about mixing sounds against each other so each can be heard clearly. well I’m not sure if all this is correct. I never read about it in books or so…I just made this up doing my own research, but it seems to work™ for me.
human hearing is divided into “critical bands”, that are like bandpass filters that can move their centers a bit up and down to focus elements. Humans can only distinguish one sound per band. the bark scale seems to be easiest to use for such purposes: https://en.wikipedia.org/wiki/Bark_scale
if in one critical band partials of a sound are like about 3-6bd louder than any other noise in that band, that sound will be clearly audible in the foreground. ofc you can mix a sound so it is in foreground in multiple critical bands. if other bands are shaddowed, the sound will be audible just like it would sound with only the foreground bands audible (you can check with filters/eqs).
this seems to be true for transients at the same time as for tonal sounds or atonal noise. so each band can yield both transients and some continuous tone, and all might be heard clearly.
regarding timbre of a sound, if its energies are mostly at the bottom range of a band, it will be percieved as “low/relaxed” while on the upper side of each band it will sound like “high pitched/pressed/energetic”. this is also true that a noise centered about the lower range of a higher band will sound more “relaxed” than one on the upper side of a band below.
for transients there seems to be a certain timeframe within they melt into one perceived hit, or if far away enough in time will be distinguishable. so shifting a track in time such that the attacks, especially of rhythmic sounds, are aligned in a succession in time in order of importance, can result in the attacks of the single hits becomming more clear leading to good separation of the sounds. To glue them together ofc you would have to do the opposite and align the peaks of the transients as exact as possible.
regarding the continuous tones, especially in the mids, although most often mixing two sound would result in mud, it seems sometimes possible to mix two tones with different timbre in a band, and get away with that band sustaining two instruments as the sounds blend.
another nice trick is clearing single bands for bringing up certain transients or reverb tails. for example, if my snare punches at 400hz, it might be a good idea to notch or sidechain that range away from other instruments, so that the snare punch will be strongly perceivable. or chop some out rigidly at around 7000hz without crippling timbre too much, and then eq some reverbs to be strong within that band -> suddenly space/depth can become perceivable even in a busy mix.
at some frequencies the human hearing seems to be more able to seperate muddy stuff from each other, this seems to be true around 1000hz and especially around 3400hz - here it seems to be possible to place vital info and delay tails etc. of multiple instruments.
a good idea to find problematic areas can be to have a sharp bandpass filter on the master and tune it for a single band, then toggle. sweeping, you can check your mix very well, and see where each instrument is in the foreground, and which bands consist of mud and would need some work out.
for stereo the possible info seems to be kind of partially doubled, so there is potentially extra info perceivable. but not all people will listen with headphones, so it might be a good idea to mix or consistently check for clarity in mono mode, and consider width and panning stuff as secondary qualities. this way you will have a clean mix as desired under any stereo circumstances. because it is possible, that something will sound very easily distinguishable in a good stereo field, but will collapse to mud on a badly placed home stereo.
ofc trying to mix with just this info and deaf ears is bullshit, you will generate ugly unnatural gibberish. you still need to eq sounds by ear, such that they sound well defined. but this info can help a lot in finding and fixing problems, when you just can’t get that mud weeded out and need a guideline about what and where to look for.