• Home
• About

Just a quick post before it is demo time.  This weekend at MIDEM Hack Day, I teamed up this weekend with the famous Mr. Doob  to build a music hack. We created the Midem Music Machine. It creates a beautiful visualization of music using The Echo Nest analyzer and Three.js.  Here’s a pic:

As you can see, our hack was inspired by the Animusic folks. Working with Mr. Doob was awesome. He did just amazing stuff.

You can see the Midem Music Machine online here:  Midem Music Machine.   You’ll need a browser that supports WebGL like Chrome.

We’ve doubled our floor space here at the Echo Nest. I now have an office with a door and a window that opens. Look at that desk. I’ll be winning the clean desk award every day for the next week at least!

In his blog post Can we kill the music business too?   James from songspin.fm has the magic formula to kill the major labels. He says:

In a nutshell, to kill the major label run music industry, startups will need to:

1.  find great music from people who aren’t assholes
2.  let people do cool things with that music
3.  let users share what they create
4.  profit!

By assholes, I mean people who will sue you for using their music in your startup, which probably makes this first step the hardest. You can’t have a great music startup without the music and more than that you need good music.

(Note that in that last quote, the first ‘sue’ link points to Grooveshark)

We are assigned a predetermined amount of weekly uploads to the system and get a small extra bonus if we manage to go above that (not easy).The assignments are assumed as direct orders from the top to the bottom, we don’t just volunteer to “enhance” the Grooveshark database.

All search results are monitored and when something is tagged as “not available”, it get’s queued up to our lists for upload. You have to visualize the database in two general sections: “known” stuff and “undiscovered/indie/underground”. The “known” stuff is taken care internally by uploads. Only for the “undiscovered” stuff are the users involved as explained in some posts above. Practically speaking, there is not much need for users to upload a major label album since we already take care of this on a daily basis.

Are the above legal, or ethical? Of course not. Don’t reply to give me a lecture. I know. But if the labels and their lawyers can’t figure out how to stop it, then I don’t feel bad for having a job. It’s tough times.

Why am I disclosing all this? Well, I have been here a while and I don’t like the attitude that the administration has acquired against the artists. They are the enemy. They are the threat. The things that are said internally about them would make you very very angry. Interns are promised getting a foot in the music industry, only to hear these people cursing and bad mouthing the whole industry all day long, to the point where you wonder what would happen if Grooveshark get’s hacked by Anonymous one day and all the emails leak on some torrent or something.

The Echo Nest engineering team just pushed out a new feature giving you more control over the artist makeup in playlists.  There is a new parameter to the playlist/static API called distribution that can be set to wandering  or focused.   When the distribution is set to wandering the artists will appear with approximately equal distribution in the playlist. If the distribution is set to focused artists that are more similar to the seed artists will appear more frequently.  When combined with the variety parameter, you have excellent control over the number and distribution of artists in a playlist.  If you want to create a playlist suitable for music discovery, create a playlist with high variety and a wandering distribution.  If you want to create a playlist that more closely mimics the radio experience choose a low variety and a focused distribution.

I’ve put together a little demo that lets you create playlists with different levels of variety and distribution settings. The demo will create a playlist given a seed artist and show you the artist distribution for the playlist.  Here’s the output of the demo with distribution set to focused:

You can see from the artist histogram that the playlist draws more from artists that are very similar to the seed artist (Weezer).  Compare to these results from a wandering playlist with the same seed and variety:

You can see that there is flatter distribution of artists in the playlist.   You can use variety and distribution to tailor playlists to the listener.  For instance, you can give the Classic Rock Radio experience to a listener by setting variety to relatively low, setting the distribution to focused and seeding with a classic rock artist like Led Zeppelin.  Here’s the artist distribution for the resulting playlist:

That looks like the artist rotation for my local classic rock radio.

Give the demo a try to see how you can use variety and distribution to match playlists to your listener’s taste.  Then read the playlist API docs to see how to use the API to start incorporating these attributes into your apps.

The Demo:  Playlist Distribution Demo (source)

This week Twitter and The Echo Nest announced a partnership where  Twitter IDs for verified accounts are incorporated into The Echo Nest’s ID mapping layer (aka Rosetta Stone).  This makes it easy for developers to get the Twitter handle for an artist.   To demonstrate just how easy it is, I wrote a little web app that displays the top 100 artists that have  verified Twitter accounts.    Here’s the core bit of code for the app:

function fetchTopTwitterArtists() {
    var url = 'developer.echonest.com/api/v4/artist/top_hottt?callback=?';

    $.getJSON(url, { 'api_key': 'GETYOUROWNAPIKEY', 'format':'jsonp',
      'results': 100, 'bucket': ['hotttnesss', 'id:twitter'], 'limit': true},
     function(data) {
       for (var i = 0; i < data.response.artists.length; i++) {
          var artist = data.response.artists[i];
          var elem = $("<li>");
          var link = $("<a>");
          var handle = artist.foreign_ids[0].foreign_id.replace('twitter:artist:', '');
          link.attr('href', 'twitter.com/' + handle);
          link.text(artist.name);
          elem.append(link);
          $("#results").append(elem);
       }
     }
  });
}

The key bits here are creating the artist/top_hottt request to The Echo Nest and  adding the id:twitter bucket and setting limit to true. This tells the Echo Nest to include the twitter handle information in the results and limit the results to only those artists that have twitter information.  After that it is just pulling the data out of the results and formatting it for the lovely display.   The Twitter ID info is returned in an ID block that looks like this:

            {"catalog": "twitter","foreign_id": "twitter:artist:LMFAO"}

Note that the Twitter ID is returned in a URN form. To get the actual Twitter URL for an artist we just need to replace the ‘twitter:artist:’ bits with ‘twitter.com/’.

You can see the app here:  Top 100 artists with verified Twitter accounts.  As you can see, I tried to make the web app as ugly as possible. The only thing it needs is some Comic Sansification.  The code is available for detailed study in this gist.

The WordPress.com stats helper monkeys prepared a 2011 annual report for this blog.

Here’s an excerpt:

London Olympic Stadium holds 80,000 people. This blog was viewed about 390,000 times in 2011. If it were competing at London Olympic Stadium, it would take about 5 sold-out events for that many people to see it.

Click here to see the complete report.

Here’s a video of an early prototype of a 3D Music Explorer that runs as a Spotify App.  The visualization shows a series of interconnected Echo Nest playlists that you can explore.  You can fly through your music,  listen to anything you see, and create a new playlist starting at any point in the visualization, all from within the Spotify App.

I’m using the awesome three.js for the 3D bits, all the rest is Echo Nest and Spotify APIs.  Frame rates with 500 albums showing is  60 FPS   Note that there’s a strange interaction between quicktime (that I used for screen capture) and Spotify audio – I wasn’t able to play new songs in Spotify while recording the demo, so there’s no song changes in the video.

On Quora, someone asked the question: Are band names getting longer? To answer that question I used the Echo Nest API to get the top 500 artists for each five year window starting in 1950 and calculated the average name length for each of the artists.

The answer to the question as to whether artist names have been getting longer is a surprising No. The average length of artist names peaked in the period from 1955-1959. This was the age when bands were called ‘XX and the YY’ like Van McCoy & The Soul City Symphony and Herb Alpert & The Tijuana Brass. Surprisingly, the era of the shortest names is 1990-1994 . The bands “… And you will know us by the trail of the dead” (1994) and “The Presidents of the United States of America” (1993) were really bucking the trend.

And the longest name, and my favorite is 2010′s Tim and Sam’s Tim and the Sam Band with Tim and Sam .

Here’s a chart showing the mean and maximum length over the years.

Overall Longest Names

Here are the longest artist names and their start year:

• 51 2010 Tim and Sam’s Tim and the Sam Band with Tim and Sam
• 48 1994 …And You Will Know Us By The Trail Of The Dead
• 47 1967 Charles Wright & The Watts 103rd St Rhythm Band
• 46 1993 The Presidents of the United States of America
• 44 1959 Van McCoy & The Soul City Symphony Orchestra
• 43 2000 Richard Cheese & Lounge Against The Machine
• 43 1950 Benedictine Monks of Santo Domingo de Silos
• 41 1981 Emir Kusturica & The No Smoking Orchestra
• 39 1972 Afrika Bambaataa & The Soul Sonic Force
• 37 2010 Antoine Dodson & The Gregory Brothers
• 37 2005 Edward Sharpe and the Magnetic Zeroes
• 37 1999 Someone Still Loves You Boris Yeltsin
• 36 1977 Gloria Estefan & Miami Sound Machine
• 36 1970 Jonathan Richman & The Modern Lovers
• 36 1964 Captain Beefheart and his Magic Band
• 35 2010 RPA and the United Nations of Sound
• 35 2005 Edward Sharpe & The Magnetic Zeroes
• 35 1975 Johnny Thunders & The Heartbreakers
• 35 1959 Academy of St. Martin in the Fields
• 34 2010 Noel Gallagher’s High Flying Birds
• 34 1978 Deutsch Amerikanische Freundschaft
• 34 1959 Van McCoy & The Soul City Symphony
• 33 2010 RPA & The United Nations Of Sound
• 33 1995 Robbie Williams And Jane Horrocks
• 33 1978 Orchestral Manoeuvres in the Dark
• 33 1974 George Thorogood & The Destroyers
• 33 1965 Big Brother & The Holding Company
• 33 1957 Cliff Bennett & The Rebel Rousers
• 33 1955 The Clancy Brothers & Tommy Makem
• 32 2010 The Ghost Of A Saber Tooth Tiger
• 32 2007 The Pains of Being Pure at Heart
• 32 2005 Margot & the Nuclear So and So’s
• 32 1990 Emily Haines & The Soft Skeleton
• 32 1967 Kenny Rogers & The First Edition
• 32 1962 Mitch Ryder & The Detroit Wheels
• 32 1961 Patti LaBelle & Michael McDonald
• 32 1960 Frankie Valli & The Four Seasons
• 32 1960 Astor Piazzolla & Kronos Quartet
• 32 1959 Maurice Williams and The Zodiacs
• 32 1954 The Sensational Alex Harvey Band
• 31 1992 Jools Holland And Stereophonics
• 31 1990 Armand van Helden’s Sampleslaya
• 31 1977 Erste Allgemeine Verunsicherung
• 31 1976 Tom Petty and The Heartbreakers
• 31 1966 Nancy Sinatra And Lee Hazlewood
• 31 1964 Björn Ulvaeus & Benny Andersson
• 31 1963 The Paul Butterfield Blues Band
• 31 1963 John Mayall & The Bluesbreakers
• 31 1962 Martha Reeves and the Vandellas
• 31 1957 Herb Alpert & The Tijuana Brass
• 31 1950 Frank Zappa & Captain Beefheart

I just finished my holiday break coding project. It’s a 3D music maze built with WebGL that runs in your browser.  You can wander around the maze and sample music and enjoy the album art.  It’s like being in 1992 with a fresh copy of Castle Wolfenstein 3D – but with music instead of Nazi