There is a fallacy in the idea:
If the recording process uses a high rez format that there is information captured that is useful in the consumer product.
The reason to use a deeper fit depth and faster sample rate is to avoid numerical nasties in the processing steps causing problems. It isn’t viable to edit and process the signal in 16 bits. You get truncation and round off errors everywhere. These cascade and you may end up with a result with much less than 16 bits of valid information and lots of distortion creating artefacts. This is true even if your ADCs are all 16 bit.
Same is true about sample rate. You can’t change the pitch or otherwise resample unless there is wiggle room in the sample rate. Also you want some room to avoid the effects of various convolutions ringing. More taps at higher sample rates for you FIR and IIR filters and so on.
The core fallacy is that the recording process ever had real human perceivable information in the higher resolution format over that that could be held in a 16/44.1 format. In the end, do the microphones, preamps, venue and instruments all exceed these bounds. The answer is that they don’t.
Properly produced and mastered audio derived from the internal high res format can and does capture everything it is possible to hear in a 16/44.1 format.
Lack of cleanliness in the production chain of the various high res formats for consumers seems to be actually delivering a result where proper care has not been taken and where, if anything, second order issues in the chain may actually result in a worse outcome. Certainly it is never any better.
