Last week I discovered what I think may be the holy grail for online music and playlisting. It's called The Echo Nest, and they're doing what I've been dreaming about for a decade now. There's no denying that online digital music has made it possible to easily pull playlists together, in a plethora of new and exciting ways, but it has mostly been a manual slog. The missing link? Having robust and detailed song-level metadata that is easily searchable. Want to create a playlist of songs in 10/4 time, or songs produced by John Boylan? You better be ready to sit down and spend some serious time on research. I understand this first hand as I've done it for hundreds of playlists here at the Y! Music Playlist blog. But The Echo Nest is finally building the platform that will take music archiving, search and discovery into the future -- we're talking Star Trek stuff here.
Here is a description of what The Echo Nest is doing as pulled from their website:
"The Echo Nest platform is powered by the world's first machine learning platform for music. Based on 12 years of R&D at MIT, Columbia and Berkeley, the Echo Nest's music intelligence platform combines large-scale data mining, natural language processing, acoustic analysis and machine learning to automatically:
LISTEN TO MUSIC
We analyze every song on the web to extract key, tempo, rhythm, timbre and other attributes — to understand every song in the same way a musician would describe it (i.e.,"swing groove, tempo = 100 BPM, 4/4 time, key of B Flat, mezzo piano").
READ ABOUT MUSIC
We crawl the web and analyze every blog post, every review, every piece of text that describes music to understand in real time what the world knows about every artist, release and song."
Computer?
Yes Captain?
Play all top 40 songs from the 1970's in the key of C that are also critically acclaimed.
Yes Captain, playing in 3...2...1..."
The possibilities are endless, and this future is quickly approaching thanks to the work of companies like The Echo Nest.
This playlist is a fine example. Echo Nest technology was able to data mine a list of songs that have a "slow build", that is, songs that begin softly, and slowly build in volume. This is a great technique to build explosive crescendos. And they've even charted the top slow build tracks as pulled from the Million Song Dataset, so you can visually see how the volume changes during each song.
Listen to this playlist and pay special attention to how the volume builds in each song. This is just one small example of how music can be searched and discovered with new technology like this.
What slow build songs would be on your list?
Songs With A Slow Build
2. The Fear Has Gone - Engineers
4. India - The Psychedelic Furs
5. Positive Jam - The Hold Steady
6. Let It Die - Foo Fighters
7. The Garden (Part I) - Tea Leaf Green
8. Intro: A Song Of Fire And Wine - ...And You Will Know Us By the Trail of Dead
9. Little Susie / Pie Jesu - Michael Jackson
10. Remember - Ella Fitzgerald
11. Song Of Praise - John Coltrane
12. Fire And Water - Free
13. Truths Arise - Red Sparowes
14. Section 6 & 7: Hanging Around The Day, Part 1 & 2 - The Polyphonic Spree
15. Asphalt (Interlude) - Amusement Parks On Fire
16. Waking the Fallen - Avenged Sevenfold
17. To All The Girls - Beastie Boys
18. Miss Teen America - Reagan Youth
19. Dream Machine (Intro Piece) - Alphaville
20. In Books - Apex Theory
Follow what Robert is listening to on Twitter