4 ms·
I was a QA intern at Echo Nest in Somerville over the summer and was super surprised to see this today! They're a really a fantastic and fun group of people and
by daturkel 13y ago
I was a QA intern at Echo Nest in Somerville over the summer and was super surprised to see this today! They're a really a fantastic and fun group of people and doing incredible work, having a fun time while they do it.
(Speaking, naturally, only my own opinions here:) I'm excited for the possibilities of growth that it'll offer the company, but also a little nervous about how it'll affect the Nest's ability to serve their other (previous?) clients. I personally signed on with Rdio because I loved the deep Echo Nest integration—I hope that's not going anywhere.
Looking hopefully, but anxiously, towards the future.
- noname123 13y agoHi, I'm curious what are the other applications of Echo Nest API besides music recommendation? Can it assist composers or performance artists in composition? (e.g., input scores of Charlie Parker, output a solo on key of X based on previous known solo's). Can it identify harmonic structures in music? (e.g., which mode is used in song X and in a composition, when a key is changed). I also live in the greater Boston area and has always viewed EchoNest with curiosity, so very curious what you can do with their API other than their "acoustic fingerprinting" (e.g., genre/bpm/key/mood/color etc.)
- sebastianwolff 13y agoNope, though you may be interested in David Cope and Experiments in Musical Intelligence. Input Bach MIDIs, receive 500 compositions in the style of Bach. Turing complete. Echo Nest has a few other fun features, like danceability, mood, tempo, and others. An API playground, too, for now.
- diydsp 13y agoThe API has some excellent stuff regarding audio segmentation. It represents any track as a list of ~2,000 segments. Each segment has energy in 12 chromatic bins measured, too. There's also information on beat times, and a Machine-Learning-based estimate of verse/chorus/solo and even more great stuff like that. Full on auto-transcription like you hint at is a hard AI problem, like speech recognition, which will never by fully solved (according to me), but I, for one, would LOVE to see a musical editor enhanced by EchoNest metadata. One of these days, i'll add to to Audacity hahaha. BTW, if you are in Boston area, come to the Music-Tech meetup. It's a "gem" and I'll be happy to intro you to the EN API (I worked there 6 years ago as a consultant).