not believe ASUS is publishing fake results. If this is the case, then what could have caused such a big discrepancy? Would the balanced out yield the much better results? Well, I can't do anymore measurements for the unit is not with me anymore, so it's up to others to solve this mystery.

Now back to the original post...

I don't know if the sharp rise at the left is due to genuine hum or poor FFT resolution at lower frequencies. Then again, is the problem of reduced FFT resolution at lower frequencies supposed to show up on a linear scale? Why did they use a linear scale anyway?

But what escapes the eye can't escape from Audio Precision, which still gave it an excellent -120dB score. Interestingly, in this AP test report of STX below, the STX has a better score but seemingly worse lower frequencies. Weird.

http://files.computeraudiophile.com/2010/1217/e4253_Xonar_Essence_STX_Audio_Test_Report.pdf

Kinda strange that the external DAC measured worse than an internal sound card in noise. But I'm more interested in something that is usually at least 10dB worse than noise. THD.

THD, is, beautiful. -110dB maximum across all frequencies. How much is that? Yup, 0.0003%. It's written there. No that's not what I mean. How about lets take a look at a reviewer's measurement of the Benchmark DAC1 with the same SYS-2722. Yup, that Benchmark DAC1. SGD$1600.

So... the Essence One beats that Benchmark DAC1 in measured performance. That Benchmark DAC1 that is well-known for measured performance at SGD$1600.

But, there is another product that also beats the DAC1. And it is even cheaper than the Essence One. And it is also from ASUS. Yea, it's the STX, you probably guessed that by the previous sentence.

The STX measures just sliiiiightly better. But here is where I noticed a problem.

Audio Precision SYS-2722's rated performance is 0.0003% THD. So both products are already at the limit of the measuring equipment. The performance of high-precision measuring equipment can change or degrade with time, I don't know if that is the case here. The SYS-2722 is still within specs though, but measurements below the rated limit needs to be consumed with vinegar. ASUS measured the Essence One years after they measured the STX, maybe the SYS-2722 has degraded by a few dBs, or maybe simply a different SYS-2722 was used?

This also explains the slightly worse noise performance of Essence One vs STX. The SYS-2722 is rated at -120dB.

Anyway, just for hitting the limit of SYS-2722, both products are wow. But this way I can't compare the difference between the two. Or can I...?

I'm going to refer to the STX's AP report above again. The STX has second and third harmonics at ~-112dB, while for the One...

It has second harmonic at -120dB, third harmonic at even lower, and higher harmonics at non-existent.

Measured noise for for STX is 4dB better than One, but measured THD+Noise @ 1kHz is the same, therefore the THD portion of the One would be better. Lower distortions is more important, since they are always audible above the noise floor.

BTW -120dB is the limit of human hearing. In the sense that the smallest sound that can be heard is taken to be 2×10⁻⁵ Pa, or 0dB. So if you want to hear something that is -120dB as loud as the main signal, the main signal has to be 120dB, which is the safety limit for hearing damage.

Crosstalk, just to complete the set of measurements.

For an Audio Precision measurement, this is missing a lot of things that the AP is capable of measuring - square wave, linearity, or at least give us IMD which even RMAA has? Oh well, already takes a lot of guts to provide the THD measurement, can't really complain.

Verdict on measurements: If those figures are true, ASUS has created a monster. If anybody knows anything that performs better at the Essence One's price range, or heck at any price range, please sound out, because I am really interested to know.

Features and functionality

On quick glance this thing doesn't look much different from most DACs, apart from the XLR out which isn't very rare either.

But it was when I started using it I repeatedly got impressed.

The built-in USB - unlike most other USB DACs that are limited to 96kHz, the CM6631 used in the Essence One is a USB 2.0 part that supports 192kHz. Needs to install driver though, but which good hardware doesn't?

The lights beside the line-out volume knob displays which sampling rate the DAC is receiving. 44.1, 48, 88.2, 96, 176.4, 192, and... bit perfect...? Time to check the manual.

Turns out, the bit perfect lights up when the USB is used via ASIO drivers. Wait... this USB has ASIO...? Lets check in foobar2000. Yupz... it has ASIO. Would make more sense to label it ASIO, but I guess the target market is more engrossed over bit perfect than ASIO.

Two independent digital volume controls. Extravagant like a king. Great for people who use both speakers and headphones. Quick testing shows that the volume controls don't induce measurable distortion, at least when measured with my Xonar DX.

When the unit is turned on, the selected source is always USB. I guess most buyers of ASUS products are computer users, and most of them use USB.

Speaking of USB, in this review by Sound News, the USB was not working well. I used the latest drivers downloaded from ASUS's website, sometimes there is noise, most of the time not. In fact I think whether the noise is there or not is decided when the unit is turned on, and once it has the noise (or not), it stays that way until the unit is turned off. I can't pinpoint the cause, because out of the many times I cycled the unit's power, only 2 or 3 times the noise was there. I leave my equipment on 24/7, so this is not an issue once I manage to make it start with no noise.

The output impedance of the headphone amplifier isn't what I consider to be negligible. Quick testing shows a drop in output levels of about ~1dB (varies with volume control position for some reason) with a 33 ohm load, which gives an output impedance of around 4 ohm @ 1kHz. I did not check if it is frequency-dependent, and it was a very rough check, a single-digit ohm should be a fair guess. Someone else would have to verify this, but until then I'm not too impressed with the headphone amp.

Additional notes on output impedance
Some equipment have a resister, for example 47 ohm, in series with the output. This is done either due to performance issues and/or to control the power into the headphones when switching between headphones of different impedances. Also the 1996 IEC 61938 standard specifies a 120 ohm output impedance for headphone amplifiers, though this standard hasn't been care about much lately.

Hence there are differing views on whether a certain value of output impedance is bad. Some believe such a value is good enough, citing equipment with much higher values, others claim to require as close to zero as possible. (I belong to the latter.) But I'll leave the number there and let you decide. That's why numbers are good - they don't have their own opinion.

Now back to the post...

There is an upsampling button which I don't really know does what apart from upsampling (but what kind of upsampling?), but it introduces a rather steep low-pass filter which also starts rather early. I don't have the graph or exact numbers, but the difference *might* be noticeable.

Trivia: At 14-15W power consumption, the sealed top of the unit gets fairly warm.

Subjective listening

I simply can't stress this enough - the human ear-brain hearing system is unreliable and can hear things there are not there. Blind testing is essential to eliminate placebo, and multiple trials are needed to ensure what is heard or "heard" did not occur by chance.

In this review, multiple-trial blind testing was not carried out. So it is possible that my observations are incorrect. This should be kept in mind when reading such reviews.

(The shootout that did not happen)

If I were to pick something to compare the Essence One against, it should be something that many people are familiar with. And something that is good of course.

So I picked Xonar DX. Which is also my choice of source among the pile up there (minus the Essence One which I do not own).

First impression of the Essence One, is that it has a sound signature that is largely similar to DX. Could be because they are both Xonars, could be because of placebo effect because I know that they are both Xonars, could be because I have gotten more conservative instead of making wild claims. But it was not long before I appreciated the spacious and less-congested treble, which is what I have always felt the DX lacks. The bass goes deep, but controlled, I could feel the initial shockwave from when the beater hits the membrane. Bass guitars were definitely enjoying the bass performance, being reproduced with greater accuracy. One would be able to hear more clearly the settings of the preamp.

Usually when I compare two pieces of audio gear, one of them would sound different from the other, but which one sounds better would be up to personal preference. Not this time. The Essence One sounds like a DX with its bass and treble weaknesses removed, an all-rounded improvement which I could find nothing to complain of, and upgraders from other Xonar products need not worry.

That was comparison against Xonar DX. As for how Xonar DX compares against others in general, the DX has a less pronounced though much cleaner extreme highs (compared against others with more distortions at high frequencies I guess), a powerful bass that may be uncontrolled at times, and an overall sound that feels like a recording studio, not as fun and lively as the rest, boring but somehow "sounds right" (I don't dare to use the word "accurate"). The Essence One keeps the humble and clean treble, and adds the life back, with improved bass, while still sounding like a Xonar.

Verdict

The price of $699 may seem steep compared to the STX, but the chassis, power supply and control circuitry of an external product easily costs a good fraction of a complete product. Add the SPDIF inputs and digital volume controls, and the Essence One isn't really that expensive compared to the STX, and for the extra money you get a pair of expensive PCM1795 and some form of upsampling, courtesy of the DSP. And lets not forget the Essence One only seems expensive compared to the STX because the STX is super-cheap compared to everything else.

It is hard to give a verdict on something that does not have an alternative for its price. It is very easy to give a verdict when its alternatives are a few times more expensive. The measurements already justify the price, and if you are a believer of measurements and/or ASUS, wants an external DAC and has SGD$700, what are you waiting for. Do consider the STX if you are cash-strapped, but in audio it costs a lot more just for that extra bit of performance, and especially if that performance is the best.

Class 4 SD

2011-10-02T10:01:00.001+08:00

The Kingston is class-less, while the Sandisk is class 4.

A reminder that not having a rating or certification does not necessarily mean it is bad. OEM PSUs with >>80% efficiency and without 80Plus certification, speakers without THX certification.

I am so going to buy a car

2011-09-29T21:44:00.003+08:00

[Flipside] The male protagonist's situation is troublesome

2011-09-18T07:57:00.001+08:00

Many people write stories or play games to transform into the main character, someone that they wish they were.

Being chased after by a bunch of pretty girls.

And they (the male protagonist) always think it is troublesome.

Well, what if we apply reality check.

The typical real woman is not pretty.

The typical highschool girl may or may not be pretty, depending on which group you hang out with.

So what if you are being chased around by a bunch of ugly ones? This is more likely to happen.

Troublesome indeed.

Gaming ruins lives

2011-09-16T21:32:00.000+08:00

I just had two full days worth of productivity wasted because I got a GTX 560. Two full days can mean a lot of progress.

But in GTX 560's defense, I completed Crysis 2.

Quote of the day: Audio snakeoils are like ghosts - they only come out if you believe in them, and they never appear when you want to show them to other people.

2011-09-08T22:14:00.001+08:00

Compaq Presario Casing Review

2011-09-02T22:21:00.006+08:00

Because reviews don't have to be always of the newest things that you can buy.

This is the casing I'm going to review today. It is one of the last if not the last Compaq casing to use the old Compaq logo. (The new logo is a single big Q, which I don't like.)

A simplistic design that is smooth to the eye but still square enough to remind you that this is a computer and not some weird spaceship things the enthusiasts like to buy.

I forgot to take a picture of the LED, so here is one taken from here. It is not that blinding to the eye in real life.

Smooth reflective black drive bay covers with 2 stealth drive bays for optical disc drives.

Most of the case is in a very nice matte black which I find rather fingerprint and scratch resistant (as in, fingerprints and scratches on this surface are not so apparent).

The front bezel is wider than the rest of the case and props the case up slightly, breaking the boring horizontal lines of PC casings. It makes it harder to put things on top though.

I like the vent holes.

The rear of the casing. Here you can see three things typical on a PC from a large vendor, are useful, but usually missing from DIY casings. The place for you to grab to pull the side panel off, a singe thumbscrew to secure the side panel, and a place for a lock.

The holes for a 92mm exhaust fan are pretty unrestrictive and is raised slightly - this allows for less restricted airflow.

The piece of metal that is supposed to be screwed onto the hole and hold the expansion cards is missing. I have mixed feelings towards this (the plate, not its missing). On one hand it hides the ugly screws of the expansion cards and makes it slightly easier to install expansion cards, on the other hand the grip strength is not strong enough to hold the graphics card sitting in the top expansion slot in place.

The panels are pretty study. This is me holding it up on one end and it is not flexing much. I never encountered flexing when installing the side panel.

The inside.

This case came with a 300W power supply by HIPRO. I remember another case came with something like 245W or 265W, so it depends on the system configuration. Good enough to handle PCI-E graphics cards without the extra power connector.

The drive mounting mechanism of this casing is what is great. At first it looks like those mounting rails-based stuff, but when you look closely there is no space for a mounting rail between the HDD and the chassis.

To install the HDD, these are the steps:

1) Install these screws onto the drive.
2) Remove the front bezel. Fortunately, it is very easily removable

3) Slide the HDD in from the front.

4) Push it all the way in until you hear a click. Those 4 x 2 rows of screws are spares I think.

To remove the HDD, in the below picture, find the screw that is among the green, erm, thingie. Lift that liftable-looking-thingy away from the HDD and push the HDD out.

The installation is similar for 5.25-inch drives.

I find that certain generic screws work also as long as they are round and smooth enough. So you will never run into the problem of missing drive rails resulting in inability to mount drives, which is always the case for casings using drive rails.

Verdict

I can't really give a verdict for things that are cannot be bought new anymore (not that it could be bought new in the first place), but this is a great study and nice-looking case. I have seen casings of this model being sold second-hand for $30 or less, and compared to $30 cases in SLS, this looks way better and is way better built too. In fact, quite a few of the modern branded PCs' cases are nice-looking and well-built, so don't let a good chance pass you by, if you can look past the brand logo.

CaseFail

2011-08-29T09:05:00.003+08:00

[Flipside Week] The internet as a source of truth

2011-08-28T16:00:00.001+08:00

Did you enjoy the week of alternative opinions? Some of them are logical fallacies, many are incomplete truth such that if you know the entire picture, the current practices will be explained. It is all troll logic, so don't message me about your intellectual superiority.

But it gets you thinking. Why are some things done some way. It need not always follow logic, or more specifically, layman's logic. Common sense differs between people with different amounts of knowledge. What is logical to a child may not be logical to a scientist.

I decided to finish up the week with a big topic that affects all of us: the internet.

The internet is always considered to be full of lies. Because of websites like this, this, this, this, this, this, and this.

However, some of the websites above have some truth in them, you just have to find hard enough.

(BTW, 4chan is always right, and rule 1.)

Why do lies exist on the internet in the place? The content of the internet is contributed by its netizens. That means it is the humans that posted the lies on the internet.

Because it is human nature to speak untruth, be it intentional or non-intentional. The real world restricts the speaking of lies via a threat called accountability. People will know you are talking bulls***, and take action against you. If given the chance, people will spout nonsense if they know they cannot be stopped. Like employees in the service sector under the protection and/or order of their bosses, and state media. I know all diploma holders earn over $7k per month, because that is what all those acquaintances of mine told me.

So where do you go when you need information to double-check a claim against?

The internet.

It is full of words. Words that can be true or false. Words that are typed by others. But compared against words from just one person whose intention you do not know (usually bad, hence the need to check), the words by multiple unrelated people who do not know you have a higher chance to contain truth.

Furthermore, most of the time we are just concerned with finding that single one truth. Or something closest to the truth.

The relationship between word quantity, truthfulness, and anonymity of the poster is quite similar to the Planck curves, the y-axis being quantity, x-axis being truthfulness, and increasing temperature representing increasing anonymity.

Image source

As you can see, as anonymity increases, word quantity increases while average truthfulness decreases. However, the absolute quantity of the number of the most truthful words also increases. And this is what we are most concerned about.

For example, when a salesman recommends you a product, he has no anonymity, but he sure is talking a lot of bull. You go onto the internet to find a bunch of Amazon/eBay user reviews which are even more bull. But then finally you find one opinion by one reviewer, who had third-party measurements he did himself to back his claims, you can be more sure of the performance of the product.

The internet may be the only way for mortals to meet saints, but there are many times more people who pretend they are saints. This is why one must be able to filter what he reads.

And when you know where and how to look for the truth, you can learn a lot of real, useful things from the internet. More than the equivalent of what you learn from a degree if you work hard enough. (Actually I don't even know why I am paying my university except for the certificate, I'm doing most of my learning by myself, can't learn anything from the lectures.) Sometimes I search the internet to solve the tutorial questions, using notes and tutorials from other universities. Yes you can get notes and tutorials of universities by Googling.

So, appreciate the internet, for it is a good source of truth. It is also a good source of useful and useless information, but that is what we like about it.

[Flipside Week] A government that keeps quarreling and gets nothing done is bad?

2011-08-27T19:00:00.009+08:00

An argument against too much opposition in the government is that people will quarrel the whole day and get nothing done.

Taiwan and USA are cited as examples.

But, is getting nothing done bad?

In today's age of heavy privatization, the government has relied on the natural forces of economics to handle decisions, citing the higher efficiency.

If forces of economics are good enough, there wouldn't be a need for goods and services that lose money. And people won't do things that are bad for them.

The government's role is to provide these things that are needed by the people, and make unpopular decisions to force people to do things at their expenses but benefit the entire society in general.

Keyword being unpopular. And efficiency of capitalism.

If government only makes unpopular decisions, and forces of economics are efficient enough to run the world on its own, why bother having the government make any decision?

In the first place, who evaluates that the government's decisions are good or just trying to make a quick buck? How would we know if the success of a country is due to its "leaders" or the exploited people?

[Flipside Week] People who get sexposed the most are the ones who get elected

2011-08-26T19:00:00.026+08:00

***I realize this post is scheduled to be posted on the evening of the cooling-off day. And due to the nature of some of the candidates involved in the upcoming Presidential Election, this post will be touching on some sensitive topic. As such, it will be postponed until the results are revealed on the night of the polling day or even scrapped depending on the results.

Please note that this themed week's postings are meant to view the world with a twisted viewpoint for entertainment purposes, and not trying to propagate any terroristic or satanic mindset. If you are going to be offended by the content, please leave this site now. By reading this post you agree to not hold wwenze, Blogger, Google, my mom, your mom, and my god directly in control of me responsible for any psychological trauma or brain damage or sudden disappearance of yourself and your family that may result from reading this post.***

During election, people have to wayang. Put on fake smiles and say they aim to do good for the people.

Some people do it worse than others. They try to wayang, but people uncover what they are really like. Kenna sexposed.

But, that does not mean they won't be elected, more like the other way round.

Because they know precisely they will be elected, that's why they don't care about getting sexposed. They can do any s***, say any s***, and still get elected. They are that confident.

Therefore, those who get sexposed are the ones who will be elected.

Just don't say something like "if you don't vote for me, you will be sorry ".

[Flipside Week] Strong national currency is bad? Yea rite.

2011-08-25T19:00:00.005+08:00

A strong currency is said to hurt the nation's competitiveness. The reason given is that the exports are now more expensive to other countries.

So there will be less demand for the nation's exports. So less jobs for the people.

Is that a bad thing?

Yea rite.

Why does a country export stuff in the first place.

Because other countries want them.

There are two cases why other countries would want a nation's exports -

1) The export is unique to that nation, or
2) The nation sells it cheaper

In the first case, we have a monopoly scenario. So price is still set by the country (or the company in the country). If the country wants, it would be easy to reduce the price of the export. IF the country wants.

In the second case it is closer to a fully-competitive market scenario, where everybody has the ability to produce the same thing, so the guy with the lowest cost sets the price to become the lowest and wins ...until the cost of production to produce any more becomes too high and matches market price.

And a country's currency strengthening is a sign of it happening. Too much demand, price of currency goes up.

Which is actually not a bad thing either - since economics is supposedly the natural way of balancing things, so the nation is not getting less competitive; it was actually getting underpaid previously. And getting underpaid is bad.

And, if you want the value of your currency to drop, just import more to make the trade surplus into a trade deficit. U.S.A. style. Spend more, enjoy yourself. Even if you do not get paid in cash, I don't mind being paid in Core i7's.

Don't forget, if a country's currency is strong, the imports become cheaper. So spam imports, especially if this nation imports everything.

So, everybody buys more stuff and lives happily ever after.

Wait... you're telling me this is not the case?

Because the prices and hence spending power locally remains the same? Despite imports becoming cheaper? And despite less manhours being available for work due to less demand due to it becoming more expensive?

Well that is the economic problem the top tells us about. But you know what, screw it. It happens only because the people up there are hoarding all the money.

And since the money does not go to you anyway, they prefer higher production. More exports. You ain't gonna get the benefit either way.

So, a strong national currency is bad. For those at the top.

As with every financial bad news. They affect the super-duper-rich more than the common people. Same goes for the financial good news too. Frankly, the past few years with the crash and recovery, I never felt anything. One thing did change within the few years - inflation. And that happened during the recovery. Inflation happens when there is more money.

So the policies are often meant for the super-duper-rich only too. So now you know why the big shots seem to always do things unfavorable to the majority. Because it is good for them. And because they can.

So, ignore those BS, all the news and idealistic theories get unrecognizably corrupted by the time they reach the ground. The reasons given for the incessant price increases are a good example - there are no reasons.

One thing is sure - online shopping is cheaper.

And that is certainly not a bad thing.

[Flipside Week] Are you really being smart?

2011-08-24T19:00:00.005+08:00

People in some jobs, some majors, we consider them smart.

But, are they really being smart?

Working hard for long hours for minimum pay.

People work for money, and if money is the goal, shouldn't a smart person aim to obtain more money for less work?

Think businessmen and management.

If you ask which are more beneficial to the society, no doubt those working hard for long hours are. But then again, if economics dictates the value of a person, then the person getting paid a lot is worth that much.

So, business majors are the smartest?

[Flipside Week] Another view of what the Earth thinks about us

2011-08-23T19:00:00.003+08:00

The majority of planets are barren or full of violent reactions.

And the majority of stars, are, on fire, but lets not talk about that.

A planet is a typically unreactive lump of substances, as opposed to the reactive lumps which we call stars.

So that is how a planet is supposed to be.

Lets look at the earth.

On earth, there is this thing that we define as living organism.

It reproduces and spreads itself to occupy larger and larger areas of of planet surface, and even releases toxins, such as oxygen, to transform the Earth into a better condition for this parasite.

Until millions of years ago when the planet got totally covered in green.

And green is a color of toxic waste.

And the decomposed black, gooey, liquid waste gets stuck underneath the surface with no way of extracting it or eliminating it (until humans came), while deposits of concentrated calcium literally scatter all over the place and scar the otherwise perfect rock strata.

Who knows, if the Earth was alive, maybe its idea of majestic beauty is the Grand Canyon, Mount Everest, Mariana Trench, the Great Red Spot of Jupiter, or the Ice Cap of Mars. And not the greenery we call 'Mother Nature'.

So, all that talk about saving the planet?

It's more like keeping the planet a favorable condition for continual existence of living organisms.

It is said that the best pathogen does not kill the host, but keeps it alive so more pathogens can be produced and spread.

Soon, the Moon, and not long afterwards, Mars, will be infected.

Always wash your hands and keep a distance from patients.

[Flipside Week] School exams grades are not useless for measuring the capabilities of employees

2011-08-22T19:00:00.064+08:00

Many people often say that doing well in schools does not always mean doing well at real work.

Well that makes sense, because the school teaches unrelated things and measures unrelated things. It teaches you how to study for exams, and measures how well you take the exams.

But, it is the most controlled way of measuring how good the employee is.

Because I can ask the same with other things - is the performance review of an employee a good indicator of how he will perform in the future, or even how he is performing now? Does the review show how good he is or how good he appears to be?

Consider two people on two projects. Project one fails while project two is a success. How to know if project one failed despite guy 1's great work because it is too hard, and project 2 is a piece of cake and guy 2 is slacking?

Trying to compare things with too many different variables at the same time yields no useful information.

This is why schools and exams are good. You have a whole bunch of students doing the same exam. The marks reflect how well they demonstrate their intellect, not how well they con people. And since everyone is doing the same exam, results are comparable against each other.

If you say that school exams are simple compared to the complex real world, then all the more reason for the good people to ace it, and it says something about those that fail.

And if you say that school exams and university degrees are about things unrelated to real work, I agree. But then the world is always hiring people whose knowledge is unrelated to the companies they are working for. We call them MBAs and consultants.

[Flipside Week] Being retrenched is good

2011-08-21T19:24:00.028+08:00

Welcome to Flipside Week!

Flipside week is another of my themed weeks. This time, I will be looking at things from a different angle and giving interpretations that are opposite of the norm, yet still makes logical sense. This begs the question - by how much have our beliefs been shaped by the common view of others, instead of logic.

So here is the first one -

Every recession, people become scared of getting retrenched.

But I always tell myself and others, only the overpaid get retrenched; those worth every single cent of their salary stay.

Which (being worth every single cent of your salary), I recently realized, is actually a problem.

That means you are doing more and getting paid less compared to those that get retrenched.

And wouldn't everyone like to be paid more than what we are actually worth? I mean, look at those people on top, running companies and countries.

So, if you get retrenched, that is a sign that you were being overpaid. And that is good.

Consider this - you may be worth only 2k per month, but you were earning 3k, that is 1k per month and many many k's over years, money that you shouldn't have gotten.

Getting retrenched is just making you draw the salary you should be drawing, in another job.

So, think about the amount of money you have stolen over the years, and laugh.

Truth hurts

2011-08-21T08:38:00.002+08:00

That's why scams based on the target's beliefs work.

[Audio] Why good engineering never wins

2011-08-17T21:47:00.000+08:00

Engineering is not just about building something - a monkey can do that

Building something that performs well.

And building the best while working within limits.

And cost is always one of the limits.

If a product of a specific performance is required, a well-designed product would be cheaper for the same performance.

Cheap parts can be used to achieve same or better performance than a design using expensive parts.

This translate to lower costs. And if the market is perfectly competitive, lower prices. Consumers are happy. Designers incapable of reducing the cost die off.

This is practiced almost everywhere.

But, when it comes to audiophile equipment, things can be different.

An amplifier with NE5532 and electrolytic caps for output, versus an amplifier with (insert discrete op-amp here) and PIO caps for output, which one would you think performs better?

Majority of fanatics would choose the latter.

Even though it may not be. Remember that engineering is about building the best for the least.

Or heck, say the two amps perform identically, but one uses cheaper parts through good design, the other uses a design from the 80's. The cheaper parts one is priced cheaper while the more expensive one is priced more expensive.

Which would people think performs better?

More expensive, more expensive parts, no doubt.

Even if the well-designed one is priced the same, people will look at it, and go for the more expensive-looking one. And then claim it is value-for-money because of the more expensive parts for the same price.

So a well-designed product is being penalized for being cost-effective.

And that is why people pay so much for wood.

Taiyo Yuden YUDEN000 T02 DVD+R: Surviving fine after five years

2011-08-14T07:05:00.002+08:00

So my PC's soundcard is less noisy when the CPU is under load? - An investigation on computer activity and audio noise

2011-08-12T22:19:00.010+08:00

A while back I discovered that my laptop's onboard audio's RMAA-measured performance was better when the CPU was under load. However I could not reproduce it on my desktop (it is still completely reproducible on the laptop) so I took it as a one-time affair and let it rest.

I was reminded of this topic again, with all those low-power CPUs, minimalistic OS builds, and SSDs, as well as linear power supplies and power filters for HDDs and fans running around. But this time, I can measure down to -120dB.

Test setup

Hardware
Realtek ALC888 (for playback)
Gigabyte GA-MA770-US3 (rev. 1.0)
AMD Phenom II X3 720 @ 800Mhz/0.75V idle 2.8GHz/1.20V load unless otherwise stated
Asus HD5670
Antec TruePower New TP-550
Western Digital Caviar Green 1TB
Xonar DX (for measurements)

Software
OCCT (to load the CPU)
Arta (for measurements)
Arta settings:
24-bit, 96kHz, Spectrum Analyzer, 1kHz Sine, FFT 65536, Hanning window, linearly averaged 20 times for each measurement

I used the onboard Realtek because I observed previously that onboard sounds exhibit more fluctuations in measured results and hence are more susceptible to fluctuations in noise levels.
The motherboard decides performance more than the codec, just as the rest of the soundcard/DAC decides performance more than the DAC chip used, you can find evidences for both on the internet.
The CPU is a beast that has been undervolted into a puppy. But when overvolted and running Linpack it draws over 100W.
Xonar DX in the same computer handles the measurements. Loopback self-tests show that it has practically no difference between when idling and when overvolted and OCCT power supply tested (more on this term later). So it does not inject additional noises as the computer is loaded and hence already busted part of the myth, but we are looking at the Realtek here.

Test 1 - CPU idle vs load

This is how the Realtek does when the CPU is idling.

And this is how it does when loaded by OCCT.

Two interesting things to note - while there are humps at around the 2kHz and 4kHz regions when under load, noise performance below 1kHz is better.

Linpack is a test that focuses on generating heat and wasting power. The hump to the right of the 1kHz test signal covers the previous two seen in the OCCT test. However the spikes to the left are still less and smaller compared to when idling.

I always remeasure my baseline to make sure it is accurate. And I am glad I practice that. This "idle again" looks different from and better than the first idling, and seems to combine the good parts of idle and load. However, closer inspection reveals a higher noise floor below 1kHz which seems to correspond to the smallest peaks (~7dB above noise floor) of the idling picture. Could be a measurement issue.

Test 2 - Power supply + CPU + graphics card a.k.a. full system loading

OCCT power supply test tests the computer's power supply by loading the CPU and graphics card at the same time for maximum power consumption. As expected, noise levels are the worst so far. But, it gets even worse later.

Test 3 - CPU voltage

CPU voltage is increased to 1.55V with everything else the same. Noise floor rose compared to 1.20V.

Test 4 - Idling at different voltages and frequencies

Noise increases just by increasing voltage or frequency alone. Idling at high frequency and high voltage results in the worst performance so far.

Test 5 - H.264 video playback

Overall it looks worse than OCCT Power Supply test. Even though H.264 video playback is much less demanding. Another case of less work, more noise.

Test 6 - Fans

It gets a little noisier with the two fans at 1150rpm and 2000rpm respectively. But at those speeds the mechanical noise will drown out any electrical noise. Most people going for fanless do it to attain true silence, but having a few fans at 600rpm is a much cheaper way of doing things, and it does not seem to have much effect electrical noise-wise.

Test 7 - Hard disk drive

It may look like there is a difference on first glance, but the difference is less than the fluctuations of idle performance. So any difference can be treated as insignificant.

Collection of idle measurements

These are all the measurements done when the system is idling at 800MHz 0.75V, comparing them gives a good idea of natural random fluctuation of performance.

Idling
Idle again
Idling 3
Idling 800MHz 0.75V
Fans off (The fans are near stalling speed in all other tests)
HDD not defragging

Interpreting the measurements

I only measured up to 20kHz because beyond that the noise levels recorded by Xonar DX according to Arta is impossibly low (below -146dB which is the theoretical limit for 24-bit). (RMAA with 96kHz sampling rate for up to 48kHz frequency shows no problem however. Could be a problem with the settings.) It is possible for ultrasonics that escaped detection to cause problem with some equipment, but in that case it would show up in the audible range.

Two patterns were spotted - increasing frequency and voltage (or power draw in general) increases noise. Loading the CPU with OCCT reduced noise compared to without, both at 800MHz 0.75V and 2.8GHz 1.55V.

It can hence be inferred that the best scenario with the lowest noise would be OCCT at 800MHz 0.75V. So I tested it, and

viola, it easily has the lowest noise. The other extreme would be idling at 2.8GHz 1.55V which has been covered above. In all the other cases, including OCCT Power Supply test, the loading of the CPU countered the rise in power draw, so the noise is somewhere in between.

In the case of my laptop, the voltage is the same at both idling and full load clocks. So there was no increase in noise due to increase in voltage. Hence the noise was lower overall.

The worst-case scenario of idle 2.8GHz 1.55V is 10-15dB worse than the best case scenario of load 800MHz 0.75V. For those unacquainted with decibels, download foobar2000 and adjust the volume control to -15dB to see (or hear) the difference.

Now, although the difference is big, in the worst case scenario the highest spike (ignore those over 8kHz) is still lower than -80dB. By right, this would not be audible. If your soundcard is outputting at maximum output level. But, if you are having it output at -20dB or even -40dB because you are using headphones, or using your soundcard as volume control and/or your system has too much gain, or for some reason this noise reaches other equipment and gets amplified, it would be audible when the system is not playing anything, and 15dB would mean the difference between audible and not.
Basically that is to say if you are not suffering from noise problems you won't suffer from it, but if you are suffering from noise problems you can try some tricks.

Also, these variations did not show up on the Xonar DX in its own self tests, hence if your soundcard has a good power supply on itself with high rejection of noise from the rest of the PC, there is nothing to care about.

Conclusion

I started out to confirm if measured noise performance of onboard sound is indeed better with CPU under load, I confirmed that to be the case.

Repairing a Yamaha keyboard

2011-08-11T09:30:00.001+08:00

So my family's keyboard had a few keys that didn't make sound when pressed. It is not a good keyboard in the first place (though it did accompany me or mostly my bro to various performances and competitions) and very old (>10 years), so I didn't want to repair it. But my mom was asking me to.

So on a free day, I took it apart to try my luck.

All the screws that are visible from the bottom need to be removed. Including a few long ones nearer to the center hidden in rectangular holes. The battery compartment cover also needed to be removed for some reason.

For you electronics porn lovers out there. That includes me.

The speaker. Doesn't look half bad.

The amp chip - an LA4705

The board.

What caused the malfunction was the large grey ribbon cable. It was probably rusted and broken inside. Because repositioning it solved the problem.

Time spent: 15 minutes

Whistling away at 600rpm...

2011-08-08T12:52:00.001+08:00

Back to being silentpcreview.com-approved.

I had to do some things to the HD5670 to achieve this though:

HD5670, 92mm fan edition.

And then undervolt the HD5670 to 1.0V @ 700MHz. The HD4770 could do 0.9V @ 675MHz.

Quote of the day:

2011-08-05T07:10:00.002+08:00

"In science, contrary evidence causes one to question a theory. In religion, contrary evidence causes one to question the evidence." --Floyd Toole

NZXT Havik 140 CPU cooler - User review - Special pseudo-fanless edition

2011-07-27T17:22:00.026+08:00

I never thought I would get my hands on an NZXT product. NZXT is famous for its chassis and case fans, typically aimed at the silent computing enthusiasts. I also am a silent computing enthusiast, but I never spend on expensive casings and fans. I have never spent on CPU heatsinks either since I got the Ninja in early 2006.

And it happened that I won a contest (or is it more of a lucky draw?) and got myself a NZXT Havik 140 CPU cooler. And being a number-hunter, I promptly proceeded to test it.

But, there are already multiple more-professionally-done reviews available on the web. So it goes against my principle to make another half-assed production of the same thing.

So, I'm making something else.

I'm going to get numbers. But the problem with numbers is, they don't mean anything if you don't understand what they mean.

...Sorry, that one applies more to audio, in most other fields people don't flak each other with illogical theories and meaningless numbers.

The problem with numbers is, they don't necessarily apply in other scenarios.

And my scenario is different.

I run my tower heatsink pseudo-fanless, with the rear exhaust casing fan acting like the CPU fan through a duct. The casing fan is plugged to the CPU fan header. Well, for a while I didn't use a duct, because the Ninja is so close to the exhaust and PSU. (But a duct still significantly improved temperatures, cool.)

Because the most effective way of making things quiet is to make less sound in the first place.

And to top it off, the exhaust has a top speed of 1100rpm and runs at 700-800rpm with motherboard fan control.

So the heatsink has to work with low airflow. And this is usually not tested in reviews. Even silentpcreview.com where people expect to find the quietest reviews does not usually test fanless or ducted conditions. This is understandable, because the fan is part of the whole heatsink product, the heatsink is usually not designed to operate without a fan (or with low airflow), and without the fan temperatures shoot skywards in the best of coolers and the CPU melts in the case of low-end coolers.

Before we start the testing, lets go for some customary photos.

The box, beside a WRT54G, for size comparison

The box

Still the box

The side of the box. At least the English has no grammar errors.

The two fans you first see when opening the box

Box contents

The heatsink itself

A 120mm fan to show the size. In Chinese, "say big not big, say small not small."

A splitter to control two fans using one CPU fan header. A nice touch.

The bracket installed. From here onwards no manual is needed, but up till this stage it was very confusing. The backplate had to be flipped the right way up depending on Intel or AMD, 4 things needed to be removed from the backplate and 4 insulating stickers pasted, 4 thumbscrews and 8 small normal screws screwed in pointing upwards with the Intel bracket below the AMD bracket. "Easy installation", yeah right.

The bracket is there for the cooler to be mounted facing two possible ways 90 degrees from each other. But at this point I noticed a potential problem - the length of the screws protruding from the metal plates are the same, resulting in different heights when two of the plates are mounted below the other two. This will result in different mounting pressures depending on which way the cooler is facing, or worse, which will be covered later.

Base is smooth to the touch, but won't be winning any prize visually.

Two ways it can be mounted,

like this

...or like this. The Ninja can be seen at the side. It is a novice, that is why it gets seen.

Comparison with the Ninja:

And now is the problem I mentioned earlier. In order to have the wide side facing the rear fan on most AMD motherboards, you have to use the screw with the lower height (if installed according to the instructions manual), and it is too difficult to press down the spring-loaded screws hard enough for them to reach. It may be possible with tools or just adequate grip, but you have to do this with the CPU cooler on the motherboard, which adds many levels to the difficulty.

A trick I used, instead of mounting the metal pieces I'm talking about below the other metal pieces, I mounted it above. I think the picture below shows better what I am talking about.

Compare it with this picture from above and spot the difference.

There are issues with this approach though. The heatsink pulls upwards, and while the original design have the metal pieces stopping each other, here only the screw threads of 8 small screws are preventing the cooler from coming off. Second, the mounting pressure may be reduced, but only if the screw cannot screw down to the same depth. So far, there has been no problem.

On retrospect, another way is to loosen the 4 thumbnuts connecting to the screw for the backplate, enough for the spring-loaded screws on the heatsink to reach, then tighten the thumbnuts, then tighten the screws.

I could reuse some of the pieces of the Ninja's duct.

Another view

A photo to show the proximity of the heatsink to the PSU. Adding a wall here to force air through the dense fins may cause too much resistance and make the PSU noisy, so I left it like that.

And now, the important part - the numbers

Unfortunately there was some problem with measuring room and ambient (in chassis) temperature. However, just with idle and load temperatures the performance of the cooler can be guessed, by temperature difference due to power difference.

This is the system power when idle (800MHz @ 0.75V)

And this is Linpack @ 2.8GHz when overvolted to 1.55V.

216 - 74 = 142. The efficiency of the 80 Plus Bronze PSU at 216W input power should be around 85%, considering it is a lower-power version of the 750W version and I am on 230V. This gives 184W of DC power.

At 74W, efficiency is maybe 78% (guesstimate), giving us 58W of DC power.

And 184 - 58 = 126. Well, there are VRM losses too and other parts like northbridge and RAM, but 100W should be a fair estimate. This is the CPU's power difference between idle and Linpack.

Notice I keep saying Linpack instead of load. Because Linpack makes your CPU hotter than most other loads including Prime95 and OCCT (non-Linpack test; I use the Linpack test inside the OCCT software package).

To get the thermal resistance of the cooler in °C/W, simply take temperature difference between idle and load and divide it by power difference of the CPU between idle and load. In this case, 100W.

With both the Ninja and Havik, I got around 31°C idling.

With the Ninja, I got around 67°C under Linpack. With the Havik, I got about 68°C. Fan is at full speed of 1100rpm.

The Havik lost to the Ninja by an inconsequential one degree, which is well within the margin of error. A degree may mean a lot in other scenarios, but not when you have a 100W heat source raising temperatures by 37 degrees. In such scenarios, a small difference in performance is all needed to generate a few degrees of difference.

Calculating, that gives us 0.36°C/W and 0.37°C/W for thermal resistances.

0.37°C/W is not bad; it is sufficient in most cases, that includes a Phenom II X3 720 at 1.55V doing Linpack. While the best oversized heatsinks today have performances of 0.10+°C/W or even less, we are doing it pseudo-fanless.

So, the Havik did not beat the Ninja, despite being newer and more expensive. But -

1) It was further away from the exhaust fan
2) It does not have as much horizontal depth
3) Its fins are more densely packed
4) It comes with two good fans which I did not use

Furthermore, the Ninja's wide fin spacing design is known to be better for low airflow and worse for high airflow, relatively speaking. So the Havik matching the Ninja in low-airflow condition is a feat, because the Havik is definitely the better performer with fans attached. Then again, most high-end coolers are better than the Ninja with fans attached.

But the Havik was disadvantaged. So I redid the duct and redid the test.

I know a good friend called paper.

With the duct v2 in place, the tables turned with the Havik one degree better than the Ninja at 66°C for a thermal resistance of 0.35°C/W.

Which is still an inconsequential difference.

With more creative ducting, particularly making better use of the PSU fan, I am positive temperatures can get even better.

So, lets summarize this thing.

The Good:
Excellent pseudo-fanless performance
I can guess it has excellent fanned performance too
Relatively small (slim)
Comes with two good fans

The Bad:
AMD installation has some issues (Intel's look straightforward though)
Not as cost-efficient as some other coolers. But hey, the size, is small, and did I mention the fans?

The Ugly:
My casing's side panel can't close with it installed.

Muse M20 EX3 TA2021 amp - User review

2011-07-26T07:39:00.010+08:00

Measurements

I finally got around to measuring this amp. And building the contraptions needed to measure amplifiers.

3 sets of numbers here - for the amplifier at various volume knob positions as stated in the descriptions. (The -1.5dB refers to what the US-144 sees after the output has gone through the US-144's input's preamp, and is of no importance here.) From the noise level numbers you can see that most of the volume-control action happens in the first half of the knob. It is usually that way.

Again, bad crosstalk numbers because I did it in mono mode because I only had one contraption for one channel. See the graph for more accurate results.

Load is a 4.7ohm resistor. At 2 o'clock, output voltage is measured to be around 6V RMS, ±2dB because I don't know at exactly what level RMAA carries out its tests. (Arta didn't want to work with US-144 for some reason) At 6V RMS into 4.7ohm, that would be 7.7 watts.

THD pretty much stays the same or increasing slightly with output volume. The THD+N figures given by the datasheet is around 0.03%, so we are pretty much near the limit of the chip already (the measured THD+N is around 0.06%). And the complete product seldom hits component datasheet performance. For a cheap small class-D (or class-T) amp, this is pretty okay already, since bigger, more expensive amps are rating their powers at 0.1% THD+N, or even 1% and 10%.

IMD improves with output volume, but that is because IMD calculation is affected by noise, and SNR improves with output volume.

The next graphs to follow are the usually most pointless ones (unless something serious can be seen happening inside the graphs, which usually does not happen) - frequency response, noise level, and dynamic range. But for the sake of completeness I am including them.

The roll-off is steeper than expected and more funky. Possibly due to a 4.7ohm resistive load? Edit: On second thought, it could be due to the low-pass-filter of the recording device.

Nothing to see here. The noise drops swiftly from 20kHz onwards due to the output filter required for class-D amps. (Or due to the LPF of the recording device, or a combination of both.) With higher-impedance loads, like a real speaker or a tweeter where that frequency range is concerned, the roll-off would start at a higher frequency.

Also nothing to see here. Except for the 40+kHz noise from switching.

0.04% worth of THD. However, most the energy is concentrated at the higher odd-order harmonics so don't expect this 0.04% to sound as good as 0.04% from, say, a tube or transistor amp. A typical case of bad chip sound which I thought had been eradicated since recording studios started using equipment with chip amps in them.

IMD would've been okay if not for the strange peak at above 10kHz, possibly due to residual switching noise.

Stereo crosstalk. That is what happens in mono mode, ignore the white one and look at the green one. Otherwise, nothing special.

IMD vs frequency. Sharp rise due to the output filter, but I do not understand how the left channel got a different result. In fact I do not fully understand how mono mode in RMAA works. In fact I do not even trust the THD calculation, but nooooo, Arta refused to work.

I also measured another Tripath amp - Yulong T-Amp using the TA2024 and a linear regulated supply for comparison, and results are pretty much similar. The M20 EX3 has advantages at high output levels, where the TA2024 probably started struggling just that much earlier (1 watt) than the TA2021. At low output levels, tables turned with the Yulong delivering a nice 0.015% at Yulong's 9 o'clock, which is my normal listening position. BTW the Yulong's gain is around 2 "o'clocks" higher (or it needs 2 o'clocks less for the same volume) than the M20 EX3.

Noise numbers, Yulong has much worse numbers due to the hum from the linear power supply. Otherwise noise level is chip-limited and hence all perform similarly.

Subjective listening

Well, note that the following observations have not been blind-tested, and after the pictures and measurements have been taken it reeks of placebo. (Even seeing the amp produces placebo effect; that is why wood is so expensive)

First impression when I plugged it in, nothing special. From personal experience, this is actually a good thing, because things that sound too different usually show up big problems when measured later.

After listening for a while, the differences between this and the Yulong are:

The highs sparkle less, and by sparkle I mean in a neutral way. It can be good or bad, depends.
Midrange - more air and lower-order harmonics observed.
Bass - Might be a tad bass shy, can't tell differences quality (a reduction can often be taken as more controlled, but this is not always the case), could be better or worse, I can't tell
Overall - sonic differences between these Tripath amps is smaller than between other amplifiers

Bottom line

It is a TA2021. It performs like a TA2021 should. Which AFAIK is just the SMD version of TA2020. (Some people even claim that TA2024 is also the same thing, but some differences in performance in the datasheets exist, particularly the THD, which the TA2024 is lower and more stable, which may explain why the MX20 EX3 got stuck at 0.03+% while the Yulong went down to half of that.) It does not measure significantly better, nor is it significantly worse. It even has that Tripath sound.

Is it the best (among TA2020/1/4 amps)? Most likely not. Is it the worst? I'm not sure, but I'm betting worse ones exist.

Is it worthy of being sold? At $60 SGD for a fully-cased T-amp and power supply shipped, it is rather cheap for something with -70dB distortion peaks. But there are cheaper amps that are based on TA2020/4 with a chance of doing the same or better. If the chip is the deciding factor, then it does not matter what other parts are used as long as they are decent enough. But looking at some TA2020 amps (tend to be the cheaper-looking ones), they visually seem to fall under the not decent enough category, so I'm gonna leave it at a 50-50. More comparisons are needed for this. (Anyone wants to donate an S.M.S.L or a Topping?)

But, among the TA2021 amps, this is one of the cheapest if not the cheapest. And the TA2021 has some advantages over TA2024. It may be a few bucks more expensive the most value-for-money-looking TA2024 amps, but consider paying a few bucks more for the TA2021 which seems to carry a price premium.

ATi/AMD AVIVO Video Converter - GPU "hardware-accelerated" encoder: June 2011 update

2011-07-17T15:34:00.025+08:00

It is July, but I typed June, because I tested the June drivers and the July drivers were not out yet.

AVIVO... ATi claimed it was their version of hardware-accelerated encoding using the GPU. The first versions were exposed to be CPU-only after they were hacked for use with other brands of GPUs, including Intel's integrated graphics. Sure it was fast, but it was CPU-only, and the quality sucked.

After some time, ATi/AMD released an update to the AVIVO Video Converter which is locked for use with HD4600-series and better cards, again promising GPU acceleration. Restricting access to it for users with older, slower cards could mean either a genuine hardware update, or another attempt to smokescreen people into thinking that way, and preventing owner of slower cards from discovering the obvious truth when their HD2000-series cards accelerated as much as a spanking new and fast (at that time) HD4850.

Due to the lack of available hardware (more than one card capable of the acceleration), I will not aim to prove if this hardware-acceleration is true or yet another bullsh*t this time. (Underclocking the GPU does not help, because GPU usage is never full and the GPU rarely increases speed from the idling state when encoding. Though that should tell you about the truthfulness of the claim.)

Instead, I shall look at, again, if this program really works or is just a piece of crap repackaged to smoke the shareholders.

Here we have the AVIVO Video Converter completing an encode as per normal. Phenom II with 3 cores at 2.8GHz, and HD5670.

Note the GPU load. Pretty much minimal. And look at the CPU usage history. Just strange.

Now we have the AVIVO Video Converter complete the same encode but with the CPU locked at 800MHz:

The much longer time indicates it is CPU-limited in this case. So does the super-low GPU load graph.

But...

1) 800MHz resulted in a time taken that is 2.55 times that of 2.8GHz. 2.8/0.8, however, gives 3.5.
2) Not one of the CPU cores was at or near 100% in either scenario.

I suspect there is GPU hardware acceleration, but latency resulted in both the CPU and GPU waiting and doing nothing. It is also possible that the GPU accelerates only specific parts of the whole transcoding process, like resizing and decoding, so for the other parts it remain CPU-limited. The GPU did increase to max speed in the first encode with the CPU at 2.8GHz and possibly be GPU-limited, but with the CPU at 800MHz, the GPU didn't even care.

So what exactly is happening in the GPU? I don't know, and I said I am not going to find out.

But lets look at if the program really works. As in, it produces something good enough to justify its existence.

40 seconds, or double real-time speed for a 720p H.264 encode with resizing involved is still too good to be true even today. Well the first versions of AVIVO did show us that it is possible, but video encoding is always a trade-off between speed and image quality (or compression ratio), and AVIVO is the best example of that.

Because, have you ever seen massive macroblocking occuring in a 720p 4mbps H.264-encoded video? (Those 300MB for 24 minutes episodes you download from the internet, those are 1.667mbps inclusive of audio. 4mbps would give 720MB.)

Well now you have.

No, I know this is not massive macroblocking. Because this is massive blocking:

And I didn't even talk about the serious problem when dealing with interlaced sources. This source is interlaced, but I picked out the problem-free frames. This is a problematic frame:

It happens once every few frames as long as there is action. Just as all interlace-related problems do. This problem has been found and talked about by others since a few years ago, so I won't talk about it in detail.

In a previous look at AVIVO Video Converter (not posted on this blog), I commented that the AVIVO's disgusting image quality is justified if you need a lot of transcoding done in a hurry. Because the speed of encode is unmatched (Any other encoder would take multiple minutes, except for CUDA-accelerated programs - CUDA is a tried-and-tested beast that works, I've tried it myself). But so is the bad quality unmatched.

So I thought, since the compression is so bad, can I use an older, less demanding codec to achieve higher encoding speeds at the cost of lower compression?

And so I did the encode again, this time in MPEG1 using TMPGEnc 4.0 XPress.

As the settings in AVIVO Video Converter are limited or non-existent and inaccurate, I trial-and-errored a few times to get TMPGEnc to encode a file with a similar file size with similar settings - resizing to 1280x720 pixels, 29.97 fps, 1 pass, quality-based, 44.1kHz audio (even though the original file is 48kHz, yes the AVIVO program is that stupid), obviously I could not select the same video and audio codec settings since different codecs were used.

The resulting time taken is in the region of 50+ to 60+ seconds. So it takes roughly 1.3 to 1.5 times the time taken by AVIVO. TMPGEnc has a slight disadvantage here due to a slow built-in resizer, speaking of which, I was forced to resize to 1280x720 because AVIVO Video Converter does not even allow H.264 to be encoded in resolutions other than 720 and 1080 vertical pixels. And it does not check for aspect ratio either, so expect your non-16:9 videos to end up looking like crap.

This is the picture when encoded in MPEG1 at a similar file size by TMPGEnc -

Seriously, this is what 4mbps MPEG1 should look like. Ok, not really, the lines are blur because the source is blur (720x480 @ 7668kbps DVD source) and there is some blocking because it is a high-action scene with regions of low contrast. And seriously, AVIVO's conversion quality sucks. How can a H.264 video look worse than an MPEG1 with similar bitrate? I know the H.264 encoded faster, but it does not mean it is going to decode faster, if any difference it would be slower and take more hardware than MPEG1. And just look at the difference in quality. Would you use 50% more time to get that better quality? I would.

I'll say it clear, the few pictures taken are relatively hard to compress, and the source itself is blur (but relatively noise-free - it is direct from DVD, and there is a limit to what 8000kbps MPEG2 can do), so it is hard work for encoders. But even with other sources and other scenes the AVIVO Video Converter does not give acceptable performance. Yes this is at the lowest setting of 3mbps, but 3mbps is like more than enough already. Yet somehow with AVIVO it is not enough.

I rest my case. The TMPGEnc MPEG1 encode has more prominent mosquito noise but I'd take that over the blurred out edges any day. At any rate it is strange to compare H.264 against MPEG1 anyway. And yes this picture is in the wrong aspect ratio. Because the aspect ratio cannot be chosen in AVIVO.

Lets ask the question of why we encode. To save space, and more often to convert it into a format which can be played by a less-capable device. (It makes no sense to re-encode for a more-capable device if the more-capable device can play the original, because encoding only reduces quality, never increase or keep the same.) And the less-capable device tends to also have limited resolution, maximum bitrate allowed and disk space, or rather, disk space has always been a problem when you are encoding entire seasons of shows, even since the VCD and DVD days. Hence there is always the push for newer more efficient codecs and competition between codecs using the same technology.

And here we have a H.264 encoder that has a worse compression efficiency than MPEG1. The default setting for 1080p is 14mbps, which for H.264 results in (or should result in, as long as you are not using AVIVO) a quality as good as, if not better than digital HDTV broadcasts, and even nearing Blu-ray. If your source is as good, why do you even bother, and if your source is not as good, why do you even bother. And if your device can already play HD H.264, why do you even bother.

And since the program does not allow you to encode standard-definition and custom-resolution H.264 videos, and H.264 videos with less than 3mbps bitrate, this limits the usefulness of the program to "people with super-high-quality sources and/or has a lot of disk space to spend on a device that accepts H.264 but somehow not the source". Which is, basically nearly none, and any one that exists would have a better time with another encoder, like x264.

To conclude, the purpose of this program is probably to show the world that the creators have something that their rival also has, and hope that nobody uses it by making it a separate download, and if somebody does use it, hide the bad compression under high bitrates.

Me doing side-by-side playback subjective comparison

Disclaimer: This review has been heavily dramatized, the encoding quality is so bad only because of issues with interlaced source; with a progressive source (sleeping people) the quality is more comparable with the MPEG1 encode, with blur but noise-free of the H.264 vs sharp but noisy of the MPEG1. In fact, if without the source to compare against, the H.264 is subjectively less offensive. But still, comparing a H.264 encode to an MPEG1 is...

Random thought: Why is an extra op-amp in the chain always bad, but an extra (any of the 101 unrequired things you can put in a system) always good?

2011-07-17T00:51:00.003+08:00

Two terms created this time while I was in the army:

2011-07-09T20:31:00.002+08:00

All the F - the response to what the F, e.g. "WTF does this thing do?" "ATF."

Ohsnoesis - the movement of ohs noes from a region of higher concentration to a region of lower concentration

Quote of the day: "Be a messenger of truth, not a creator of lies"

2011-07-09T20:29:00.002+08:00

Quote of the day: "The value of a product is inversely proportional to the amount of marketing speak you see on its webpage."

2011-07-08T21:53:00.001+08:00

What you can get for SGD$50 nowadays

2011-07-02T21:38:00.004+08:00

Nokia 1800. Got it for $50.

Getting the power indicator LED(s) working on Muse M20 EX3

2011-07-01T00:04:00.005+08:00

At first I was cursing about how the Muse M20 EX3 does not have a power indicator LED. Turns out I was wrong.

There is, but it is so faint, it might as well don't be there.

I only noticed it because I saw a faint blue glow coming from the bottom side of the volume knob when the room is dark. In fact, even with the lights off, the light is so dim it cannot be captured on my camera.

So there is light behind the knob. So I pulled the knob out.

Four LEDs do the job of lighting up the blue glow.

But the light is blocked by the knob which is too fitting.

So pull the knob out until you get the amount of light you want.

How it should look like.

I don't know if this is a design flaw, problem during transport, or intentional. But easily adjustable power LED indicator is useful for me. You know how the newer monitors can set the power indicator light to off, so that you can enjoy the movie in total darkness?

Muse M20 EX3 TA2021 amp - a pictorial

2011-06-29T16:41:00.006+08:00

Around the interwebs

2011-06-24T19:44:00.002+08:00

"2) 1 pair Ultimate Bybee rca filters with 4 small Bybees slipstream purifiers. Connectors are Eichmann Silver bullet plugs. $300

These are meant to be connected to your interconnects, near the destination of your interconnect. For example, from your source to amp/preamp, then these are to be installed near your amp/preamp side.

The Ultimate version has a slipstream purifier on BOTH the signal AND ground connection. Excellent results o for digital audio data transmission (SPDIF from your DVD to your processor)."

I hate stories where there is only one lead guy and many girls

2011-06-17T10:08:00.004+08:00

But, I like Anime adapted from Visual Novels because the stories are generally good. By that, I mean most of the Visual Novels (or text adventure games, some of them don't count as Visual Novel, but the idea... you get it) that get adapted into Anime have generally good stories, although exceptions do happen. (Because I can safely say that, for each good one that are 99 rubbishes in the industry) Note: Does not apply for Anime based on games like strategy, action, RPG, SRPG, ARPG, even if they have any amount of dating sim or sex inside. You already know the usual outcome for those. Disaster.

The good story generally outweighs the pimp-master setting, or actually in many games, it's really ultimately 1 guy 1 girl per story, just that there are many stories for the protagonist to choose. Or the ultimate style of needing to go through all or at least half of the stories to know what is going on. Think mystery/suspense adventure games, but I've seen this being used for romance stories, multiple times.

So when I see a romance story with multiple guys and multiple girls and multiple relationships, I get interested. Not the first time I've seen, and not limited to Visual Novels, but is still rare.

This may have something to do with some Anime that I may be giving my views on, because I think it is not getting the appreciation it deserves (as with every time I post my views on an Anime).

Random

2011-06-15T19:52:00.002+08:00

Improve your ATi / AMD graphics card's video playback quality by DISABLING all the enhancements

2011-06-06T10:43:00.015+08:00

It happened one day (a long time ago actually), I acquired Nanatsuiro Drop's DVD creditless opening video, so when I got home I played it on my main PC with a HD5670 and SyncMaster 913V.

What I saw, was a putrid mush of ****.

It should be noted, at that point in time, my 913V was not calibrated. But that only makes the problem even more obvious.

I've seen the TV version of that video so I know what it looks like. My experience from comparing different encodes tells me how some fluctuation in color is normal, due to different tuner and tuning settings (more an analogue problem than a digital one nowadays), different broadcasted and subsequently captured raws, encoder and filter colorspace issue, and any enhancement the encoder (person) wants to add.

But even if you haven't seen the original video before, you can tell which one looks the most and least s**tty -

Four photos, with AMD video enhancements on and off, and display calibration on and off, in various combinations. +5 brightness and +25 contrast added in Photoshop as a noobish attempt to compensate for camera color response.

I'd expect some fight between 2 and 4, but 4 is closer to the original video, and 2 looks worse when big (click for 800x600). Photo 4 is with display calibration on and video enhancements off. As you can see this screen is fairly inaccurate while uncalibrated (photo 3), it is one of the earliest 19-inchers and only has a VGA input.

With an uncalibrated screen and enhancements on by default, I was greeted with the oversaturated POS of photo 1.

Fortunately I already knew that ATi / AMD cards have video enhancements on by default all along, and that it only works with MPEG2 (now also H.264 and VC-1 as long as you are using DXVA), so for all the MPEG4 ASP and AVC videos it didn't have any effect so I ignored it. But this video is MPEG2.

Speaking of oversaturation, this is what can happen with enhancements on, at least with an uncalibrated screen -

You know something is wrong when dark hair turns from being a light absorber to a light emitter.

Enhancements can be good sometimes, but they can be very bad many other times. So it should be left off by default unless the user wants it. However this is not the case with consumer video products, where upscalers, more vibrant colors, sharper edges, and motion-enhancement-whatever flood the market.

Xonar DX RMAA

2011-05-26T13:34:00.003+08:00

My new measurement card.

SGD$128 for this kind of performance. Some manufacturers should be ashamed.

The D2X is rated almost twice as good at twice the price. That is the card to go for if you have a bit extra moolah. For now, this is enough to help weed out the bad products.

List of the amplifier chips used in some studio monitor speakers - thoughts

2011-05-18T23:43:00.005+08:00

List of the amplifier chips used in some studio monitor speakers

Pro-audio is one place I expect to find balance of component performance. Because with so many competitors as well as using measured numbers to reflect performance, it is about the best performance for the same money, at least on paper. And to do that, bottlenecks must be minimized, and the performance of each part matches the performance of the rest.

With so many gainclones being used, this makes me wonder about the performance of T-amps and such.

Gainclones aren't bad, in fact they perform very well for the price.

And that is actually the problem. Look at the number of LM3886 being used. And usually two of them are used for a bi-amp configuration.

And take a look at the amplifiers themselves:

(Source: Everywhere on the internet, I just visited webpages that were given to me by Google Image Search, if you demand credit or need it removed just inform me)

M-Audio BM5A

Mackie HR824

Yamaha HS80M

ESI nEar05

Tannoy Reveal 5A

Genelec 8050A

Behringer B2031A

I could go on forever, but you should get the idea already. Even if we ignore the complicated preamp section, looking at the power amp sections themselves, well they are not fancy with multi-colored components and "magic materials" or what the ****s, but they certainly are no slouch either. Well you can see some amps being bigger than others due to price differences of the speakers, and some amps more worth the money than others, but in general, a 2xLM3886 amp, complete with parts of those size, costs quite a bit already materials alone.

I don't expect cheap gainclone kits to do any better. Or even the same, those cheap kits using cheaper or fake chips and tiny capacitors.

And that is the problem. If pro-audio designers are using this quality of gainclones with their $300-1000 active speakers, then passives speakers in the $300-1000 range should demand even better amplifiers.

And Creative T20 Series II and T40 using TPA3123D2 and TDA8932 respectively tells more about the value of these amps.

Now I'm doubting that <$100 amps are sufficient.

Random thought: Hearing loss and frequency range

2011-05-17T10:10:00.006+08:00

Since I was doing a frequency sweep might as well check my hearing.

Reliable detection - 18kHz. Don't know at how many dB down. Not bad considering my ailment. Might have dropped I dunno, I can't remember my 18kHz loudness from last time, and it can be due to equipment. (The Atom's 18kHz is rather high compared to the Diamond 8.2's)

Just to make sure I also did all the way till 14kHz in 1kHz intervals. I'll never understand how the ear interprets these high frequencies.

But one thing in common as I listened to all those tones.

They all sound like ****.

They are meant to be a screech alright. But the similarity...

That is when this struck me. A basic concept in music and acoustics that had escaped me. (Well because nobody mentions this during discussions involving hearing range, because hearing range translates into phallus size.)

The octave.

The ear differentiates tones by their relative differences.

An octave = doubling of frequency
A semitone - 1.059463 times frequency

So, if we take 15kHz as the reference,

A person with an over-perfect upper limit of 22.5kHz can hear 7 semitones higher, also equals to a fifth (e.g., C to G), or 1.5x frequency
+6 semitones = 21.2kHz
+5 = 20.0
+4 = 18.9
+3 = 17.8
+2 = 16.8
+1 = 15.9
0 = 15.0
-1 = 14.2
-2 = 13.4
-3 = 12.6
-4 = 11.9
-5 = 11.2
-6 = 10.6 = 1 octave lower than 21.2

So even if your hearing range has deteriorated until the seemingly bad 10.6kHz, you have only lost 1 octave compared to a small child.

And human hearing range has 20 to 20.5kHz = 11 octaves to start with.

So you would've lost 1 octave out of the original 11.

Big deal.

This argument supports the phenomenon that musicians and old listeners, although their hearing range could have deteriorated to below 10kHz, it is a relatively minor handicap and their experience more than makes up for it to make them better listeners.

(Note: This is partial bullsh!t. Yes the maths make sense, but you can try the effect of a high-pass filter via a parametric equalizer and cut the highs at different frequencies and see the result. It is not unnoticeable.)

Lord praise the guy who invented Li-ion

2011-05-11T18:45:00.005+08:00

...Or we would still be using 3 NiMh cells and getting two hours of runtime out of hand-held gaming consoles.

Now if only someone can standardize the form-factor of Li-ion cells, may lord praise him too.

List of the amplifier chips used in some studio monitor speakers

2011-04-30T07:15:00.002+08:00

Got this from here, now to make a copy before it disappears
http://toki.2ch.net/test/read.cgi/av/1245925463/l50

Alesis M1Active / M1Active 620 - LM3886 x2 - National Semiconductor
Behringer B2030A / B2031A - LM3886 x2
Dynaudio Acoustic BM5A / BM5A Compact - LM3886 x2
Fostex NF1A - LM3886 x2
Event - PS、TRシリーズ LM3886 x2 / ASP6 ASP8 (LM3886 x2) x2
ESI nEar05 eXperience - TDA2052 + ？？ - STMicro
Genelec 1029A - LM3886 x2
Genelec 1030A - STK4036V x2 - Sanyo
Genelec 1031A - STK4241V - Sanyo
Genelec 8030A - LM4780 - National Semiconductor
Genelec 8050A - STK442-130 - Sanyo
KRK V4 V6 V8 Series2 / VXT6 - LM3886 + ？？
KRK Rokit Powered RP5 RP6 RP8 - TDA7294(LF) + TDA2052(HF) - STMicro
M-Audio EX66 - TDA8927 (Class-D) - NXP
Roland DS-5 - LM3886 + ？？
SONY SMS-1P - UPC2505 - NEC
Tannoy Reveal 5A 6D - LM3886 x2
Yamaha HS50M HS80M - LM3886 x2
Yamaha MSP7 Studio - STK415-130A - Sanyo

Go gainclones.

Energy conservation

2011-04-23T13:33:00.002+08:00

As I read this topic, cost savings and ROI keep coming up.

Use less electricity = lower electricity bill.

And if the whole world uses less electricity, demand goes down, price per unit goes down also, because it is cheaper to generate.

Right?

My ass.

Cost-based pricing is never used in the real world.

And if OPEC realizes they are selling less oil, they will simply increase the price instead so they earn the same amount of money.

And locally we have good examples too.

When oil prices increase, electricity and transport prices go up.

When oil prices drop back, electricity and transport prices go up, citing "other costs".

That's why you can't trust the system.

I just discovered I can tell the difference of 3cm of head positioning

2011-04-22T21:20:00.009+08:00

I've been feeling that my left speaker's a bit louder than the right. Thinking it could be either due to room acoustic, supported by the fact that the perceived bass is stronger on the left even though the subwoofer is on the right, I ignored it.

Today I was thinking that the left/right volume can be altered by the positioning of the head, so I decided to shift my head towards the right to get the perfect balance. When I got the balance I want, I looked at my head's position...

...to find it right in the middle of the two satellites.

You see, due to some strange situation of my table, my monitor is placed 6cm closer to the left satellite than the right. And I always make sure my head is at the middle of the screen. Which resulted in my head being 3cm off the sweet spot.

But what's a concept without testing.

Test methodology

Music is played through the speakers. Subwoofer volume turned to zero to prevent it from influencing the balance. The room lighting and computer screen are turned off (it's night btw). Eyes remained closed throughout each test until each measurement is taken.

1. Stand up from the chair, pulling it a short distance away from the original position.
2. Walk a metre or two away and back, pulling back the chair and sitting down at a random position. The direction to face is decided by the chest feeling against the table edge (my table is fairly large enough to do this).
3. The positioning of the head is shifted left right forward backward while facing the same direction as much as possible as the center position is attempted to be found.
4. After deciding position of head, a torchlight, one end placed against the chin and the other end pointing at and close to the table is switched on. There is a piece of paper fixed onto the table and where the torchlight is pointing at (the brightest part at the center) is marked.
5. Repeat from step 1 for next measurement.

A total of 6 readings were taken with the speakers at the original positions untouched. Then the left speaker is shifted 6cm to the left such that the distance between each satellite and the screen is both approximately 9.5±0.5cm. (Originally the left speaker is approximately 3.5cm away) This results in the sweet spot shifting left by 3cm. Another set of 7 readings was taken.

The screen is 53cm wide, so the original distance between the speakers is 66cm while the new distance is 72cm - an increase of 9.1%. The speakers are approximately 65cm from the edge of the table. The resulting change in distance between speaker and "sweet spot" - taken to be the edge of the table, equidistance from each speaker, is 1.9%, and resulting increase in angle is 7.7%.

Results

The first set of readings are those directly above the horizontal line while that of the second set are indicated by a circle.

It is not a lot of readings and variance is fairly huge. However, two important observations - the average distance (from the left side of the paper) of each set of readings differed by approximately 3cm, while the differences between both maximums and both minimums are both 3.5cm. The difference between max and min values of each set is 4cm, resulting in a 0.5cm overlap.

Which means that it is fairly possible for the human ear(s) to detect changes in sound due to the sweet spot shifting by 3cm. 5cm would be a safe estimate.

Note that this little test does not reveal which position is the more accurate one, since there are other factors that affect balance of both channels - channel imbalance of equipment, room acoustics, hearing loss etc, but the main objective is to see if audible differences in perceived sound can result from small shifts in the positioning of the head relative to the speakers.

Granted, I used some sort of averaging technique to increase my precision - I shifted my head left and right to find the positions at which one channel is a certain degree louder than the other, and took the center. Like the way you tune the TV and focus the camera. But my head only moved by a few cm in the process (although it feels like a lot). This adds more weight to the idea that our hearing is precise to a few cm.

I'm not really surprised by the result - animals are able to pinpoint the location of predators and preys that are far away by just hearing, and some are even able to use them to navigate and eat prey. This would require even better precision. Way, way better. If I go and search I believe I would find better technical articles on the precision of human hearing.

This has implications on speaker positioning and how subjective listening tests are being carried out. I already know that some instruments in a particular song sound relatively louder and softer depending on the position and angle of my head, but having numbers puts things in perspective. We would need to be accurate to a few cm and a few degrees when positioning our speakers, and in subjective listening tests you may be hearing differences that are not actually there due to the head not being in the same position. And how many people actually make sure their heads are in the same position throughout the test? It also questions the credibility of auditions carried out simultaneously by a group of people, for no two heads can be in the same position at the same time, although subjective bias caused by the first guy to open his big mouth is more likely to cause everybody to "hear the same thing".

Looks like we may need, as it is jokingly mentioned on some forums, a vice clamp for our heads.

[Update] JDS Labs cMoyBB v2.03 OPA2604 measurements

2011-04-15T17:29:00.008+08:00

I noticed some peculiarity with the OPA2604's distortion graph. This is a chip that is used in some high-end designs achieving lower than -100dB THD+N, so the graph I had wasn't reflective of its true performance. And I was checking this blog's stats on a hunch, and found many traffic sources from Google image search, with the search term OPA2604.

People may get the wrong idea reading that graph, so to reduce the amount of rubbish on the internet, I decided to do a retest with a 12V power adapter this time.

Results are shocking.

The EMU wasn't cooperative that day, and I was in a rush, so I didn't bother getting 0.001% THD. But this is good enough to show the previous measured performance of the OPA2604, which was shocking bad as >0.01%.

The first thing to notice is the numbers for THD dropped by an order of magnitude.

Now it measures okay. Most of the spikes are EMU's "noise floor" (for the lack of a better term that I know).

And yup, this is with 9V battery. So the issue wasn't insufficient voltage. This is an example of why control measurements are important, for if I had measured only 12V scenario, I would had gotten a different and wrong conclusion.

The main difference is the greatly-reduced even harmonics.

My guess for the cause? Is not going to be something groundbreaking - probably a bad contact.

Other graphs are similar, 12V and/or 33Ω load, so I won't go into that. Into 33Ω for both cases THD is 0.005+% but how much of it would be due the the EMU? Even-order harmonics did increase with current draw as with the case of OPA2132. Speaking of which, now I wonder how accurate are the measurements for OPA2132, particularly the 0.015% reading.

So here I declare the previous OPA2604 measurements invalid. Some others might also be questionable.

Boston Acoustics HPS 10HO subwoofer repair - Day 2

2011-04-14T18:48:00.005+08:00

Day 2 start - got my fuses, lets go.

The symptoms on day 1 got me feeling that it is the power amp section that is spoilt. So I have to disconnect just the power amp section. A good way is via removing the rectifier. But before I do that, lets look at a strange thing I saw on day 1:

Yes, one leg of the rectifier is missing. In replacement a wire is soldered on half that leg and the solder pad on the underside of the board. Why would anyone do that is out of my comprehension.

This does hint of something wrong with this portion of the circuit, or maybe the rectifier. Unfortunately the rectifier measures fine out of circuit and unpowered, plus the humming when turned on (after the capacitors get their power before the fuse blew) indicates something wrong further downstream. There is also a previously unnoticed burn mark on one of the connector pins for the wire from the transformer, which further hints that the power amp section is the problematic part.

The 20W of my Goot iron really shows its inadequacy here, I am unable to desolder the bridge. Thick leads and traces conducted heat away too fast. I had to cut the legs off to remove the rectifier then desolder the remaining sections of the legs.

The right side of the board, the blue capacitors and the parts below it, is the power supply for the pre section. This tested working fine, and nothing blew up with the pre board connected either. Now I just need a way to tap the pre-out to ensure it is working, then I'll stuff an amp inside and finish it up. But this is enough for the day.

Bugger has 50VAC x 2, wooorr. But that is also a problem because I don't remember any chipamp that can handle ±70VDC. (Erm yes those big capacitors are 6800uF 80V)

God I want one

2011-04-14T18:39:00.002+08:00

For some reason this lab also has many Audio Precision - the early 2700 series units. Many faulty ones - burning smell when power on - all lined up on the shelf.

Also has Neutrik audio analyzers. Just what does this lab do?

Juvenile

2011-04-14T18:19:00.004+08:00

One day I heard lots of mynah chirping (as in a lot of the chirping, not mynahs). I looked to the source thinking it was another mynah fight. But no, all I saw was four mynahs standing close to each other. Two of them were slightly smaller than the other pair. Upon close look I realized they were juveniles.

They can already fly and are walking around pecking at things, but following close to their parent and chirping incessantly. Probably learning how to hunt.

On first look they can be easily mistaken for adult mynahs, since they can already fly and such. But they are not adult yet and probably still stay in the same nest as their parents. Isn't it the same for humans; some of us appear to have reached adulthood, but are missing one critical step to really becoming one. That includes NSFs, uni students, businessmen, investors, inventors and entrepreneurs - those who stay with their parents and throw their savings away via their businesses because they can't do real jobs.

"Wouldn't the cheap ring terminal bottleneck the performance of the faceplate?"

2011-04-12T16:57:00.001+08:00

I was chuckling hard when I typed that.

On a serious note, hire a licensed electrician for in-wall mains wiring, or check with your local authorities on required permit and inspection. Also, make sure the connections are rated for mains and safety earth, though that isn't an issue if a licensed guy carried out the job. You wouldn't want to void your insurance.

[Rant] Pokemon - Infernape

2011-04-10T21:38:00.000+08:00

How the hell does a monkey know stealth rock?

Non-fiction fail or win?

2011-04-06T11:32:00.002+08:00

They are real!

Tweens 250mAh 9V (8.4V) rechargeable battery - first charge/discharge cycle (almost) complete

2011-04-04T20:14:00.005+08:00

Just to prove that not all China batteries are crap.

Tweens 250mAh 9V (actual 8.4V) 250mAh battery, why did I get the cheapest battery that is 250mAh instead of paying a few tens of cents more for 280mAh?

One reason is it has the highest number sold among the sub-USD$3 batteries.

Second reason is that you wouldn't believe that these batteries have over 200mAh anyway. And the same manufacturer probably made them and put them in different packagings.

And the charge rate is a giveaway (ya the manufacturers are nice enough to print the charge rate on the sticker) - 15 hours at 20mA. Usually the formula for charge rate and time is 1.5 x 10 hours x 0.1C. 10 hours @ 0.1C gives full charge but some extra is needed to make up for losses and give it a bit of overcharge. That 20mA pretty much tells us the actual capacity.

Well, there are batteries on eBay stating 300mAh and 30mA charge rate. Maybe I'll try that out when I need another 9V battery. But by then Chinese LSD 9Vs would most likely be out.

Ok, enough talking, the most important part of this post:

Is not the graph. But the note that I only charged it at a rate of 14-15mA on average for 700 minutes, and the charge rate slowed down towards the end before I terminated the charging. There are reasons for this but I'm not going to waste bandwidth. That gives us roughly 160-170mAh of charge, after losses I'm inclined to think the former.

Without further ado, here is the graph for the first discharge after the first recharge (or the second discharge if you count the discharge I did right out of the box):

Discharge rate is 10.1mA, which is the current the LM4562 draws. That gives us about 150mAh.

Pardon the lack of readings at the start which makes the graph too linear at the start. I didn't have the opportunity to check it. I'm guessing it should be linear up till 8.8-8.9V before it shoots upwards (counting from the back).

If you take into consideration that full capacity could be higher, not bad for $3 SGD. At least it is not crap. Now that I know, I can do a 200mAh test, if I decide to do one.

Untitled

2011-04-03T16:35:00.001+08:00

Measured current draw of some op-amps

2011-03-29T21:47:00.004+08:00

As I was going to use my cMoyBB on a regular basis, I was curious as to which op-amp to use for maximum battery life.

Of course I can search the datasheets, but what's the fun in that. Also, I want to try to find clues that can tell me whether my OPA627BP's are fake.

LM4562 10.1mA
OPA-Earth 49.1mA
OPA627 14.0mA for two
LM6172 >60mA (error)
OPA2604 10.0mA
OPA2227 7.3mA
JRC4580 5.1mA
OPA2132 8.8mA

Yup, the numbers are as the datasheets say, mostly. LM6172's is over 60mA and increasing, and it was getting warm fast. It still has a very high DC-offset (>100mV!) despite the unity-gain mod - LM4562 dropped from 20 to 2.3, JRC4580 from 40+ to 10+, and OPA2227 from... I forgot from how high/low, but now it is 0.1 and 0.3mV.

OPA627 taking 7mA each as according to the datasheet, low DC-offset of... 0.0 and 0.0mV which was expected when reading the datasheet too. Things are looking better for these unknown-authenticity chips, some THD and load tests (to test the 45mA output current) later will wrap things up.

Subjective listening yielded good results too - particularly the bass, which is powerful yet clean, as though all the other op-amps I've tried so far have difficulty driving the 32-ohm MS1 (I haven't tried high-current ones like AD8397. And the background instruments come out more too. The main instruments are also more forward and wider but less deep soundstage which is not to my liking.

I wasn't specifically listening for anything nor having any expectations so it pleasantly surprised me. But hey, I've demonstrated to myself that I can hear tweaks even before doing them, so take these results with a pinch of salt. I've also failed a recent "Name that op-amp in the dark" test, although I have more confidence of identifying this OPA627 but these are not possible to do that test with.

Well, these chips are looking to look like the real stuff, pity I sold off two pairs cheaply at SGD25 per pair because I wasn't sure if they were real.

Random thought: Amplifier output power ratings, 4Ω vs 8Ω

2011-03-27T10:04:00.003+08:00

Typing the Ω symbol reminded me that I have a lab report to complete. It's week 9 yet I feel like zero hour.

Many if not most amplifiers specify max output power ratings at 4Ω or 6Ω for those that cannot handle the high current, rarely 8Ω. Amplifiers and limited by both the voltage their power supplies provide and the current their output stage can put out. However, in marketing speak amplifiers are always be able to source or sink 1000A of current for a nanosecond. Hence current is not a problem and output power is limited by voltage instead. Since V²/R, the smaller the R, the bigger the power.

Hence the Tripath TP2050 power stage chip is a peculiarity. It is highly current-limited such that it is rated at 8Ω in the datasheet (although it is also mentioned that the minimum it can go is 6Ω).

However, there is something interesting about this chip - in stereo mode, it is a pair of bridged amplifiers. Change some resistors and switch it to mono mode, the two bridged outputs combine together to form a 100W into 4Ω amp that is capable of twice the current. Hooray for FET output stage. These numbers also suggest that the max power into 8Ω in this mode is still 50W, but hey extra headroom is always nice.

Now, this thing has my interest because some completed product amp manufacturers misuse those numbers, dunno if it is by their marketing department, or a careless designer. 100W and 50W output powers were quoted with different combinations of number of amp chips and load impedance. You can see that some combinations don't work.

Official specs for this chip is 50W into 8Ω in stereo mode and 100Ω into 4Ω in mono mode. But nothing was mentioned about power into 4Ω in stereo mode.

So a few calculations using Ohm's Law later,

Current limited = fixed current = power proportional to resistance = half of power into 8Ω = 25W

Well, that is veryyyyyyyy small compared to 50W.

For a few seconds it didn't make sense, until I remember the other form of the formula P = I²R. In paralleled mode it has twice the current capacity. Hence four times the power.

Well, the amount of thinking that was carried out would take just a few seconds, but this reminded me that a good amplifier should never be current-limited. And amplifier power ratings are screwed up.

Evil truck is evil

2011-03-24T18:15:00.001+08:00

What's the difference between a programmer and a manga artist?

2011-03-22T21:04:00.002+08:00

The manga artist chose to torture himself for fun and pride.

JDS Labs cMoyBB v2.03 Review: Measurements Section

2011-03-17T08:06:00.025+08:00

Update: The measurement results for OPA2604 have been deemed invalid. Update on this here.

E-MU 0404 PCI - Detailed RMAA Results

The card I used to measure - E-MU 0404 PCI

Note that the first measurement is taken a long while back while the other two are taken on the day the cMoyBB is being measured and with the same setup, minus the cMoyBB of course. Also those voltage values are RMS voltages out from the E-MU 0404 line-out and into the E-MU 0404 line-in, which are the voltages at which the cMoyBB will be outputting for this review. The small things to be done that avoids false bad measurements of the thing to be measured.
For the above measurements the E-MU's output voltage is lowered by software, but counter-checking measurements show no appreciable decrease in performance figures. When feeding the cMoyBB, the E-MU is at max output.

Also note the strangely bad crosstalk numbers. This is because I only measured one channel (the left) and the right channel was duplicated from that one. Ignore those numbers for now. (Although you can also use it to find, not absolute numbers, the relative performance between the different things.) Or, you can look at the stereo graph versions.

The test

The cMoyBB will be tested in four conditions:

1) 1V into no load
2) 1V into 33ohm resistor
3) 0.2V into no load
4) 0.2V into 33ohm resistor

The voltages (in RMS) are chosen as such because 1V is the standard 0dBV line voltage, while 0.2V into 33ohm gives V^2/R = about 1mW of power into the load, and most ear/headphones' sensitivities are rated @ 1mW.
0.2V into no load acts as a control, to see how much of the distortion is caused by the 33ohm load.

Lower impedances require higher current draw, so 33ohm was chosen, and 1V into no load tests for scenarios where high voltage swings but low currents are needed for high-impedance headphones. Even for a 600ohm headphone, 1V into it yields 16.7mW, so unless its sensitivity is in the 80dB's and you have bad hearing, that is more than sufficient.

1V into 33ohm is a scenario you will never see in real life unless you want to destroy your earphones and/or your ear. But I'm including it anyway just to see how the amp handles such a scenario and how much headroom it has. But in the event of a fight-to-the-death between your earphones and your ear, my money goes to the earphones.

cMoyBB v2.03 - Default OPA2227

cMoyBB v2.03 OPA2227 - Detailed RMAA Results
cMoyBB v2.03 OPA2227 - Detailed RMAA Results - 0.2V into 33ohm

Starting off we have the cMoyBB v2.03 with its default op-amp OPA2227. Well, since this review is about the cMoyBB.

No load 1V and 0.2V give the same numbers as E-MU 0404 PCI, which means the performance of this amp is most likely better than what E-MU 0404 PCI can measure.

1V into 33ohm proved too much for the amp to handle with rocketing distortion numbers and huge signs of clipping, but even so it is below the 10% THD that many amp manufacturers like to rate their products at. Not that you should stand for that, but just saying.

0.2V/1mW into 33ohm, now we see much better numbers. However, compared to 0.2V without any load, we still can see some harmonic distortions that get significant from 10kHz and up. By significant, I mean there are more spikes that are still less than -100dB each in the audible frequency. Maybe you can hear a little bit more "detail", maybe not.

Notes on op-amp rolling

What's the use of the DIP socket if you're not going to roll op-amps?

So far I tried LM4562, LM6172, JRC4580, OPA2132, and OPA2604. LM6172 and JRC4580 have too high DC offset when used in this circuit, while LM4562 have around 20mV, which is high but still usable with care.

If in doubt, use FET-input op-amps.

LM4562

LM4562 - Detailed RMAA Results
LM4562 - Detailed RMAA Results - 0.2V into 33ohm

There seems to be some problem when measuring LM4562, which resulted in abnormal and fluctuating noise levels, which affected the distortions results to some extent. But you can always see the detailed results.

Unlike the OPA2227, the LM4562 had almost no problem driving the 33ohm load at 0.2V, with numbers close to the E-MU 0404's limits. Well, almost, because as you can see below, there is still some extra spiking at above 20kHz, and between 10 and 20kHz the spikes are taller than the EMU's, albeit by a few slightly audible dB's.

Known for its accuracy and price, the LM4562 would have been an ideal drop-in replacement in the cMoyBB if not for its DC offset, and some noise issues and frequency spikes that might have been caused by the noise. Needs to be retested when possible. Granted, it was not in a proper circuit for it anyway. But still, overall performance, minus the DC offset, is excellent.

OPA2132

OPA2132 - Detailed RMAA Results
OPA2132 - Detailed RMAA Results - 0.2V into 33ohm

Burr-Brown's "audio op-amp" pays a visit. "Audio grade" means bad performance on the whole, so there was not much expectation for it. On the other hand, OPA2227 is said to suck too, so there is hope for OPA2132 after all.

The distortion pattern from 10kHz onwards look similar to the OPA2227's, only smaller in magnitude. However between 1kHz to 10kHz there is much more lower-order harmonic distortion. Some may like the sound, some not, but in terms of accuracy, OPA2227 would come out ahead of the two.

OPA2604

Update: The measurement results for OPA2604 have been deemed invalid. Update on this here.

OPA2604 - Detailed RMAA Results
OPA2604 - Detailed RMAA Results - 0.2V into 33ohm

At SGD$6, this isn't much more expensive then the rest of the contenders. However, it's minimum rated supply voltage of ±4.5V is right on the voltage that a 9V alkaline battery can provide. I was slightly worried about this, but my worries proved unfounded.

The higher current output capability over LM4562 (35 vs 26mA) proved to be useful -no sign of clipping anywhere. Even though it was working at its lowest voltage limit. Well, I did read somewhere that the absolute lowest required is a bit lower. So it was not the TLE2426 that was bottlenecking. (It shouldn't, by right, because 0.2V / 33ohm = 6.1mA) But that also means all the other chips have a bit of current issues in this particular circuit, specifically, CMoys that are powered by a 9V battery pushing 0.2V into a 33ohm load. Further experimentation is required.
But then again, most of the shitz is above the audible range, so should it be cared?

Not only that, OPA2604 also gives a very nice hill of harmonics, highest at the better-sounding lower-orders, and decreasing as the order increase. OPA2132 did have something like that, but the gradient was less steep so there was more of the nasty higher-order. Also, OPA2604's hill is there even without any load, evident from the no load 1V test. It's as if OPA2604 synergized with this circuit to give a warm sound. Measured THD numbers are bad of course, but it may sound much better than what the numbers say.

A summary of the performance numbers (OPA2132 omitted):

Conclusion - cMoyBB

There are amps that measure bad, amps that are expensive, and amps that are both. With no-load @ 1V performance better than what the E-MU 0404 can measure, 33ohm @ 1mW just slightly worse, real-life usage can only result in even better numbers. At these numbers, the bottleneck is likely to be your whatever player if not headphones. Or, pop a suitable op-amp like OPA2604 in to change the sound while still retaining great performance. This thing definitely comes out great in the measurement grading, and the only reason I'm not giving it excellent is because it didn't get below E-MU 0404 performance while driving 33ohm @ 1mW. Though, with signs that it is the chip that is being the bottleneck, I doubt other similar designs using the same chip would do better. Or you can slap an AD8397 onto this amp, but do you really need that much output current while sacrificing the ability to roll op-amps to suite your taste in sound?

But you know, I'm an anal guy. So great, no excellent. And I blame it on bad initial op-amp choice. Wakakaka.

As for the price grading, I know of...... no portable amp that gives this type of performance for SGD$80. Those that probably do don't have measurements. Only many bad ones that cost less or around this price. Even if you go higher, there are some decent measurements of some amps but the measurements do not mention voltage and load, so it is difficult to judge.

Engineering is the art of compromise, and this amp did it right. Don't want output capacitors because they audibly and measurably add distortion, check. A TLE2426 instead of a bad resistor divider in order to avoid output capacitors, check. No need for extra current capability via ground buffer and certain op-amps (you can't do it without either), saving costs, check. The numbers don't lie, and this thing is perfect for most sane situations as a portable headphone amplifier. Sure, you can get better performance by paying more, and worse performance while paying less. But you don't really need higher performance, and you shouldn't stand for worse performance since it is easier to design an amp that performs better than a DAC (that includes portable players and soundcards - anything doing digital-to-analog conversion), so the amp should never be the bottleneck. And this thing smacks right at that point. Getting the job done right at the lowest price. Okay, maybe not really, I would've wanted slightly more current. But then again, that's the op-amp's fault.

So unless I find an amp that either 1) achieves the same performance for less money, or 2) achieves better for the same money, this thing will be at the top for value. And with measurements of other amps hard to come by, there is a lack of evidence to bring this guy down.

Conclusion - op-amp rolling

Being able to use only FET-input op-amps reduces op-amp choices, but there are many high-performance op-amps with FET-input.

LM4562 is the sure choice if you want clean and accurate, and if you can live with the DC offset. Changing some resistor values will solve that problem, I will need to look into that.
OPA2604 also works pretty nicely too if you want that kind of sound.
The default OPA2227 with its super low DC offset isn't bad either, with somewhat decent performance. But there might be better alternatives out there.
AD8397 (untested) might be a good choice, if you really need that current. You most likely won't, but hey.

Random thought: Why do "The Smartest Gurus" only mod others' products instead of designing their own?

2011-03-13T21:45:00.002+08:00

Boston Acoustics HPS 10HO subwoofer repair - Day 1

2011-03-10T21:35:00.005+08:00

Note to self: a 15.7kg white cube is very hard to carry by hand

Fuse intact.

Power on, nothing exploded. Nothing happened at all either. Fuses intact.

Took it apart.

Transformer secondary continuity check pass.

Transformer primary continuity check pass.

Ok, now this is strange.

Power cable continuity check fail. And at neutral somemore.

Knn.

Changed cable. Power on. Voltages on secondary side okay. At this moment the transformer is not connected to the power supply cum amp board.

Connected transformer to PSU board. Power on. Hum for a few seconds, fuse break.

So the good news is - transformer is okay. And the bad news is - the power supply/amp board isn't. Well, something has to be broken somewhere.

But the worse news is - there isn't any sign of damage on the power supply board.

Well, an amp board with built-in power supply is easy to get.

But the even worse news is - the stupid preamp stage board is fking hot-melt-glued to the metal plate like a motherf*cker. That means to keep things clean I'd better reuse the preamp board. But the even worse news - I can't even tell how that shit is connected to the power supply/amp board via a 7-pin connector, and what voltage it uses. And if it's working.

Well, that's enough for day 1. Diagnosis will continue after I get more fuses.

Random thought/quote:

2011-03-08T13:43:00.001+08:00

The size of the audio system owner's penis is proportional to the size of the low-frequency driver.

Almost completed - Linear post regulator

2011-03-05T21:44:00.007+08:00

I needed a 15V power supply to replace a dead one. And something at least 2.5A.

But there are two issues:

- 15V power supplies of >2.5A are significantly more expensive than the more common 19V (at the time of my searching)
- Always expect worst-case scenario when not proven otherwise - the DC quality of these power supplies are crap. And when I say crap, it is more crap than a typical desktop computer power supply. Much more crap.

Hence the idea of using a linear regulator + filter to clean up the circuit and drop the voltage at the same time came up.

Thanks to a group of helpful and knowledgeable enthusiasts over at diyAudio forums a good design was created very quickly. I implemented one of the earlier designs but the topic has progressed even further.

I will not post the design I used out of respect for the community by keeping the internet free of bad circuits when superior ones exist, but basically the idea I used was capacitor multiplier with one RC stage at the base of the darlington, fed to a LM338 linear regulator.

I so need a camera. PC Show coming soon.

Missing input connector and output strain relief. Grommet I got was too small.

P2p build, active devices bolted to the tin case which doubles as a heatsink.

I'm getting 38°C with no load and 46°C with load for the LM338, and 34°C no load 38°C load for the transistor, taken directly from metal part of the TO-220 packaging. Same 29°C air temperature. Compare it with the 47/54 heatsink temp with the Zhaolu power supply.

I don't know if this or having the cover closed, and the surface with the components bolted-on flipped over to face the top would provide better cooling, but for certain, at these temperatures and power draw, it would not matter.

Fun with power

2011-03-03T22:13:00.004+08:00

Since my regulator is coming soon, I had to check how much heat will be coming out of it. So I had to see how much the amp draws and calculate from that.

0.24A. No change whether playing music or not.

Lets make it a quarter of an ampere for easy calculation. The regulator drops 4V. 4V @ 0.25A = ...1W.

That's pretty easy to dissipate with a mints tin. Heck, even 2W would be possible.

Just that the full power dissipation of 12W would be really impossible. Lets hope I'll never come to that. Can that amp draw 3A anyway?

Since the Zhaolu is now open I might as well take a few temperature readings... I've always known its power supply to be inefficient, lets take more evidence of it.

Power draw should have a somewhat linear relationship with temperature rise. In practice, the relationship between rate of heat transfer and temperature may not be linear, but the error is reduced by having the heatsink in the open with the fan blowing onto it - this minimizes the issue of different surround air temperature screwing the calculation.

Power supply power off - air temperature

Power supply on, amp off

Power supply on, amp on

18 vs 25 degree rise.

Output voltage = 15V, input voltage = 14V x √2 = 19.8V, voltage drop = ~5V, current draw = 0.25A, additional power dissipated = 1.25W, additional temperature rise = 7 degrees, thermal resistance = ~5.6ºC/W

Sounds logical. It is actually way lower than what it should be, but I had a fan blowing at it.

Using that thermal resistance, power used by regulator doing nothing = ~3.2W

Sounds logical.

I did remember this bugger to use 7 to 8W idling. 3.2W is for one LM317 at 20V in/15V out. There is also the negative -15V guy and a +5V guy too, so roughly would add up to around there alright, assuming power draw by the DAC + headphone amp circuit to be negligible, which it should be.

Hmm... this does not look good

2011-02-27T21:32:00.002+08:00

Hmm what does that number mean?

I paid a few bucks extra for this multimeter. For the thermometer function. It has an accuracy of ±5 degrees but that is enough for me - I just want a general idea of the kind of temperature the thing is running at.

And that is in degree Celsius.

Well, actually I was using the linear power supply of the Zhaolu D1.3 to power an amp.

A linear power supply of a DAC powering an amp, so, ya.

"Overclocking" anyone?

Actually, not really. I measured my Yulong T-Amp's power draw to be 4W, and that thing also uses a linear power supply.

Remember I measured the D2.5A to draw 10W.

Nonetheless, it is not good to create something without going through flame-test 3 - long term stability.

How I measured - probe inside closed chassis.

92 degrees is not good, but it's not too bad either, at least it is stable instead of increasing all the way to ∞.

It's just a bit close to the maximum operating temperature of 125°C of the LM317. But that is junction temperature - I'm measuring that of the heatsink. Include junction-to-case thermal resistance, and don't forget contact resistance, consider inaccuracy, and it is darn close to max temperature.

Fortunately LM317 has overheat shutdown. Hooray for integrated circuits.

But I don't like stuff running at near boiling point of water anyway. Maybe 80°C is the max I would go.

So, is this thing so hot because I'm "overclocking" it?

Well it's too unscientific to jump to conclusion at this point. Need to measure temperature of regulator with no load connected.

I got... 80°C.

Screw this design.

RCA couplers have arrived

2011-02-27T12:11:00.007+08:00

The last part of the in-line attenuator has arrived.

This thing converts the female connector on the attenuator to a male to save using one extra cable, which has benefits cost and performance wise.

Why didn't I use a male connector on the attenuator in the first place and instead used two female connectors instead?, you may ask.

Well, the dimensions of the front and back are 15 x 13mm. It's actually rather compact for a DIY-ed solution. And it has to be in order to have the most compatibility. These dimensions are pretty tight for a male RCA connector. I did try to use a male connector, but the moment I on-ed the 1cm drill it cracked through the whole thing.

Even for the socket I had to look for one with a really thin body, so that the nut can fit inside the enclosure.

How the finished thing looks.

The thing in action.

So far I've tried it on two systems - the one that you see in the pic, and another one which I shall not disclose now. Lets call it system X for now.

Why am I mentioning this? Well, there are unexpected side findings. Apart from the obvious 20dB reduction in volume.

On system X, it solved a bit of balance issues (due to volume control at too low positions) as expected, at the same time the soundstage opened out/the sound became wide/sue me my audiophile language sucks.

I attributed this to the balance issue being solved - when more sound is coming from one speaker, the image feels less spacious. And/or maybe having two speakers at the same volume brings out mids and some bass better.

On the system you see in the pic, the bass became more controlled and more powerful. Could be due to the same thing as above.

*Note - none of these claims have been blind-tested.

If avoiding channel imbalance issues results in these benefits, there could be some weight in why stepped attenuators sound better than potentiometers (notice the commonalities between claimed improvements of using stepped attenuators vs my thing).

However, again, this is untested, and those who calibrate their speakers to +-0.5dB would know the truth.

RMAA results of 16-bit/44.1kHz MP3 at various compressions

2011-02-24T18:03:00.003+08:00

http://wwenze.netfirms.com/mp3.htm

I'm not going to do any explaining... too tired... sucks if you don't know how to read the graphs, go learn.

Incorrect Assumptions

2011-02-20T07:32:00.000+08:00

Music Core

2011-02-16T09:19:00.000+08:00

- A musical genre that includes everytype of music. This is usually a sarcastic response given when some lame scenester asks what type of music you listen to after listing 20+ genres that may or maynot exist.

Scenester: Yeah so I listen to gothcore, disco ska, plip hop, darkwave, lightwave, gabber, and science rock. So what kind of music are you into?

Source: Urban Dictionary: music core

What I found after trying to find out what exactly does "Slash Core" mean.

http://en.wikipedia.org/wiki/Music_of_DJ_Max#DJ_Max_Portable_3

Big-Oh

2011-02-11T08:31:00.002+08:00

Ok sorry about that bye.

That's what he said

2011-02-10T06:56:00.006+08:00

"If u supply 150 ohm to the the headphone, other then lack of bass, the mid and high freq is clearly define. in fact it is one of the best gaming headset, if u try to drive with a ipod, well, it's crap."

Me: What happens if I supply 10kohm or 2 ohm to the headphone?

"for 10 kohm, the headphone will be gone.

for 2 ohm, if you are lucky, you can still hear some sounds with the volume at full.

Never use your speaker output to connect to yr headphone."

ragnarok95: Nice question to test someone.

"in fact a lot of people are not aware of it, got hole just plug.

Esp those high end headphone (300 and above), you need to match the ohm in order for it to performance.

From some of the reviews complaining PC350 is cmi, you can see that they just use 64 ohm to drive the headset (which need 150 ohm)."

Me: Okay final question, pardon me noob, you said can see others just use 64 ohm to drive the headset, while you know 150 ohm is good for the PC350, but how do you know how many ohm your soundcard is supplying to the headphone?

"u need a dac or a sound card come with a headphone amp to drive high ohm headphone.

for dac, there is a switch for u to adjust the gain, sound card is control by software."

He didn't answer my question, but that's enough from me.

Source:
http://forums.hardwarezone.com.sg/showthread.php?t=3097327&page=3

[Rant] "WTS: HDMI cable"

2011-02-07T11:10:00.009+08:00

Selling brand new premium HDMI cables @ $10 each!

Comes in a classy blue sleeve to match your decor and for better shielding

Ok, seriously, I got it for SGD$4 from eBay, and one without a sleeve would be even cheaper. And my philosophy is, if I can get something, you too can.

So why the f should you let me earn extra $6 for doing nothing?

I'm not going to do anything about what happens in the forums though; there is a rule for documentarists, that an outsider is not supposed to interfere with the course of nature - if you spot a wolf eating a rabbit, you cannot go save the rabbit, even if you like the rabbit. (Exceptions, like endangered species, occur.) If people are stupid enough to pay 200% for the same thing, that is their problem. I believe stupidity is punishable.

Well, I hope none of you reading this blog have been suckered to pay more for less, and if you didn't know about this kind of get-rich-quick scheme (by others) before, well you are lucky.

And if you still don't believe that you can get this for $4 on eBay, and if you want to buy this HDMI cable @ $10, well, we can discuss.

And, I'm not joking about the better shielding part -

Braided for better shielding

[Anime] "Best of the season/year"

2011-01-28T07:15:00.005+08:00

I've went through a few of the previous season(ending in Dec 2010)'s offerings.

Some of them were commented to be "best (genre) of the season/year".

But frankly, I am not amused.

Because so far, there was nothing spectacular. In fact quite a number of them fall into the "only watch when you've watched everything else and are still bored" category. Comedies with low-level humor, love stories with no feelings. There were some that were average, but shouldn't get the title of "best".

But, if you compare them against the pile-o'-junk created in that season, well, that puts things in perspective.

I was wondering about the lack of high ratings and attributed it to uncertain AniDB rating system, now I know the real reason.

The problem is some of the reviewers. I also find myself a possible victim of this - habituation. Being exposed to bad shows after bad shows, eventually one finds even the mediocre good.

It also hurts my radar that detects the quality of the show. Ironic that I have to expose myself to crap and possibly degrade my radar in order to prove its accuracy, but so far it has still been accurate - things that I thought would suck still suck.

Random thought/quote:

2011-01-27T11:35:00.001+08:00

"I also think inside the box. But my box is bigger."