suppose you have discovered a new musical piece, and it turns out you like it.
I am looking for software that can help to answer the question: What makes this music good?
Besides pure listening (learning to analyze it by ear or just brute force listening to it many times) what software do you know that helps us to understand the music better? In any way.
There have been some algori(/y)thm aided statistical projects that has trawled through thousands of charts and provided data to the academics studying the field. Like how length, tempo, keys, intervals, structure have changed over time. Sorry, can't recall any details, only the gist of it but I believe it was used against commercial billboard charts and thus easier to define and rank than 'good'. Rick Beato's 'What Makes This Song Great?'
is Great since he has the master tapes to solo out and zoom in on precisely so there's a tool to learn from. He's very detailed and knowledgeable about harmonic structure, instrument choices and traits of the different songwriters. Still, his definitions of 'Great' are as arbitrary as anyone's. Someone on another channel with the same tools could just as well present a strong case for why Beato's song pick isn't that great.
Services like shazam, last.fm. spotify must also have some robots crawling around, with all kinds of parameters applied to the songs in their database and their listeners habits. Probably not something they share on github but maybe will license out to songwriters and taste makers one day.
Pretty sure that if something like that existed our pattern recognition ears would filter out the machine music from the human touched, creative, personal immediately.
Even when the Turing test gets passed and Watson has tracking data from millions of peoples lives to hook into, what should skynet gorillaz write good songs about?