Hacker Newsnew | past | comments | ask | show | jobs | submitlogin

Fascinating. I wonder if this can be applied to sounds as well ? What makes it more interesting with sound is we can "time travel" so to speak, and sample across hundreds of years of music where as with imagery, you're just stuck with what Google streetview can show right now..


Work on music (not necessary sounds) is carried out by the Music Genome Project (http://en.wikipedia.org/wiki/Music_Genome_Project). It's the engine behind Pandora. There's an incomplete list of attributes used in this classification here: http://en.wikipedia.org/wiki/List_of_Music_Genome_Project_at....

They still use manpower to classify all those tracks which makes it very hard to scale.




Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: