Skip to main content ↑ ↓

Notice

Please note that most of the software linked on this forum is likely to be safe to use. If you are unsure, feel free to ask in the relevant topics, or send a private message to an administrator or moderator. To help curb the problems of false positives, or in the event that you do find actual malware, you can contribute through the article linked here.
Topic: PCI soundcard w audiophile Analog stereo output (Read 33659 times) previous topic - next topic
0 Members and 1 Guest are viewing this topic.

PCI soundcard w audiophile Analog stereo output

Reply #50 –
Quote from: Dawnrazor-age on
< Nelson Pass's commitment to single ended MOSFET designs is strange, but his amps end up sounding very good - not necessarily transparent, but nice.>
Pass doesn't only make SE designs, the one I have is the X series.  What Double blind test are you refering to that concludes his amps are audibly not transparent?
I have a headphone amp using one of Pass's Zen circuits which is obviously not transparent. I haven't ABXed it (I would need recording equipment better than what I have) but it's measured THD is around 5% - which is lots. It sounds great though.

My point was more than many of Pass's designs measure poorly (and, based on those measurements, one would expect a certain ABX result) but sound nice. I don't use that headphone amp any more, but could do some testing if anybody was interested.

Quote from: Dawnrazor-age on
I made some 192 vbr mp3s using lame and have done a comparison with the original .wavs.  I scored 14/15 in the Foobar ABX test you recommended.  That is with crappy onboard sound into some decent head phones.  It was fairly close, but once you hear the differences, it is easy to figure it out with consistency. The bass was not as full and rich on the MP3s, and there was a bit more slam on the original.

I am assuming you want me to post the results and the files.  Where should I do that, in the file upload section?
Uploading the files is uneccesary, but a quick note of what music it was and your ABX report is generally appreciated. The important thing is that you did an ABX test.
Quote from: Dawnrazor-age on
I was able to easily ABX files at different rates 44.1 vs. 96khz.  maybe it was the distortion you are talking about, but they did sound different.

I'm still busy reading that article, but for most hardware upsampling a signal to a higher rate before the DAC (not counting oversampling DACs, which do it for different reasons) will tend to be worse, rather than better in quality. Maybe you could try something like ABC/HR to rank the results.

Quote from: Dawnrazor-age on
Also, I remember a gentleman who had a dac that lit up when HDCD files were played, and when the Kmixer was involved, the dac would not light up, but when using ASIO, it lit up.  HIs conclusion was that the bits were not identical through the kmixer.
Right, kmixer is not bit perfect in many situations - I think it's driver dependent but I am not sure, and don't use windows so I can't test it. If you want bit perfect playback then bypassing kmixer is perfectly reasonable.
Quote from: Dawnrazor-age on
Thanks again for getting me to do the ABX test.  It really solidified what I thought I knew...and that was on just average equipment!  I can only image how much easier it would have been on my main rig where imaging could be part of the test.
If you succeeded with this test, please try some more. Tests with the recommended LAME settings would be a good start. Good luck with future tests.



From http://www.positive-feedback.com/Issue22/nugent.htm
Quote
Upsampling is one of the BIG advantages of using a computer to drive your music. Digital music is recorded at a particular sample-rate, generally 44.1 kilosamples per second on CD's. Upsampling or resampling adds more samples that were not part of the original recording process, but try to approximate the samples that would have been recorded had the music been recorded at the higher sample rate initially. The added samples are computed with various mathematical algorithms that examine the music waveform prior to and after the time-slots where the new samples will be inserted. These new samples not only add more detail to the music, but improve the dynamics as well.
This isn't an accurate representation of what is going on with resampling. I suspect Steve Nugent knows that, but is writing for a non-technical audience.

Ok, in general, resampling adds no more information into a signal. The new samples that are inserted are inserted in positions predicted by interpolation algorithms which add nothing to the signal - the original sampling rate sets the bandwith and the original bit depth sets the SNR. Adding more samples to the signal with some algorithm that is not an oracle will add one of two things:
  • Nothing, or
  • Distortion

Of course when you take digital to analog conversion into account the picture is more complex than that, but in general resampling is just a waste of cycles (if your DAC supports the orginal source rate).

PCI soundcard w audiophile Analog stereo output

Reply #51 –
Quote from: cabbagerat on
Not really, no. Most oversampling DACs don't have a "native frequency" but can sync to multiple frequencies, albeit with different amounts of oversampling. The performance of a DAC might change with sample rate but in most cases I have seen performance is better (within band) for 48kHz than 96kHz anyways.

That's interesting. What do you mean by better performance at 48khz than 96khz? Why would a sound card implement 96khz if it had technical problems that made it perform even worse than at 96khz? (Just for show?) This can't surely apply to top notch sound card dacs?

PCI soundcard w audiophile Analog stereo output

Reply #52 –
Quote from: cabbagerat on
Ok, in general, resampling adds no more information into a signal. The new samples that are inserted are inserted in positions predicted by interpolation algorithms which add nothing to the signal - the original sampling rate sets the bandwith and the original bit depth sets the SNR. Adding more samples to the signal with some algorithm that is not an oracle will add one of two things:

Is it true that if you consider the digital signal as measurements at points in time that if you resample to a higher sampling rate the samples corresponding to the same times are the same?
That would be true if the analog-digital process measured voltages at those time points but isn't it a more complex function?
In any case interpolation might be quite complex, mightn't it? Certainly not just smoothing (thinking about pure tones close to the Nyquist frequency shows it can't be smoothing). Can DACs (without upsampling) be designed to get this interpolation just right?

PCI soundcard w audiophile Analog stereo output

Reply #53 –
----Uploading the files is uneccesary, but a quick note of what music it was and your ABX report is generally appreciated. The important thing is that you did an ABX test.---

Ok, the file was a song from the Fields of the Nephilim named "Love Under Will"...not exactly mainstream.

Anyhow, here is the foobar report:

foo_abx 1.3 report
foobar2000 v0.9.3.1
2006/10/13 17:49:17

File A: H:\Nephilim\Love Under Will.wav
File B: C:\Documents and Settings\Dawnrazor\My Documents\test mp3\Love Under Will.mp3

17:49:17 : Test started.
17:58:14 : 01/01  50.0%
17:58:30 : 02/02  25.0%
17:58:48 : 03/03  12.5%
17:59:16 : 04/04  6.3%
18:00:06 : 05/05  3.1%
18:00:18 : 06/06  1.6%
18:00:33 : 07/07  0.8%
18:00:55 : 08/08  0.4%
18:01:12 : 08/09  2.0%
18:02:03 : 09/10  1.1%
18:02:22 : 10/11  0.6%
18:02:37 : 11/12  0.3%
18:03:13 : 12/13  0.2%
18:03:28 : 13/14  0.1%
18:03:44 : 14/15  0.0%
18:04:00 : Test finished.

----------
Total: 14/15 (0.0%)

I had done a test before with an engineer friend who works for a large audio company.  He had some files on his laptop and played them balanced through a "pro" headphone amp.  He was really surprised when I could easily identify the mp3 files.  This was his field of expertise, and he seemed concerned.


---If you succeeded with this test, please try some more. Tests with the recommended LAME settings would be a good start. Good luck with future tests.---

What are those settings?

Perhaps Steve is oversimplifying upsampling for the audience, but I can say that he does at least believe himself that upsampling is important.  SO much so that he is till using the old 8.3 version of Foobar since his pet upsampler (secret rabbit code) doesn't work on 09.

Yet there are some that totally agree with what you are saying.

PCI soundcard w audiophile Analog stereo output

Reply #54 –
Dawnrazor - which release of lame, and what command line settings (or corresponding program settings) were used?

The HA recommended settings can be found here:  http://wiki.hydrogenaudio.org/index.php?title=LAME

EDIT: answered "What are those settings?" test above.

-brendan

PCI soundcard w audiophile Analog stereo output

Reply #55 –
Quote from: CSMR on
That's interesting. What do you mean by better performance at 48khz than 96khz? Why would a sound card implement 96khz if it had technical problems that made it perform even worse than at 96khz? (Just for show?) This can't surely apply to top notch sound card dacs?
What I meant was that many DACs will perform no better on 48kHz material resampled to 96kHz than on native 48kHz material. Obviously material that is recorded at 96kHz will offer twice the bandwidth - and effect that may or may not be audible. What I mean by better performance is common DAC measurements such as passband ripple, stopband ripple and THD+N.

For a case study, consider the PCM1702A, a fairly widely used chip from TI/BB supporting sample rates up to 192kHz at 24 bit. On 44.1kHz native material with 384 times resampling, the SNR peaks at 106dB and the THD+N sits at 0.002%. On 96kHz native material with 384 times resampling the SNR peaks at 104dB and the THD+N sits at 0.008%. Other measures, such as dynamic range, and channel seperation are similarly improved by lowering the sample rate. Before you say that this is a single isolated case - it's pretty typical of most DACs I have seen. The figures are from TI's data sheet, not my own measurements.

These changes are probably not audible, but a decrease in measured performance with increasing sample rates suggests that resampling material to a higher rate for playback is misguided. You are best off (in 90% of cases) feeding material to the DAC at it's native sample rate.
Quote from: CSMR on
Is it true that if you consider the digital signal as measurements at points in time that if you resample to a higher sampling rate the samples corresponding to the same times are the same?
That would be true if the analog-digital process measured voltages at those time points but isn't it a more complex function?
In any case interpolation might be quite complex, mightn't it? Certainly not just smoothing (thinking about pure tones close to the Nyquist frequency shows it can't be smoothing). Can DACs (without upsampling) be designed to get this interpolation just right?
No, DACs can't be designed to get this just right - it requires a brick wall lowpass filter. DACs generally cheat by using delta-sigma (or sigma-delta, depending on who's data sheets you read) and cutting off below 0.5Fs. For example the TI/BB PCM1720 cuts of at 0.454Fs.

You might improve the frequency cutoff (from 20kHz to 22kHz) by resampling 44.1 material to 96kHz before playback, but you won't be doing your THD+N, SNR or dynamic range any favours. Your DAC, of course, may vary.

PCI soundcard w audiophile Analog stereo output

Reply #56 –
Quote from: bhoar on
Dawnrazor - which release of lame, and what command line settings (or corresponding program settings) were used?

The HA recommended settings can be found here:  http://wiki.hydrogenaudio.org/index.php?title=LAME

EDIT: answered "What are those settings?" test above.

-brendan

HI B,

It was the most recent Lame version...I believe it 3.9.7 (i am posting on a different computer and am going by memory) or the one that I could have downloaded 2 days ago.

I used EAC to compress the existing .wavs, and selected 192 for the bit rate.  I remember VBR being better, so I picked that.  I'll have to check what the exact command lines were, but I don't remember using anything besides 192.  I am pretty sure I left what ever lines were there alone.  BUt I'll see if I can get them for you.

The link supplied doesn't really spell out exact settings.  What are they?  Would 192vbr not suffice for a valid test?

THis maybe a moot point in that the link also clearly says:

"However, 'archiving' music using a lossy format like MP3 is never recommended – no matter how transparent the resulting files might be. The alternative is to use Lossless formats like WavPack, FLAC etc. that allow true archiving bit for bit like on original CD. "

THis was essentially the point I was making to the original poster who by my reading was trying to create an archival quality front end for his hifi system.  Maybe I misunderstood what he was saying, but I definately don't think my reading was too far off.

PCI soundcard w audiophile Analog stereo output

Reply #57 –
-V 2 (alt preset standard) has been LAME's historical high quality baseline ever since presets were added to the encoder.
No one can be told what Ogg Vorbis is...you have to hear it for yourself
- Morpheus

PCI soundcard w audiophile Analog stereo output

Reply #58 –
Quote from: Xenno on
-V 2 (alt preset standard) has been LAME's historical high quality baseline ever since presets were added to the encoder.


I DID use 3.97 192 vbr...I believe that IS -V 2.

WOW.

PCI soundcard w audiophile Analog stereo output

Reply #59 –
Hi, I'm new here.
I'm searching for a good pci soundcard for my pc since I don't like the onboard soundcard.
I'd like to connect the soundcard to my amplifier with 2 diffusers AND, in the future, to an active subwoofer, since my amplifier doesn't offer this possibility.

What do you suggest?
I've seen these cards: Onkyo SE-150 - ESI Juli@ - ESI maya 44 and some others.

Thanks!
Francesco

PCI soundcard w audiophile Analog stereo output

Reply #60 –
EMU 1212m.
Break The Rules!!!

PCI soundcard w audiophile Analog stereo output

Reply #63 –
Quote from: Dawnrazor-age on
The link supplied doesn't really spell out exact settings.  What are they?  Would 192vbr not suffice for a valid test?

THis maybe a moot point in that the link also clearly says:

"However, 'archiving' music using a lossy format like MP3 is never recommended – no matter how transparent the resulting files might be. The alternative is to use Lossless formats like WavPack, FLAC etc. that allow true archiving bit for bit like on original CD. "

THis was essentially the point I was making to the original poster who by my reading was trying to create an archival quality front end for his hifi system.  Maybe I misunderstood what he was saying, but I definately don't think my reading was too far off.


I think you're misundestanind the recommendation.  MP3 is not recommended for archiving because an archive is meant to be the 'master source' for any conversion to other formats.  You would not want to convert MP3 to another *lossy* format (it would stand a good chance of audible degradation, even if the 'archive' MP3 sounded transparent compared to its original .wav), so MP3 is not a good archiving format.


Your performance on 192 VBR with a modern LAME compile is outstanding, though not without precedent...IIRC at least one person here has reported ABXing even 320 CBR from source.  ABXing 44.1 from 96 is even more impressive.  It could be you are way out at the right end of the bell curve in audio acuity, but claiming both together makes me wonder of something's  funky in your ripping/compression/listening test setups.

Perhaps someone here could whip up some mp3s and flacs and/or samples at 44.1 and 96 kHz, for you to download and ABX. That at least takes the recording/compression part out of the equation.  Then just make sure you aren't using replaygain or other differential DSP during ABX.

PCI soundcard w audiophile Analog stereo output

Reply #64 –
Quote from: Dawnrazor-age on
I was able to easily ABX files at different rates 44.1 vs. 96khz.  maybe it was the distortion you are talking about, but they did sound different.


ABXing 44.1 KHz from 96 KHz is not an easy test to set up for many reasons, nothing to do with ABXing wav vs. mp3. Please give us more information about the exact procedure you employed to carry out this test if you want to be taken seriously.

PCI soundcard w audiophile Analog stereo output

Reply #65 –
Quote from: KikeG on
Quote from: Dawnrazor-age on

I was able to easily ABX files at different rates 44.1 vs. 96khz.  maybe it was the distortion you are talking about, but they did sound different.


ABXing 44.1 KHz from 96 KHz is not an easy test to set up for many reasons, nothing to do with ABXing wav vs. mp3. Please give us more information about the exact procedure you employed to carry out this test if you want to be taken seriously.


Certainly.

I simply took the same .wav file and made a 96k version and abxd them.  The file was created with Dbpoweramp.  It really was night and day. 

I am not necessarily saying that what I heard was 100% due to the differences between 44.1 and 96khz.  See, the digital out on the Lynx card feeds a dac that upsamples to 192.  SO perhaps that is a major reason it was so clear.  Maybe not. 

I'll try to see if I can repeat the test without the external dac using the analog outs from the Lynx.  Before I go to that trouble, can you let me know if such a test would suffice, or is there some other hoop I need to go through?

PCI soundcard w audiophile Analog stereo output

Reply #66 –
Quote from: krabapple on
Your performance on 192 VBR with a modern LAME compile is outstanding, though not without precedent...IIRC at least one person here has reported ABXing even 320 CBR from source.  ABXing 44.1 from 96 is even more impressive.  It could be you are way out at the right end of the bell curve in audio acuity, but claiming both together makes me wonder of something's  funky in your ripping/compression/listening test setups.

Perhaps someone here could whip up some mp3s and flacs and/or samples at 44.1 and 96 kHz, for you to download and ABX. That at least takes the recording/compression part out of the equation.  Then just make sure you aren't using replaygain or other differential DSP during ABX.


I must tell you that the mp3 test was done on my non audio PC...just the onboard sound.  It won't support 96khz, so that test was done on the audio PC.  PC rig is not networked or connected

I think I do have decent hearing, and a fairly resolving system, but hardly a golden ear.  I do find that I can be sensitive to bass and sound stage though.

Anyhow, I would be happy to upload the actual files that I used for you to test ( I had offered this, but was told it wasn't necessary). Let me know and I'll do it.

An engineer friend also thought that I was crazy to think I could hear the difference between 192 mp3s, but I easily did when he had me listen to 2 different files.

PCI soundcard w audiophile Analog stereo output

Reply #67 –
Quote from: Dawnrazor-age on
Quote from: krabapple on


Your performance on 192 VBR with a modern LAME compile is outstanding, though not without precedent...IIRC at least one person here has reported ABXing even 320 CBR from source.  ABXing 44.1 from 96 is even more impressive.  It could be you are way out at the right end of the bell curve in audio acuity, but claiming both together makes me wonder of something's  funky in your ripping/compression/listening test setups.

Perhaps someone here could whip up some mp3s and flacs and/or samples at 44.1 and 96 kHz, for you to download and ABX. That at least takes the recording/compression part out of the equation.  Then just make sure you aren't using replaygain or other differential DSP during ABX.


I must tell you that the mp3 test was done on my non audio PC...just the onboard sound.  It won't support 96khz, so that test was done on the audio PC.  PC rig is not networked or connected

I think I do have decent hearing, and a fairly resolving system, but hardly a golden ear.  I do find that I can be sensitive to bass and sound stage though.

Anyhow, I would be happy to upload the actual files that I used for you to test ( I had offered this, but was told it wasn't necessary). Let me know and I'll do it.


Sure, let me know when and where they're up.

PCI soundcard w audiophile Analog stereo output

Reply #68 –
Quote from: krabapple on
Quote from: Dawnrazor-age on

Quote from: krabapple on


Your performance on 192 VBR with a modern LAME compile is outstanding, though not without precedent...IIRC at least one person here has reported ABXing even 320 CBR from source.  ABXing 44.1 from 96 is even more impressive.  It could be you are way out at the right end of the bell curve in audio acuity, but claiming both together makes me wonder of something's  funky in your ripping/compression/listening test setups.

Perhaps someone here could whip up some mp3s and flacs and/or samples at 44.1 and 96 kHz, for you to download and ABX. That at least takes the recording/compression part out of the equation.  Then just make sure you aren't using replaygain or other differential DSP during ABX.


I must tell you that the mp3 test was done on my non audio PC...just the onboard sound.  It won't support 96khz, so that test was done on the audio PC.  PC rig is not networked or connected

I think I do have decent hearing, and a fairly resolving system, but hardly a golden ear.  I do find that I can be sensitive to bass and sound stage though.

Anyhow, I would be happy to upload the actual files that I used for you to test ( I had offered this, but was told it wasn't necessary). Let me know and I'll do it.


Sure, let me know when and where they're up.



The MP3 is no  problem, but the .wav file is big.  is ther a place that will handle over 70mb?  Even if I change it to flac or some other lossless format, it will still be fairly big.  Any ideas?

PCI soundcard w audiophile Analog stereo output

Reply #69 –
Quote
The MP3 is no  problem, but the .wav file is big.  is ther a place that will handle over 70mb?  Even if I change it to flac or some other lossless format, it will still be fairly big.  Any ideas?



You should be able to find an online site that will host a flac file.

PCI soundcard w audiophile Analog stereo output

Reply #71 –
oh dear, was about to ask if there were any "high quality 5.1 DACs out there" >_> it's late and i'm still at work ;-p
 

PCI soundcard w audiophile Analog stereo output

Reply #72 –
Quote from: Mike Giacomelli on
Quote from: Dawnrazor-age on

Quote from: Xenno on

-V 2 (alt preset standard) has been LAME's historical high quality baseline ever since presets were added to the encoder.


I DID use 3.97 192 vbr...I believe that IS -V 2.




Can you please confirm that you did -V 2?



Since there is no V2 setting, I am not 100% certain.

But, I did use EAC with Lame 3.97 and a setting of high quality and a bit rate of 192, which is the same as -V 2 AFAIK