Implement optional automatic real-time dynamic range correction for all hardware audio outputs
All recording are typically mastered to about ~90db of volume and everything in the system is mixed for full DR of chosen encoding format (96/144/1528 db for 16/24/32f bits) but in reality such high volume is never used (and in case of 32f, impossible and would explode you if it wasn't), so significant amount of minute details are lost because of that. Using pre-output RT DRCor stage that matches volume of the DAC to "fit" stream's "virtual" loudness with real one could alleviate that, assuming that post-DAC amplifier (if DAC isn't integrated into the amp, be it power amp for speakers or op-amp for headphones) is constantly set to its optimal highest volume/power and real volume is fully software-controlled. Similarly to colour & gamma correction stage of video output pipeline.
Linux Studio Plugins' author had an idea on how to implement that in more correct way than usual DR compression allows but it's better to be done directly in OS' audio management system to account for system volume controls. When determining real DR of the output, it may also be practical to account for noise floor of user's environment (40db for typical "quiet" room but that may be completely irrelevant for sealed headphone outputs) in addition to volume/power ceiling. Moreover, "DAC" devices/chips mostly use overall "Master" volume for DAC circuit itself and separate volume per each port's op-amp meaning that real DR's ceiling for each physical port is combination of two figures. The resulting real DR often will be quite a measly figure.
There is also another way to make things more pronounced, "equal-loudness compensation", which is discussed there but that one is based on assumptions about generalized human sound perception capabilities and changes frequency response that should instead be left as "flat" as possible. However, both approaches could be used in conjunction independently to squeeze maximum audio details using minimal signal volume.
This should greatly increase "clarity" of sound in all practical use cases and for hardware of any build quality, except for physically dysfunctional "distortion generators". But sound engineers would not like that being enabled while they work.