I don't know if you guys know this but there is new resampler dsp and good news is that I did some abx tests against default resampler and in both cases (slow and normal) i got 10/10 in favor of new resampler (set at top quality).
(testing done with Audigy, some headphone amp and Sennheiser HD580)

You can find it at otachan's site.

Now some bad news, it consumes more cpu power and may have some problems, it's still in developement.

Combined with kx drivers and otachan's asio output and you have probably the best combination for Live/Audigy series of soundcards.

yeah that does use a LOT of CPU and makes nasty mouse lag on my system. probably not worth it for me/most people. any info on what he's done?

For a laggy mouse solution is to reduce priority in Foobars properties (playback thread priority) and reduce resampler's quality to high.

(i don't know technical specifications on resampler)

Thanks for the link nuhi. It takes up a little cpu but I think it's worth it.

Top quality also creates a small lag when jumping around a song.

Hi nuhi.

That's interesting. Can you please give some more detailed information about how/what you've ABXed:
- What was the original sample rate?
- What output sample rate did you use?
- What output bit depth did you use?
- What setting did you use with default resampler? (slow mode on/off)
- How often did you resample (only once or several times to emphasize the difference)?
- What did you do to prevent clipping?
- What software/settings did you use for ABXing?
- Was replaygain enabled (or some other way to match volume)?
- Did you make sure there's no offset?
- What was your reference (since audigy cards resample I wonder how you determined what sounds closer to original)?
- Related: What did you ABX against what?
- What samples did you use for ABXing? Real World music or some test signals? Can you upload < 30 seconds samples, please?
- Can you please describe the difference you've heard and ABXed?
...

Cheers tigre

P.S.: I don't want to (nit)pick on you, there are just too many points where something could have gone wrong in your test, so I want to be sure...

Offtopic posts moved to recycle bin.

nuhi: my guess is that you have an audigy1 with creative's drivers, and you have compared resampling to 96 kHz (which afaik isn't supported natively by this combo; resampling to 48 kHz is performed before passing it to the DAC). This could make the normally inaudible differences noticable.

People generally can't ABX SSRC vs original signal, so if you ABX'd old SSRC vs "new" SSRC then there is something seriously wrong with this "new" version. Anyway, follow tigre's post.

Hi, i just got home.

Talking about interrogation...lets go

I used it on Audigy with kx drivers, and i know what resampling is and what are the limits of ac97 codec.

- What was the original sample rate?
44100

- What output sample rate did you use?
48000

- What output bit depth did you use?
16bit, like the song.

- What setting did you use with default resampler? (slow mode on/off)
Both

- How often did you resample (only once or several times to emphasize the difference)?
Because of all that placebo "danger" i test myself every now and then with resampler turned off, and there is a clear difference, it all gets "blurry" without it.

- What did you do to prevent clipping?
Replaygain

- What software/settings did you use for ABXing?
Foobar plugin with special version.

- Was replaygain enabled (or some other way to match volume)?
It was at the same volume, both of the tracks.
I made them from a same track, only changed dsp setting when converting to wav.
(then i played it abx-ed without any dsp because now they are 48khz)

- Did you make sure there's no offset?
No.

- What was your reference (since audigy cards resample I wonder how you determined what sounds closer to original)?
Hm....that's a good question, i did no such a thing...i only compared two samples blindly and decided which one was better and at the end it was the other resampler.

- Related: What did you ABX against what?
Same track with replaygain, different dsp resampler.

- What samples did you use for ABXing? Real World music or some test signals? Can you upload < 30 seconds samples, please?
Pain of salvation....metal
I don't see the point of uploading, i'm sure i can pick difference with any song.
It took me some time till i got used to the difference, at first 2 minutes of the song they were same to me, but after i calmed myself down and turned up the volume (maybe earned some tinnitus) there was this slight difference in clarity

- Can you please describe the difference you've heard and ABXed?
I know it's subjective oppinion and it doesn't matter to you and it shouldn't.
It is quite simmiliar sound but more i listenned the quicker i got the difference, but of course only if you try hard (doesn't matter what song, only let it have some clean and rich sounds, vocal especially.Concentrate on vocal clarity.

If you are so sure that it was some error on my side please let me know.

Thanks, nuhi. Possible problems can be caused by:
For testing write 32bit float .wav files. Fb2k ABX tool will convert to the output bit depth you've selected. This will avoid problems that might be caused by dithering.
It's possible that the resampling algorithm changes volume. (SSRC standalone does this in 2pass mode too to avoid clipping.) Use trackgain of converted files for ABXing to make sure this doesn't cause audible differences.
I think this is no big issue since you didn't say you focussed on the beginning/end of a certain selection.

Biggest problem I see here... You could try this: resample to 48kHz/32bit float and back to 44.1kHz/32bit dithered. Now either:

• Apply trackgain and convert to 16bit (flat dither). Burn all versions to CD and ABX using a decent standalone CD player (random mode, someone else writing down what track has been played and what you think it was),
  
  
• Get your card to play back without resampling somehow (e.g. using digital out + external DAC - I'm not sure if this works, there are some threads covering this) and use fb2k ABX tool or
  
  
• Do the same with a non-resampling soundcard (some friend's PC?)
  

Uploading shouldn't be a problem. Use hydrogenaudio upload forum (< 30 seconds lossless sample). The point is that with your sample and a good description of the difference, it's much easier for others to verify your results (or to find out what's wrong). Additonally, there've been some ppl before who've claimed they ABXed, but when ppl ask to provide a sample, they refused and/or disappeared. So a sample will increase the probability that ppl take you seriously and are ready to spend time on this.

I tryed to ABX them but with no luck.

tigre, thanx for all that instructions but i'm not ready for all that testing.
I'll upload samples later when i find a part of some song to be easily recognized in 30sec.

Then you can do some spectrum analysis or something to tell me is it maybe dithering applied without my knowledge.

(i'll do it in couple of hours, i can't right now)

Damn I can't tell the difference...

There, i uploaded part of some song where i can distinguish those two DSPs.
But 28seconds is nothing, so this test is doomed to disaster .
You have to make converts, i couldn't have failed there because i did them both the same.
(today i noticed that they are not the same in size, one is few kb's bigger then the other)

clean sample here

Thanks, nuhi. I haven't been able to hear any difference, I also tried with some of my music. In the upload thread I've uploaded 2 versions of your sample resampled with CoolEditPro. After resampling (saved at 32bit float resolution) they were replaygained and converted to 24bit with foobar2000. Could you please test and tell if you hear a difference / what they sound like compared to fb2k native resampler and otachan's SSRC resampler (applying trackgain (to the sample, not the whole thread) -> converting with fb2k diskwriter with replaygain enabled should be enough to get the same volume for ABXing)? Thanks in advance.

Finaly i did the test, problem with quiet moments and college.

Results are...

Native vs Cooledit01, same
Native vs Cooledit02, same
Otachan vs Cooledit01, same
Otachan vs Cooledit02, Otachan wins 8/9 (with favor on some more "air" or maybe it produces more hiss, so then it sounds like more crisp (maybe). It acctually can be a bad thing...so you tell me what you did with the other sample then we can draw some conclusions.

And the song is to calm, to slow...i know i choose it, but now i regret it...because on the whole song i can more easily tell the difference...but i wonder is it all just luck because they all sound 99,9% same.

Normally, now i hate the song.


edit: grammar corrections