Have been thinking about how to experiment to get a better understanding of this, maybe roughly 3 ways:
- use a "good" home audio system and add delays/distortion until II get nausea.
- use a known "bad" car audio system and correct until it feels good enough.
- measure distortion and time alignment etc in audio systems, instruments, human voice, combustion engines vs electric vehicles etc and record together with subjective experience..
Just got more into measuring distortion in the car audio systems I am working on, and noticed I hadn't really analyzed what I disliked in the pianos that I hated as a child or later when I attempted to tune by myself until I noticed that that's not it.
In addition to delay and distortion, it could also be L/R or Front/Rear balance, in a car. Also, what do you mean by delay for a car environment? Or do you mean extra reverberation? There's also noise (e.g. pink) and distortion has different flavors too. It would be useful to write out all the different dimensions and then experiment with each individually. If those are all null, then try in pairs, etc. Google "factorial experimental design." Of course, it could be biomechanical -- some combination of RPM rumble or road/suspension giggle with audio stimuli. Experiments are hard to construct that disentangle all the independent variables and confounding parameters.
So, maybe it would be better to start with a known bad, measure it, and see what the dominating characteristics are (e.g. balance 60/40, high reverb), then come up with a hypothesis about those and do a "known good" with hypothesis testing to see what the true interactions are.