I'll try and keep this to the point:
Let's say you import 32 stereo tracks into the vse. Stack them ontop of eachother, occupying all 32 channels in the vse. You mess around with the volume levels in blender (anywhere from 0 - 9.0) and get it nice sounding on ur earphones in the vse playback. And now you want to mix and render them out of blender as a single stereo track. HOw on earth would you do this in blender to ensure highest quality possible?
You have to use mixdown, that's a given to get acceptable quality. So is it a case of setting your accuracy to 1 and choosing a high quality format like .wav f32 for your render settings? The only higher format is f64, with ridiculous file size, no?
The big issue though is how do I export it knowing blender amplifies audio waveforms when it combines more than 1 track together rather than doing the merge you can get in programs like audacity. Coupled with the fact that you have no idea outside of guessing and feeling, whether you are clipping anywhere if not everywhere in ur track, as the vse playback gives 0 indication and there is no setting to confirm in the vse. So would you lower the master volume (or gain, or amplitude, whatever word you refer to it by) to a very low setting and render out that way? (You can do this under properties->scene the mastervolume is available there) But do I damage quality if I render out at say 0.1 or 0.2 of the original 1.0 master volume? Do I need to choose the highest mastervolume possible without clipping to get the best export quality?
Curious if there are any other methods out there or settings I missed.