The Upsampling Your Mix Thread

[bduffy]

YAY!

[mauseoleum]

I reinstalled my ancient FLs, too. It really works as expected. My opinion is that it's the best suited for such upsampling/mix-oversampling methods.

But - I could't get the sonogram plugin to display properly, maybe a graphic issue or something to do with time constants. Did the 5second 20Hz-20kHz sweep, conversions, etc, but couldn't get the sweep to display nicely in the window. Nevermind.

What is next to find out - could be FLs then used as a VSTi oversampler? f.e. within cubase/nuendo? Wild guess, but possible?

In other words, at least with eXT1.4, you'd get a funny syncopated mess if you export 44k at say 96k and have arps hooked to midichannels.

I think somebody should transfer this to the developer. Such behaviour should be prevented/corrected in XT2 (imho).

Which is the same basic issue with cubase audio file alignement when changing rates.. Something to look out for in your individual host/effects setup and audio alignment scheme..

L

[iShawn]

How do I set Tracktion up for 32bit float? Or is it automatically 32 bit if the 64 bit isn't check?

[kylen]

How do I set Tracktion up for 32bit float? Or is it automatically 32 bit if the 64 bit isn't check?

Yes that's it 32bit internal processing by default in Tracktion2.

[ermi]

I haven't read the whole thread in detail (too much off topic content ) but I went on and did some tests with plugins at different SRs.
Here's a few of the popular ones (converted the 96k files to 44k for your convenience):

Ambience:
http://files.myopera.com/ermi/files/hh-ambience-44.wav
http://files.myopera.com/ermi/files/hh- ... 96_r8b.wav
GlaceVerb
http://files.myopera.com/ermi/files/hh-glaceverb-44.wav
http://files.myopera.com/ermi/files/hh- ... 96_r8b.wav
FreeAmp2:
http://files.myopera.com/ermi/files/fa244.wav
http://files.myopera.com/ermi/files/fa296_r8b.wav

I was blown away by the differences in sound. So much that I'd even call some plugins broken at a certain SR (see the FreeAmp2 example). Some other plugins seem to sound more or less the same at both SRs, tho.

I'd like to hear some opinions... which of these statements would you say is true:
1) The plugins at different SRs just sound DIFFERENT. They may sound either better or worse at a higher SR, depending on how they are programmed.
2) The plugins in theory just sound DIFFERENT at different SRs, but in practice they almost always sound better at higher SRs.
3) The plugins sound BETTER at higher SRs, both in theory and in practice. Always.

Also: if a plugin has an oversampling option (like GClip)... does this mean it will work exactly like if it worked at a (2x) higher SR?

[Kingston]

I was blown away by the differences in sound. So much that I'd even call some plugins broken at a certain SR (see the FreeAmp2 example).

some plugins are broken at high sample rates.

Also: if a plugin has an oversampling option (like GClip)... does this mean it will work exactly like if it worked at a (2x) higher SR?

the oversampling algorithm of any real time plugin will be of limited quality, due to CPU constraints. a plugin working natively at a high sample rate will always perform better (somewhat depending on what tehcniques you used to upsample your mix, or what rate your source material is).





The real issue here is aliasing in all it's forms. It's never a question of what sounds better or worse, but what techniques you use to avoid aliasing - or if you want to avoid it at all!. working at a high sample rate is one of them, but as always, YMMV.

[bduffy]

ermi, are those reverb examples with the exact same settings at each sample rate? I don't recall hearing that big a difference in reverbs between sample rates; those sound like totally different settings! (not accusing you of anything, of course)

[ermi]

Thanks for the explanation, Kingston.

bduffy: Yes, same exact settings. But those are some of the most drastic examples, especially the FreeAmp2 one. But also on some choruses... it sounds just like the settings are different. I also tried with two hosts (Renoise and Cubase), just to make sure.

[bduffy]

Thanks for the explanation, Kingston.

bduffy: Yes, same exact settings. But those are some of the most drastic examples, especially the FreeAmp2 one. But also on some choruses... it sounds just like the settings are different. I also tried with two hosts (Renoise and Cubase), just to make sure.

Interesting. I guess the plugins I use don't display that big a difference!

[dimitar]

Wow!!!!! the last test blow me away!!!

[ermi]

I tried some synths as well... Some sounded more or less the same, some sounded a little bit dull at 44k.
But then some sounded like they were played at a higher pitch when played at a higher sample rate. These were Anna, Chimera and Ethereal.
http://files.myopera.com/ermi/files/ann ... nd_96k.wav

Is this perhaps normal for certain kinds of synths, maybe those that use some samples/waves or something? Or are they simply bad? Didn't mean to offend anyone, but I'm curious.

[jackson]

Yeah, it's the same with some vocoders I tested and mentioned a few pages back; they play higher at faster rates. 88.2k plays exactly an octave higher than at 44.1k.

[bduffy]

Some plugins simply aren't coded to work at higher sample rates; they not bad, they're just built that way.

[Kingston]

Some plugins simply aren't coded to work at higher sample rates; they not bad, they're just built that way.

it's still a bloody grave error on the programmers part.

It's not like it's hard, just that some people are lazy or don't know better.

there's simply no excuse for it.

[laputa_sync]

YAY!

get a group buy on more "YAY", save 25%.