One of the Echo Nest attributes calculated for every song is ‘speechiness’. This is an estimate of the amount of spoken word in a particular track. High values indicate that there’s a good deal of speech in the track, and low values indicate that there is very little speech. This attribute can be used to help create interesting playlists. For example, a music service like Spotify has hundreds of stand-up comedy albums in their collection. If you wanted to use the Echo Nest API to create a playlist of these routines you could create an artist-description playlist with a call like so:
However, this call wouldn’t generate the playlist that you want. Intermixed with stand-up routines would be comedy musical numbers by Tenacious D, The Lonely Island or “Weird Al”. That’s where the ‘speechiness’ attribute comes in. We can add a speechiness filter to our playlist call to give us spoken-word comedy tracks like so:
It is a pretty effective way to generate comedy playlists.
I made a demo app that shows this called The Comedy Playlister. It generates a Spotify playlist of comedy routines.
You can look at the tempo distribution of songs by The Rolling Stones:
Or you can look at scatter plots that show 4 attributes at once (X, Y, size and color). Here’s a plot of all of Muse’s songs showing the energy, loudness, hotttnesss and liveness:
You can interact with the plots – click on a bar or point in a plot to listen to songs (via Rdio).
The app lets you explore across 11 different song parameters: energy, loudness, danceability, liveness, speechiness, hotttnesss, tempo, duration, key, time signature and mode. You can use the app to find all sorts of interesting things. Want to listen all the stage patter for an artist? Create scatter plot for the artist with liveness and speechiness as the X, Y parameters. The songs in the upper right-hand corner of the plot will be the ones you are looking for. Try it with an artist like Elvis Presley or Dean Martin.
Sunday at SXSW was the Artist’s Hack - where passionate developers from around the world gathered to build cool stuff. Artist’s Hack was organized by Backplane and Spotify and is dedicated to building the future of music, art, video and collaborative though on the web and mobile during SXSW.
The hack was held at Raptor House – a short walk from downtown Austin. There was plenty of bandwidth, good food and beverages for the 8 hour hackathon. APIs were in abundance: Spotify, The Echo Nest, SendGrid, Twilio, Youtube, Klout, Paypal, Gimbal, SeatGeek, Aviary, Etsy, Topspin, Chute, Dropbox, Music Dealers and others were all there in force offering their technology for hackers to use.
Hackers built around 20 hacks during the event. Some of my favorites are:
- biomuse – creates playlists based upon your biometrics. This was built on top of the biobeats platform. Quite neat stuff. Winner of one of the Echo Nest prizes.
- Jamblot – visualize your song history in a creative way to commemorate any period of your life that affected your music choice. Jamblot draws your song history for you. Winner of one of the Echo Nest prizes.
- Party Together - ambient automatic shared playlists for your party. Winner of one of the Echo Nest prizes.
- We browse in public – Stream all of your browser activity live to others. Chat with others based on their activity.
- Bundio – Monetize dropbox.
My hack is A longer life for post-rock fans. This was my first time using the Twilio API. It was a lot of fun to build. The Twilio API and whole developer experience is awesome. Any company with an API should try to emulate what Twilio does.
One novel aspect of the event was that Cory Booker was one of the judges. Here he is watching Danny Kirschner give the Bundio demo
Cory is a pro – when there was a power outage that delayed some of the demos, Cory conducted an impromptu ’interview’ with one of the founders of Backplane while the crew scurried to restore the power.
All in all, the Artist’s Hack was great fun, with lots of creative hacks. Well done Spotify and Backplane!
I like to listen to post-rock. Unfortunately, post-rock bands tend to have very long names like ‘Explosions in the Sky’, ‘God Speed you black emperor’, and ‘This will Destroy You’. I have a long commute and I will find that I am frequently risking my life trying to type a long band name into my music player. I wish Siri supported non-itunes players like Spotify, but until then I need a way to tell Spotify to play music by bands with long names. If I don’t, I will die in a fiery crash on Route 3 in Lowell Mass. A horrible way to go. So this weekend at the Artists Hack I built something to solve this problem. It lets you play music in Spotify without having to type long artist names. Here’s how it works.
I used Twilio to set up a phone number such that if you text it an artist name, it will respond with a spotify link to a song by that artist. You can add the phone number to your contacts as “music player”, You can then use Siri in a dialog like so:
Me: Send a text to Music Player
Siri: What would you like it to say
Me: Explosion in the Sky
Siri: OK, I’ll send it
A few seconds later I get a text message back with a link to a popular track by Explosions in the Sky. I tap the link and Spotify opens and plays the song. It is about as simple a hack as can be, but it solves a real problem for me. Here’s the screenshot:
If you want to use it – send a text with the artist name (and nothing else) to 603 821 4328. The code is on github. Update … much to my surprise this hack won two prizes at the hackathon – the Twilio 1st prize and the overall 3rd prize.
Meet Mandy. She’s a musician and a multimedia artist. She’s also losing her vision. She is on a quest – to push forward interfaces for non-visual music production to make it easier for those with vision impairments to use modern technology to create music.
We in the music tech community, especially the music hackers among us are in an excellent position to help Mandy take some steps toward reaching her goal. Anyone who has been to a Music Hack Day has seen the wide range of non-traditional music interfaces that we create. We’ve made Music gloves, leap-motion-based mixers and orchestras, invisible violins made with iphones, remixing tools that use makey-makey, music controllers made out of kinects, arduinos, webcams and even neckties and coat racks. We build things that make music. With a little guidance we should be able to build things that will help the visually impaired make music too. Mandy has just started a blog: Hacking Blindness where she is writing about her journey to help develop new and immersive ways for low vision and blind musicians to perform and manipulate. sound. If you are a music techie/hacker and are interested in learning more about this go to Mandy’s blog and join in the discussion. It is just getting started. I hope that in an upcoming Music Hack Day we can have Mandy come and help us understand how we can help her #hackblindness.
Last weekend, Barbara Duckworth and Jennie Lamere teamed up at the Tufts Hackathon to build a music hack. Here’s Barbara’s report from the hackathon:
Cinemusic – created at Tufts Hackathon
For our second hack day, Jen Lamere and I were wildly successful. Going into the Tufts hackathon, we knew that we wanted to create a hack involving music, but we didn’t want the hassle of having to make hardware to go along with it, like in our last hack, HighFive Hero.
As we were walking to the building in which the hackathon was held, we decided on making a program that would suggest a movie based on its soundtrack. The user would tell us their favorite artists, and we would find a movie soundtrack that contained similar music, the idea being that if you like the soundtrack, the movie would also be of your tastes. So, lets say you have an unnatural love for Miley Cyrus. Type that in, and our music-to-movie program would tell you to watch Another Cinderella Story, with Selena Gomez on the soundtrack. With Selena also being a Disney Channel star and of similar singing caliber, the suggestion makes sense.
We used The Echo Nest API to search for similar artists, and with the help of Paul Lamere, utilized Spotify’s fantastic tagging system to compile a huge data file of artists and soundtracks, which we then sorted through. We also added a cool last-minute feature using the Spotify API, which would start playing the soundtrack right as the movie suggestion was given. Jen and I hope to iron out any bugs that are currently in our program, and turn it into a web app.
Our (if I do say so myself) pretty awesome hack, combined with our amateur status, won us the rookie award at Tufts Hackathon! Jen and I will both be proudly wearing our new “GitHub swag” and we will hopefully find a way to put the AWS credits to good use. Thank you to everyone at Tufts, for organizing such a fantastic event!
Control your radio with your mouth
- Play music by Carly Rae Jepsen
- Play music like Weezer
- Play some brutal death metal
- Play some christmas music
- Play slow music by Beyoncé
- Play fast music by Beyoncé
- Play chill music in the style of smooth jazz
- Play some screamo
Pro tip – the artist or genre should always be at the end of your utterance.
The hack is an exploration of how well an off-the-shelf speech large vocabulary speech recognizer would work in the music domain. Music has lots of hard names like deadmau5, p!nk, !!! and many domain-specific terms like ‘screamo’, ‘hip hop’, ‘shoegaze’. I am actually quite surprised at how well this works. The Google speech recognizer does a good job at understanding most of the neologism like ‘screamo’ and ‘shoegaze’, and does an excellent job at recognizing popular artist names like Jay-Z and Beyonce. For unusual artist names, The Echo Nest artist search does a really good job of finding what you meant. So when the speech recognizer returns “play music by chick chick chick”, The Echo Nest artist search can turn the artist search for “chick chick chick” into “!!!” with no problems. Similarly the speech recognizer will return “dead mouse” which The Echo Nest will resolve to ‘deadmau5′.
We can also field more general music queries. If a style query returns no results, it is re-submitted as a general artist-description query. This lets you find more esoteric music “big hair bands”.
You have to grant the app permission to access the microphone for every utterance. This can be alleviated in the near future after a few API issues are sorted out. Until then, the app is all Cancel or Allow. (And yes, it is incredibly annoying). This is all sorted now.
This hack was built at the Tufts Hackathon 2013. For me, it was a half-a-hackday with lots of time spent supporting The Echo Nest APIs to folks who had never used it before and traveling in the snow. Still, it was fun to use the nifty new Web Speech API that just shipped this week in Chrome Version 25.