Print

Please login or register.
1 Hour 1 Day 1 Week 1 Month Forever
Login with username, password and session length

[6233638]

8. Changed: Switched the resampler to high precision mode (testing appreciated).

Did some preliminary testing, since it was requested:
 
IMD: Performance seems largely unchanged using a standard IMD test. (60Hz -5dBFS tone, 7kHz -17dBFS tone)
Which is to say that it's mostly fine.

Aliasing: The old SRC did not seem to properly filter out the highest frequencies.
A 23kHz tone showed some aliasing with a 48kHz test file resampled to 44.1kHz. This is fixed with the new resampler.
I will say, however, that this aliasing did not seem to show up in the music tests that I did, only tests specifically designed to demonstrate aliasing.
 
Upsampling: It seems like there is less distortion than before.
A 1kHz tone at 44.1kHz upsampled to 192kHz:

• 20.0.66
  
• 20.0.68
  
• Reference
  

 
I must point out that neither of the VST plug-ins that I have for dither seem to work with Media Center's conversion tool, so I am reliant on MC's RPDF dither in these tests.
So it is difficult to say whether these are actually resampler artifacts, or if they would be eliminated with TPDF dither. (I suspect the latter)
 
I haven't done more in-depth testing like impulse response tests or testing a 20-20k sweep for example.
I'm a bit out of my element when it comes to testing resamplers to be honest.
 
Rather than getting into the minutiae of one resampler vs another, where it could be argued that someone may prefer to use a resampler other than SSRC, the tests at Infinite Wave Mastering indicate that SSRC High Precision should perform better than regular SSRC, and my own testing generally seems to agree.
So without requesting a big overhaul or a switch to another resampler, it seems like it should be of equal or better performance to what was previously in use.

[Dr Tone]

I'm awaiting a response to whether the updated resampler is in the Mac version yet.  I down-sample some stuff for my Dragonfly at work.

[6233638]

Other than some potential aliasing, which does not seem to be an issue with actual music in my testing, there really doesn't seem to be much of a difference between SSRC and SSRC High Precision.
It seems better in some very specific situations, but overall I do wonder whether anyone will notice.
 
I did a handful of other tests which did not really show anything other than highlight some differences between SSRC and other resamplers, rather than finding differences between SSRC and SSRC High Precision.
And that mostly seemed to illustrate differing amounts of ultrasonic noise created when upsampling (SSRC, High Precision, Other) rather than differences when downsampling, and those were, again, far less noticeable with music than specific tests. I was only testing linear phase resamplers though.
 
But this shouldn't be a surprise. SSRC is known to be an excellent tool for downsampling.
 
I will say again, that this is far from an area of expertise for me.
I'm just trying to run through some of the tests I have seen performed on resamplers on other sites.

[Dr Tone]

I may or may not be able to hear a difference but I just want to know it's being used.  It's already driving me nuts knowing it wasn't using the high precision option before.  I know, I have a problem!

[Fabith]

Right now, i'm upsampling all my 44,1khz files to 88,2khz and no doubt there are audible improvements in the sound.

 Thanks for this improvement