Noel Borthwick [Cakewalk]

Doktor Avalanche
Well I don't know of any global settings that can get changed in Sonar, that can alter sound, other than via preferences and plugin manager. Correct me if I'm wrong? Plenty of studios share DAW computers as you know...

If you have a multi user setup and care about someone changing your settings, you shouldn't be sharing the same user profile. That said all the resampling settings are stored in aud.ini (which is per user) so you can easily backup and restore the config settings if you need to. And even latency settings can "alter sound". There are some plugins that sound different at different buffer sizes. 

Tom Riggs
I can state that TruePianos benefits from the upsampling. I read that many of the guitar amp sims might benefit but which ones?  I have guitar rig 3 and 4 that came with Sonar versions.
 
What about the reverbs, chorus and phasers etc that come with Sonar? Lets make a list of the plugs that seem to be improved by this new feature.
 

gustabo
It would be nice if you could right click the synth name in the synth rack and select upsample on render...

slartabartfast
An interesting commentary that suggests local re-sampling (which seems to be the point of this new feature) is inherently inferior to doing the whole project end to end at a higher sample rate:
https://varietyofsound.wordpress.com/2012/11/02/working-itb-at-higher-sampling-rates/

Can someone explain how one can determine (without using the notoriously unreliable ear) that foldover distortion has occurred during processing?

drewfx1 
Some have stated that in the free Reaktor Player the upsampling options were grayed out, but I'm not 100% certain whether that's a configuration issue or a limitation of the free player as I have the full version.

Tom Riggs
A bit further testing with truepianos seems to indicate that the only module that has a noticeable difference (at least to me) is the Atlantis module which is a preview for the next version. The other modules I tested I could not tell a difference.
 
I should also note that where I hear the most difference is in the attack of the note especially on the higher velocities. The upsample seems to have a more rounded attack and the regular seems harsh to me. Of course depending on what you want out of the track may determine weather you prefer the upsample or not.
 
YMMV
 
 

slartabartfast

Anderton
 
Been there, done that. Here's an example of a file I played at my address to the New Music Seminar last week in New York. This download link is good for 7 days.
 
The file is at 44.1 kHz. The second half was upsampled to 96 kHz then downsampled back to 44.1 kHz. If you can't hear a difference between the first and second sections, especially with headphones, you should probably consider a career path that involves something other than audio. 

I can certainly hear a difference. The second half has a much fuller richer sound especially in the highs. Is that because of the removal of foldover distortion? Assuming that something has been removed, the upsampled then downsampled again version should have less power at audible frequencies. Conversely if foldover or another form of distortion is introduced by the operation, there should be new frequencies or more power at existing frequencies introduced by the procedure. 
 

 
The above is a sonogram of your sample file. Notice the striking addition of new power in the top of the audible spectrum. Unless I am misunderstanding your description this second half is of the processed (upsampled/downsampled) signal. If so the original signal is the one which should contain only the original frequencies. Where did the new frequencies come from? If they are not present in the original and are present in the processed signal are they not artifacts? 

Notecrusher

slartabartfast
 
I can certainly hear a difference. The second half has a much fuller richer sound especially in the highs. Is that because of the removal of foldover distortion? Assuming that something has been removed, the upsampled then downsampled again version should have less power at audible frequencies. Conversely if foldover or another form of distortion is introduced by the operation, there should be new frequencies or more power at existing frequencies introduced by the procedure. 
 

 
The above is a sonogram of your sample file. Notice the striking addition of new power in the top of the audible spectrum. Unless I am misunderstanding your description this second half is of the processed (upsampled/downsampled) signal. If so the original signal is the one which should contain only the original frequencies. Where did the new frequencies come from? If they are not present in the original and are present in the processed signal are they not artifacts? 

Yes, seeing it here too w/ SPAN. The oversampling is adding distortion at 19.4k and above.