I gave up on ffmpeg and switched to sox_ng and now I've got it decoding, but is the quality perfect? would a slightly different set of options make it sound better?

I DON'T KNOW! I DON'T KNOW HOW TO KNOW

to make this better, I either need to listen to a bunch of WAVs with different ADPCM decoding options, OR learn the internals of the PS2's audio chip AND MIPS assembly.

To give you an idea of how much I hate audio, I am seriously considering the latter

I also can't compare two audio files! how can I tell which is better?

I listen to file 1, then I listen to file 2, but by the time I listen to file 2 I have forgotten file 1.

I can easily visually diff two images or show them side by side but I can't do that with audio

so using sox_ng with -r 24k -c 1 -t ima -e ima-adpcm seems to be "close enough".

Comparing to playing the game in the emulator, it's seems Very similar in shitiness.