[fenixdown110]

Using SoX here as well and it's amazing for the same reasons the OP stated. However, how does this compare to the Secret Rabbit resampler?

 

[thuantran]

@fenix: http://src.infinitewave.ca/ . It's one of the best at Very High setting. BTW you should not use any lower quality setting, I found that it barely made a difference in speed compared tio the quality you compromised (on paper, whether that is noticable is a different matter).

 

[fenixdown110]

Thanks. However, it's a 96 to 44.1 resample. I guess I should have specifically asked for the differences in upsampling.

 

[bmcelvan]

Quick question here...when downsampling from 24bit either 96/192kHz to 16bit either 44.1/48kHz...should I enable aliasing/imaging or keep it off? Or does it really not matter at all because any difference between the two are going to be differences above 20kHz hence I won't hear them?

The full settings I'm using are 50% (linear), 95% passband, and dithering. Are these the recommended/best settings? Obviously dithering is because of the bit change.

I'm asking these questions because I've been reading a bunch of posts about aliasing and right when I think I shouldn't use it because a great encoder should not introduce much noise, I read 5-10 posts in a row about how it does add noise but in doing so it actually reduces a bunch of bad noise, ringing, etc.. So then I think I should use it. So I guess my above question is the most important...does it doing anything noticeable at all or is it above the hearing threshold anyway and so it doesn't matter (to me) who will only be listening to it.

Thanks
Ben

 

[Yuri Korzunov]

Quick question here...when downsampling from 24bit either 96/192kHz to 16bit either 44.1/48kHz...should I enable aliasing/imaging or keep it off? Or does it really not matter at all because any difference between the two are going to be differences above 20kHz hence I won't hear them?

The full settings I'm using are 50% (linear), 95% passband, and dithering. Are these the recommended/best settings? Obviously dithering is because of the bit change.

I'm asking these questions because I've been reading a bunch of posts about aliasing and right when I think I shouldn't use it because a great encoder should not introduce much noise, I read 5-10 posts in a row about how it does add noise but in doing so it actually reduces a bunch of bad noise, ringing, etc.. So then I think I should use it. So I guess my above question is the most important...does it doing anything noticeable at all or is it above the hearing threshold anyway and so it doesn't matter (to me) who will only be listening to it.

Thanks
Ben

Hi Ben,
 
All aliases that above 1/2 sample rate as input as output must be suppressed. Otherwise it will be mirrored to audible range:
 
http://samplerateconverter.com/content/how-convert-sample-rate-divide
 
http://samplerateconverter.com/content/how-convert-sample-rate-oversampling
 
Pro resampling algorithms has supresion level -160 ... -180 dB and more (for 32- and 64-bit float bit depth).
 
When you convert 24-bit and more to 16-bit, dither must be turned ON.
 
More about dithering in pictures:

 
http://samplerateconverter.com/content/what-dithering-audio
 
If you have anything in ultrasound frequency range, it may be shifted/mirrored to audible range too, due non-linearity analog part of apparatus.
 
Ultrasound components may appear due not proper filtration of DAC (as analog as digital).
 
Check shifting/mirroring troubles we can on hear with sweep sine 0 ... 1/2 sample rate of test signal.
 
We can check all components (player software - operation system driver - DAC) as system.
 
If system work properly we must hear only growing pure tone from zero to 17....20 kHz.
 
If we listen growing tone again (after disappearing sound at 17 .. 20 kHz) - it's shifted ultrasound. If we hear downward tone, it's mirroring.
 
Here need be careful with ultrasound. We don't hear it, but ultrasound can damage our ears.
 
Best regards,
Yuri

 

[Soundsgoodtome]

Hello guys, I'm resampling some hi-res flac (88, 96, 176, 192khz - 24bit) to 44.1 or 48khz - 16bit files on foobar. I'm using SOX resampler which I've heard has built-in dither (please confirm).

Question is if SOX resampler plugin does have built-in dither, do I need to apply dither at the file output setting as well (pic below) or is that redundant and not necessary? If the latter, how does it affect sound quality?

 

[Dobrescu George]

Hello guys, I'm resampling some hi-res flac (88, 96, 176, 192khz - 24bit) to 44.1 or 48khz - 16bit files on foobar. I'm using SOX resampler which I've heard has built-in dither (please confirm).

Question is if SOX resampler plugin does have built-in dither, do I need to apply dither at the file output setting as well (pic below) or is that redundant and not necessary? If the latter, how does it affect sound quality?

You should never apply dither in that point of conversion, it will worsen SQ most of times.
 
Theoretically, there should have been applied a layer of dithering earlier, and applying one more adds noise, and sometimes it is quite audible. 

 

[Soundsgoodtome]

So you're saying that SOX resampler does add dither?

You should never apply dither in that point of conversion, it will worsen SQ most of times.

Theoretically, there should have been applied a layer of dithering earlier, and applying one more adds noise, and sometimes it is quite audible. 

 

[Dobrescu George]

So you're saying that SOX resampler does add dither?

I am not sure about this. But if you use a normal music file, already recorded and extracted from a CD, your original file should already have a layer of dither from what I know. Adding dither over dither degrades SQ generally. The best way to test is to do it 2 times, with the same file and see if you hear the differences, if you do not hear or with dither sounds worse, avoid dither. (dither basically adds noise to mask any artifacts that were induced by processing, but resampling generally should not add artifacts).

 

[Soundsgoodtome]

Hello guys, I'm resampling some hi-res flac (88, 96, 176, 192khz - 24bit) to 44.1 or 48khz - 16bit files on foobar. I'm using SOX resampler which I've heard has built-in dither (please confirm).

Question is if SOX resampler plugin does have built-in dither, do I need to apply dither at the file output setting as well (pic below) or is that redundant and not necessary? If the latter, how does it affect sound quality?

-------------:--------------

The files are originally 24bit 88-192khz not from CD. I don't apply dither to CD extracted files.

I am not sure about this. But if you use a normal music file, already recorded and extracted from a CD, your original file should already have a layer of dither from what I know. Adding dither over dither degrades SQ generally. The best way to test is to do it 2 times, with the same file and see if you hear the differences, if you do not hear or with dither sounds worse, avoid dither. (dither basically adds noise to mask any artifacts that were induced by processing, but resampling generally should not add artifacts).

 

[Reggy]

Quote:

Originally Posted by dex85 /img/forum/go_quote.gif i gave it a try, i second the better focus, clearer vocals, deeper soundstage, better imaging, ambience details retrieval and a sense of "faster" sound and musical flow. but i noticed the sound significantly lost on the weight, cello and violins immediately struck me as too thin sounding and unnatural. i don't doubt it has to do with headphones, equipment and recordings' characteristics as well. with some songs i preferred it with Sox but with most i didn't.

Matches my findings exactly. SoX makes the sound a bit thinner and less bassy, but makes it faster. Depending on the setup, this could be a good or bad thing. In my case, it wasn't a positive change, though it was outweighed by the benefits.

Really interesting you found any difference let alone significant. What USB driver output do you have? I am using Combo384 Amanero which my DAC takes.
Settings I have are:
Upsample x4
44100;48000;88200;96000
Allow aliasing/imaging.

I found no difference in using aliasing. I can not really compare to without resampling but I can say the sound is "technically" 99.9% transparent to my ears without any added tubes or added color in the chain or what not. Very clear and accurate transient sound absolutely no veil for excellent recordings to say, soundstage is phenomenal, I could imagine a very slightly coherent one in my imagination though (I don't have any Hifi Speakers to compare to).

I'm guessing the differences your hearing could be to some other limitation your experiencing. Just curious how.