The most you should worry about this, is if your source is is capable of processing or outputting bit-perfect (which shouldn't be an issue if you run WASAPI or ASIO with no DSP in the pipeline). Volume control on-device though is pretty pointless to worry about with respect to it being "bit perfect" the moment you don't output full-scale. I wouldn't even understand how that would even work exactly speaking since all volume is essentially 'attenuated' at some point before it reaches you in practice, since you usually set volume to a certain level - unless you output full scale with no voltage gain or "unity gain" from an amp section, which sounds like a way to blow your ears away or something.
So if your device has digital-out, I suppose bit-perfect processing and handling makes sense while it's in the digital domain. Once it's out of this domain, or something besides full-scale output (whether due to volume setting, or due to DSP, or whatever else) bit-perfect ceases to exist for all intents and purposes.
At least that how it seems to be when I try to comprehend how it all works.