unfortunately, at this point, what i've been able to confirm is inconsistency across both cards over asio. it's the same type of error - the bass moves from clean and crisp (which is what i want) to muddy and undefined, the high mids cut in a way that makes the guitars sound flat and a few other things. it's at least repeatable, if random. there seems to be about five different settings that come in and out without warning. i can get it to go back to "normal" by cycling, but it doesn't stay there.
i don't think this is repeatable with the m-audio card using the windows mixer, but i need to test it a little more strenuously. if i can't repeat it, i have to conclude it's a broken asio. this at least suggests the hardware is not broken, which is positive.
you normally fix this by reinstalling drivers. but, it's starting to seem as though the codec install broke something a little deeper that repeated driver reinstalls aren't fixing and that a full os reinstall is necessary to address. i mean, i'm pretty good in the registry, but it's a labyrinth that one cannot really navigate.
i can't mix on the m-audio card through the windows mixer - even if i'm able to determine consistency. it's just a lower quality signal path that just doesn't sound as good. i can't finish this and then reinstall, i have to reinstall first.
an os reinstall will cost me about a week because i have some scripts to update. but my hands are tied. i still have to determine consistency over the m-audio card via windows, but it's pretty clear at this point that the codecs fucked the registry and that's what i'm going to have to do to fix it.
i guess if i installed windows streaming stuff, it might have prioritized itself over asio and then not removed itself fully. best guess, anyways. if that's right then i should in theory be able to find all the right values in the registry and reverse them by hand. but, that would likely take longer to figure out than a reinstall will take.
the way it's cutting in and out is like it's being mapped wrong somewhere.