Adq
Offline upsampling has no much sense without real-time upsampling, because if there is real difference, it is often so huge, that sounding changes significantly, and if one set up some sound without upsampling, it would very often suit worse with upsampling, just because it would sound very different, I think so.

For the time being, you'll want to freeze the synth with upsampling applied before mixing.

twelvetone
Craig, some time back there was a thread, I think started by your good self, concerning recording VST instruments and in some cases the result sounding better/different at higher sample rates.
Is this what this is intended to address?

Yes, exactly. FWIW I played an audio example at the Mix/AES event in Nashville during my keynote address, and even in a trade show/noisy context, people could hear the difference. It also makes a difference with amp sims, some dynamics processors, etc.
 
Whether there's an improvement or not is hit-or-miss - it depends on the plug-in itself, and what's being processed. That's why the real-time upsampling is a welcome improvement.
 
Also note that this is kind of like the Console Emulator in the sense that you may not notice much difference on an individual track, but multiplied over multiple tracks where there's a benefit, the overall mix is cleaner. Gasp!

Interesting... if upsampling a 44.1k project, does it unconditionally get downsampled on the master output? What if the project is being exported at the sample rate of 88.2? Will the master output be downsampled and then re-upsampled?

yevster
Interesting... if upsampling a 44.1k project, does it unconditionally get downsampled on the master output? What if the project is being exported at the sample rate of 88.2? Will the master output be downsampled and then re-upsampled?

yevster
Interesting... if upsampling a 44.1k project, does it unconditionally get downsampled on the master output?

 
You don't really upsample an entire project per se (nor do you necessarily need to), you upsample/render/downsample a track or clip that contains a plug-in when you render it to audio.
 
The advantage of real-time upsampling is you can determine whether the upsampling process makes a difference or not (most of the time it doesn't, but when it does...). Upsampling/rendering/downsampling back down takes more time than simply rendering, so you don't want to do it unless you need to.
 
I have a few projects that use multiple plug-ins that benefit from upsampling. When I bounce down to two tracks, it takes a lot longer as the individual tracks are upsampled/rendered/downsampled during the rendering process. The audible results are definitely worth it, but I wouldn't want to spend the time if it wasn't necessary.
 
 
 

yevster
Interesting... if upsampling a 44.1k project, does it unconditionally get downsampled on the master output? What if the project is being exported at the sample rate of 88.2? Will the master output be downsampled and then re-upsampled?

If exporting a 44.1 project at 88.2 yes because other plugins in the project are still being rendered at 44.1.
However it doesn't make much sense to export to a higher sample rate since you aren't going to get any gains. Upsampling there is only done at the final stage. A

Noel Borthwick [Cakewalk]

yevster
Interesting... if upsampling a 44.1k project, does it unconditionally get downsampled on the master output? What if the project is being exported at the sample rate of 88.2? Will the master output be downsampled and then re-upsampled?

If exporting a 44.1 project at 88.2 yes because other plugins in the project are still being rendered at 44.1.
However it doesn't make much sense to export to a higher sample rate since you aren't going to get any gains. Upsampling there is only done at the final stage. A

Then, since individual tracks can be upsampled, wouldn't it make sense to upsample as far down the mixing chain as possible? Why not downsample only on the busses that contain plugins that can't handle the higher sample rate? Then, if no downsampling occurs before the master bus, the project need only be downsampled to the export rate.
 
This is particularly important in Sonar as, unlike StudioOne and Cubase, the sample rate for a project is immutable. A project created for CD at 44.1Khz cannot be easily re-mixed for video release at 48khz, or for any higher sample rates. Upsampling that goes maximally far up the routing graph could ameliorate that deficiency.

As understand it, it isn't taking place at the level of tracks and buses, but only within individual plug-ins.

Correct this is at the plugin level not bus or track.
What you are talking about is a completely different request. To resample at a project level.

yevster
A project created for CD at 44.1Khz cannot be easily re-mixed for video release at 48khz, or for any higher sample rates. 

The easiest and highest-quality solution for any program is to mix in the project's native format, then sample-rate convert the mixed file. That way you need do only one sample rate conversion to one file.
 
If for some reason you must re-mix at a different sample rate, and therefore convert the entire original project:
 
1. Export the existing tracks using SONAR's ability to export individual tracks.
2. Change SONAR's sample rate, and create a new project at the new sample rate (e.g., 48 kHz).
3. Import the tracks you exported.
 
When you import the tracks SONAR will sample-rate convert them using a very high quality sample rate conversion algorithm. Now your project is at the new sample rate and you can mix/render at that sample rate.