I've wondered if it's possible to train exact, frequency-based pitch, instead of this relative tuning-based absolute pitch. I wonder if that would be as discordant to the trained ear when switching to another tuning.