Why Are Computer Soundcards So Crap?!?!?!?!

[soundcheck]

Hi there.

I also think the Asus card family is quite a good choice. You can easily feed them externally with a quality supply. (People are reporting great results )

If you go for the HDAV Deluxe you get on top several channels for activating systems,
you can tap off I2S or you can easily swap the OpAmps on the output.
Of course you can easily pullout the OpAmps and put up a LC-Audio Zapfilter .
This would get you even a balanced output option at reference level.

You can even fire your multichannel amp with linear multichannel pcm via HDMI.

What else do you want.

Cheers

 

[Segran]

This is a great card but the Auzentech forte is a card in the same level.

Close, but not quite up to par. The dynamic range is 15dB better on STX, distortion a decade better (0.0003% compared to 0.0019%).

I compared STX to Auzentec and found noise level to be noticeable better with STX. Distortion at those low levels I can't hear.

I also compared ESI Juli@, which IMO is equal to Auzentec in noise/DNR and have lower distortion. Juli@ also offers balanced I/O and lower price than STX.

I use the STX for head phone listening and, more importantly, to measure my DIY amps. The high dynamic range/low noise is very important when measuring distortion levels >0.001%!

 

[JZatopa]

Where did you get your numbers? The stx does not have 15db better dynamic range then the forte as tested on rightmark. At 16/48 they are almost exactly the same.

http://www.elitebastards.com/cms/in...sk=view&id=696&Itemid=27&limit=1&limitstart=4

In fact at 24/96 the forte tested to have about 30db (I say about because the STX is far from smooth in the measurement) better dynamic range then the STX. The forte also has about 30db better noise floor. The forte seems to beat it in almost all tests at this point. What surprises me is how big the difference in cross talk is, there is no way that can't have a large impact on the sound stage and imaging.

*edit*
If you are basing your info off of the test here I can now see where you got your numbers.

http://www.extremetech.com/article2/0,2845,2339949,00.asp

My problem is that if you look at each test on this site the low end on three different unrelated tests have the bottom end tilting up starting at 200hz. I think there was something wrong with the cable or setup they used because I don't see that on the elitebastards site or here.

http://www.guru3d.com/article/auzentech-xfi-forte-71-soundcard-review-test/9

 

[gabrielbecheanu]

Because we talk about PC sound cards I have a question for you guys:
I own an Asus M2N32 SLI Deluxe WiFi edition with optical and digital coaxial out.
Who knows how to unlock the digital out from 48kh to all frequency? The sound chip is AD1988B. In the specification of the AD sound chip tell us that this chip allow to exit via digital all the frequency from 32khz to 192khz.
If is not possible I subscribe: All the onboard soundcard are crap

 

[Segran]

Where did you get your numbers?

The Auzentec figures come from their web site, STX figures from the Audio Precision Test Report that comes with the STX (that must be a "one-off", or is anyone else shipping AP test reports with their boards?).

Might not be quite fair to Auzentec and ESI to compare only "paper numbers", but remember that I did test both Auzentec and Juli@ in the same setup as the STX.

I have seen several tests by various magazines using RMA, they all come out with different numbers.

I don't trust RMA test reports to give more than a hint, there are too many uncertainties related to test environment that are cause for concern. Loop back cable being one, the SMPS and MB a couple of others. By simply changing loop back cables I can get anything from 98dB to 124.9dB SNR from my STX! The level settings of input and output also have great impact on the test results.

Would be interesting to see a comprehensive test under equal and controlled conditions on all sound cards with a claim to fame (audiophile, that is..).

I'm sure that both Auzentec and ESi Juli@ are great sound cards in tehir price bracket, I just have found STX to suit my needs (measurements) the best

 

[Segran]

In fact at 24/96 the forte tested to have about 30db

Where did you get these numbers?

I didn't look at Extremetech, but I have checked Guru3D now.
Guru3D have tested both boards and I guess that's as close as a comparative test we can come for now. I at least guess that they are using RMA with some insight and knowledge.

Check the attached file and let me know what you think about the test results

The forte also has about 30db better noise floor

Same thing, where did you get those figures? Again, Guru3D are showing much different figures. In fact, Asus comes out better - not much, but better...

three different unrelated tests have the bottom end tilting up starting at 200hz.

Sure, that's what I said about the test environment and you skill to set up the test object correctly! But in all honesty, the Auzentec is noticeably more noisy below 100Hz. I think this is where the bigger EMI shield and more ellaborate filtering is giving the STX the advantage.

What surprises me is how big the difference in cross talk is

I don't see what you mean, the Guru3D test are showing better (lower) cross talk figures for STX. I agree that STX shows a strange behaviour in 24-bit/192kHz mode. I don't see that on my STX when i run RMA.

All in all, I don't see that any of these figures would render one or the other useless or "crap" as this thread suggests. The relatively small differences in Measured performance is not likely to be audible.

Auzentec uses AKM 4396 DAC (for the front channels which is the only I'm interested in) and Asus uses BB PCM1792A. Both are "audiophile" quality in my mind although Auzentec give slightly more distortion. The differerence in distortion is NOT aidible, but important when using the board as a mesuring device.

Auzentec use Creative DSP and I feel strongly that all things labeled Creative are "crap". Auzentec have other boards, using C Media DSP's which are far better. Asus actually bought all rights to C Media's top of the range DSP, did some "improvements" and re-labeled it AV100 and AV200.

The difference are small and IMO not audible. Well, the noise level below 100Hz IS audible if you are using headphones.

Let's agree that both boards are good in their price bracket! And so is ESI Juli@. I don't know about M-Audio, they seem to be pretty good as well.

 

[dzzmiller]

I think Sergen is avoiding the Asus Xonar, which in analog output rates higher than the Auzen in direct tests.

Perhaps the meaning that can be taken from this thread is not to assume that medium quality D/A aimed at the professional recording market is better than the best of the multimedia cards. Especially for people looking for more than two outputs. It might be worth purchasing both types and returning the "loser".
I can also say that in my experience at least the Xonar is superior to any less expensive (<$350?) USB DAC.

 

[Segran]

I think Sergen is avoiding the Asus Xonar

Avoiding???!!! I LOVE Asus Xonar and am only trying to let people know that there is at least one sound card that's OK!!

It might be worth purchasing both types and returning the "loser".

I did! And also ESI Juli@...

But yes you are right, I am ONLY interested in two channels, and mainly for measurements and headphone listening. Otherwise I listen to my (DIY) stereo system. I don't do gaming either, and for 5.1/7.1 I use my Yamaha system.

That's why I've been going on about "audiophile", not multimedia!

If I needed 7.2 I would choose Asus Xonar D2PM, for gamin probably Asus Xonar D1. Or Auzentech - who knows!

But I don't get this:

medium quality D/A aimed at the professional recording market

My impression is that Xonar Essence STX was aimed for audio freaks and media PC's, not prefessional recording market?

 

[JZatopa]

I go those number in the graph here from the elite bastards site.

http://www.elitebastards.com/hanners/asus/xonar-stx/charts/rmaa/noise96.png

You are right a lot of tests seem to have different numbers. The differences at these levels are most likely on the subtle level and not glaring. Once you add in the mods that most of us will do (upgrading opamps, upgrade or remove dc blocking caps, upgrade clock, external PS) I am sure the differences will become almost nonexistent. I do believe that at that point there are very few audio sources that can compete with the level of sound quality with any traditional source.

 

[Segran]

I go those number in the graph here from the elite bastards site.

The Elite Bastards test just proves my point - RMA tests can be totally screwed up!

If you look at the results for Auzentech it shows a noise level (24bit/96kHz) around -170dB. That is not ver subtle and simply not realistic! Compare the noise level to the distortion graphs taken at 16bit/48kHz and you will see that Auzentech noise level is around -120dB, which is a very good result.

Looking at Asus in the 24bit/96kHz graph, the level is all over the place - and even worse than Xonar HDAV! There is a serious flaw in this measurement. In fact all of the 24bit/96kHz measurements are way off. It could be driver problem, cable problem or more likely - operator error!

I have had similar test reults using RMA my self - but that was because of heavy distortion (clipping) as a result of fried input stage opamps.

BTW, I have changed my mind regarding what sound card I would buy if I wanted 7.1 - the NEW Xonar Essence ST with the H6 DAC expansion board plugged into it!
http://www.guru3d.com/article/asus-xonar-essence-st-sneak-preview--test/1

So that's why the STX is obsolete...

 

[JZatopa]

Segran seems like we are both on the same page even if we are fans of different products. Personnally, compared to most on this board, I think we are ahead of the curve on this subject. I was wondering if you are planning on doing any mods and if you could tell me a bit more about the computer and sound system you have your sonar setup in. Did you ever look into any pro audio solutions such as the Hammerfall or any products from Apogee?

 

[Segran]

Hi DJNUBZ,

Did you ever look into any pro audio

I have never looked at anything near the Hammerfall or Apogee products, I only need simple 2-channel stereo for my measurements!

The only thing I'm planning to modify on the STX is the power supply for the analog section, as I have a -120dB peak at 50Hz (followed by low harmonics at 150 and 300Hz) while the rest of the noise level is at -128dB. I havent' really investigated the root cause yet. It could be that I just need better cabling!

The PC is nothing fancy, a Core2Duo Intel in a ASUS P5QDL motherboard and a vanilla flavour 450W SMPS.

And yes - the sound card function on the motherboard (Analog Devices) is really crap! I have a old Compaq N400 laptop as well, the sound card and that is really terrible. The ThinkPad I have is kind of OK for making MS Windows noices, but that's about it.

 

[fotios]

Asus Xonar Essence STX FFT analysis

Hi to all

I have read all posts, and i agree with the estimation of Segran for the essence stx. I changed my EMU0404 USB, with the essence stx because its lower output noise. Xonar Essence STX is powered via the traditional 4pole Molex connector which offers +12V, 0V, 0V, +5V. This is the standard connection inside PCs for peripherals. The convertion to -12V is made from the +12V supply rail by an on board PWM positive to negative MPS - MP9141ES converter. A very good thing, it is that yet the +12V supply rail it pass thru a 7812 voltage regulator first, which ensures a very good PSRR.
Usually, from the FFT software providers an external USB sound card is suggested. This is not a general rule. I ascertained this with a DSO, by measuring in the output of EMU0404 USB a noise floor of 12mVrms, instead in the output of Essence STX 3,6mVrms. I don't know if this is applicable only in my PC which is builded as a workstation. Its power supply is a Zalman ZM750-HP, which offers 3 seperate sockets for power distributing. I use one of them exclusivelly for essence STX. Another one important thing, it is the shielding quality of VGA card which is very close to sound card. MY Sapphire X1950GT ATI, it is covered at whole from an enormous heatsink. I attach a picture of inside of my PC:

An externally hosted image should be here but it was not working when we last tested it.

In this picture is presented the noise floor of essence stx

An externally hosted image should be here but it was not working when we last tested it.

The rest hardware stuff included in PC is:
M/B: Intel DQ35JO 1333MHz FSB
DRAM: 8GB / 800MHz synchronous
CPU: Intel C2D 3GHz
The operating system it is WinXP x64.
As for the Xonar Essence STX, it offers the top ADC converter of Cirrus CS5381. For the rest, you are informed allready. PCM1792 it is the second better DAC after the ESS Sabre.
The only drawback it is the sensitive Line In of essence stx. I have destroy one essence stx during the calibration procedure with my FFT analyzer. Enough people have the same problem. Fortunatelly the card was covered from the guarantee during the damage, and i got a new as replacement. The second time i was very cautious in the input level settings, so as it does not exceed the +16dB in the patch mix spectrum analyzer.
I had a significant reason to buy the essence stx. This was the offerd test report from an Audio Precision SYS-2722 FFT analyzer. I had never a better opportunity to estimate my Virtins M.I. Pro 3.1 FFT analyzer - oscilloscope. This analyzer, it is a proffesional tool compared to RMAA and others. If it is calibrated correctly according to I/O level of sound card, then its results are very reliable. Especially its internal signal generator, it is quite sophisticated, and it is individual from the Windows mixer or the patch mix of sound card.
In the following pictures is presented the performance of essence stx in hardwired loopback test via external cables:

An externally hosted image should be here but it was not working when we last tested it.

An externally hosted image should be here but it was not working when we last tested it.

You can download the test report of essence stx via the Audio Precision from here:

http://dlcdnet.asus.com/pub/ASUS/Audio_Card/Xonar_Essence_STX/e4253_Xonar_Essence_STX_Audio_Test_Report.pdf

We should take into account, that the plots obtained from Virtins MI includes simultaneously the Input and Output performance of essence stx, instead the plots obtained from Audio Precision are seperate from Input and Output. For this reason the results obtained from Virtins MI are a little worse from Audio Precision.
I have tried the analysis as well with the Spectra Plus, but the results obtained was better from those of Audio Precision!!! Consequently, i rejected these.
Summarizing, i think that we are very close (at least!) now to have a reliable audio measuring system of low cost (about 350 euro).
I am in waiting of your comments.

Fotios

 

[fotios]

REAL STUFF

Here is presented a FFT analysis of a real device with the Virtins MI Pro 3.1 FFT analyzer using as ADC-DAC interface the ASUS Xonar Essence STX. The DUT it is a diy microphone preamplifier designed by me for using with condenser or dynamic microphones. The settings of MI analyzer are the same as in the previous plots.
Oh my! Wonderfull and trustable results at least! Look please:

An externally hosted image should be here but it was not working when we last tested it.

An externally hosted image should be here but it was not working when we last tested it.

This is a picture of the device under test:

An externally hosted image should be here but it was not working when we last tested it.

Fotios

 

[fotios]

FFT size is changed at 16384

To ensure you that the results of measurement are trustable, i did another one analysis with FFT = 16384. The results are almost the same, except that the plot it is not so clear because the lot of FFT points. I think that the FFT size of 2048 points, it is the better and i think as well, that the expensive Audio Precision it uses the same size! Here is the plot:

An externally hosted image should be here but it was not working when we last tested it.

Don't waste your life in repeated loopback tests of your sound card to examine if it is trustable. It is! As you can see, the device under test it removes by alone the most of the harmonics produced from the signal generator driving the DAC of sound card. Only two remains, one at 2KHz and one at 13KHz. Although the plot is presented most smooth from the loopback test, the numbers are worse and says the TRUTH!

Fotios

 

[duekfx]

Many will disagree with me when I say these Asus cards are just hype. Don't misunderstand me, I love their motherboards, but when it comes to soundcards they give you a whole lot of sh*t you definitely don't need. Noise... those "audiophiles" mentioned on the first page are absolutely correct about their theory. You can of course waste loads of money to buy ultra-hyper-super low noise stuff etc. but that's the kind of logic like touching your left ear with your right hand going around your head instead of just using your left hand if you know what I mean. I've had a couple of soundcards, now I stick to a Creative X-Fi Xtreme Music card which I was probably the first to buy in my country when it came out. But it's nothing special in terms of sound quality, it's a tiny little better than my onboard Realtek ALC888, and I guarantee any ASUS card won't do any better either. The reason I like it is, because it has realy great features in terms of positioning, gaming stuff (EAX 5 and alike), a realy great control panel, and it's driver is finaly bug free Comparing it to some Asus card: I would loose all the great features in exchange for something I can easily achieve with an external DAC or I don't need it at all.

I got bored of reading all those "great" and "fantastic" reviews about soundcards. The difference in sound (and I mean the "above average" category) is negligible. Our mind can easily hear better measurements as a better sound reproduction, but is it realy better or is it just our mind playing with us not realizing that hearing any difference is beyond our limits. These limits might vary from man to man. Some have better ears, some don't. A good example would be the height of a room. An average ceiling is as high as approx. 3 meters. Would it make any sense to make it 10 meters if there is no human being taller than lets say 2.5 meters? No, of course not, but you are free to waste your money.

So my advise for all those who don't want to suffer as much as I did (before I enlightened) with choosing a proper sound card is: stick to the one that suites your needs in terms of features and use an external DAC, so you will have all you REALY need (and spare yourself from a bunch of useless junk just to make the card more expensive).
i.e. you are a gamer and you are addicted to sound quality, buy a gaming card and a DAC. It's just a pain in the a** to buy an EMU1212M and see all those "not supported" messages and grayed out options in games.

Some of you might share my view, some of you won't. It's my opinion, and everyone is free to ignore it

Edit: an external sound card is also a very good option if it's not for recording purposes.

 

[Segran]

duekfx,

you are totally missing the point, at least the one I made - I use the soundcard for measurements on hifi amplifiers I design and build! It is extremely important to use a soundcard that adds as little as possible to the measurement, both noise and distortion.


As you are obviously using the soundcard mainly for gaming I might agree with you that Creative Xi-Fi is good enough.

For measurements, as well as for any serious listening purposes I still regard Creative products completely useless.

 

[duekfx]

Gaming was just an example (since most soundcards are being designed for gaming and home theater purposes). I'm not a gamer at all. As for measuring purposes I don't think the engineers of Mark Levinson's would ever use a 500$ (or even the best available) soundcard to do the job. You'll probably need some realy precise instruments or stuff like NASA has for good results.

Edit: Creative cards (despite their annoying 48kHz resampling and buggy drivers in early stages) are quite good cards in total. I've had quite a few soundcards and heard/saw even an RME HDSP 9632 in action. Differences are quite negligible since there is no atomically card available. Every product has it's own weaknesses. You can go to a shop and say: hey, I'd like a pro soundcard for measuring purposes, and if the vendor knows his job, he'd probably say: "any or none"

 

[fotios]

Gaming was just an example (since most soundcards are being designed for gaming and home theater purposes). I'm not a gamer at all. As for measuring purposes I don't think the engineers of Mark Levinson's would ever use a 500$ (or even the best available) soundcard to do the job. You'll probably need some realy precise instruments or stuff like NASA has for good results.

Edit: Creative cards (despite their annoying 48kHz resampling and buggy drivers in early stages) are quite good cards in total. I've had quite a few soundcards and heard/saw even an RME HDSP 9632 in action. Differences are quite negligible since there is no atomically card available. Every product has it's own weaknesses. You can go to a shop and say: hey, I'd like a pro soundcard for measuring purposes, and if the vendor knows his job, he'd probably say: "any or none"

You do or you do not believe that, there is many people in this forum which is using sound cards for measurements. Likewise, there are enough proffesional stand alone FFT analyzers of 6000$, in which is used the same ADC converter CS5381 like in Asus Xonar. If you take a look in the application notes of CS5381 in Cirrus Logic site, you can find that if two CS5381 are combined can give a better performance by 3 to 6dB. This conjuction is used inside pro FFT analyzers. In Asus Xonar, simply is used a single ADC. Creative, does not paid so much attention in output noise (yet in its EMU products) because is addressed mainly in musicians or gamers. Asus xonar essence stx is designed mainly for audiophiles, for this reason its superb performance compared to EMU1212M which is the closest project of Creative to Asus xonar. If i had the possibility to increase the supply level of analog section of essence stx from +/-12V to +/-18V, then i believe that this sound card could obtain an unbelievable performance.
Segran and me, we are builders of audio devices. For him i did the labour to post the above reports. Not for gamers or MP3 class audiophiles. For building an audio device, the traditional instruments like the DSO and the function generator are the most indispensable devices. The FFT analyzer it is for examining the details, and to give a test report for consumer use. For audio designers (not for reviewers) the first tools are the DSO and the Function Generator and after those a FFT analyzer. I can confirm you that my DSO and my function generator together have a cost of 3800 euros. With the same ammount, you can buy a low cost stand alone FFT analyzer, new, like those offered from Sound Technology, or a used Audio Precision from e-bay. No, thanks! I preffer my trusty instruments, and for FFT analysis an economic solution. You will see in the near future (acording to fast improvement of PC hardware) that sound cards will be competitive with the expensive FFT analyzers. Big role in this will play as well the improvement of sound card based FFT analyzer softwares. A good example it is this of Virtins, which already supports as well 2 - 3 ADC-DAC interfaces of National Instruments except sound cards.

Fotios

 

[Segran]

dukefx wrote:

I don't think the engineers of Mark Levinson's would ever use a 500$ (or even the best available) soundcard

Neither do I, but I don't have the financial means to pay for an Audio Precision analyzer since I am engaged in audio design purely for my own pleasure and knowledge gain. I don't sell my amps at outrageous prices either... I have to make do with what I can afford rather than dream unrealistic dreams.

I'd like to thank fotios for your ellaborate work. I think I'll buy Virtins now.

Have you made a input protection (selectable levels) for your soundcard? I have made a few attempts but I'm not satisfied with any of them.