Wednesday Jan 16, 2008

Music sites like Last.fm and Pandora that offer you a customized radio based upon your listening tastes have been extremely popular over the last few years.  These sites are like listening to a radio station where the DJ knows exactly what kind of music you like.  These customized radio stations are great for general listening and as a way to discover new music.  But  just like traditional radio, if you in the mood for a particular song by a particular artist, you are out of luck - Internet radio providers have to follow a strict set of rules to make sure that listeners aren't too happy with their Internet listening.  You can't play an artist or track on demand and  the number of 'skips' are limited.  If you really want to listen to Deep Purple's "Smoke on the Water" you are out of luck.

In 2008, this is all going to change.  We are now starting to see the next generation of music apps hit the web that no longer are beholden to the rules of Internet radio. These  music apps allow you to play any song you want, at anytime for free. Some notable next generation music apps that offer play on demand are Finetune, Jango, The Hype Machine, Grooveshark, and  seeqpodSpotify (in private beta) is extremely nice.  Over the next 6 months, we will see even more music apps released that let you listen to any song you want.

In the book 'The future of music', authors Kusek and Leonhard write about a future where music 'flows like water'.  Where all songs are playable, on demand, anytime.  This is the Celestial Jukebox - the great jukebox in the sky that is ready to play any song you want.  2008 will be the year of the Celestial Jukebox - at least if you are sitting at your computer, You will be able to listen to virtually any song that you want, on demand.  The only time you will need to pay for music is if you want to listen to it on your iPod or your living room stereo.  And of course, in the not too distant future,  all of our devices will be connected to the net - when this happens, the Celestial Jukebox will really exist.  We will be able to listen to any song, at any time, at any place, on any device.

Tuesday Jan 15, 2008

If you are located outside of the US and you liked to listen to Pandora's wonderful radio, you've been out of luck as copyright licensing issues have forced Pandora to close its doors in non-US locales.  But now,  you can listen to Pandora via globalPandora.  GlobalPandora finds a way of getting around the great wall of copyright that surrounds the U.S.   Who knows how long it will last before it is turned off.  Enjoy it while you can.

 

At the end of February, I'll be flying out to San Francisco to attend the SanFran MusicTech Summit.  I'll be meeting up with the movers and shakers in the music 2.0 space.  It looks like all of the Bay Area players will be there including folks from Yahoo, YouTube, MusicIP, Pandora, Rhapsody, Mp3Tunes, Apple, MusicBrainz, iLike, Gracenote and more.  The event is being organized by Brian Zisk, the co-founder of the Future of Music Coalition.  I'll be involved in a panel on recommendations and playlists.  I am especially looking forward to meeting face to face, all of those smart folks that I have exchanged emails and phone calls with over the years.  It should be real fun.

Monday Jan 14, 2008

Last year, I wrote  a few posts about the top 5 things in the digital music work that didn't happen in 2006.  At the top of the list was "Apple didn't innovate".  In that post I wrote:

Apple, the undisputed leader in the digital music world, with its stunning 85% market share, with more than a billion tracks sold, and the sole supplier of the coveted 'perfect thing'  is spending less time on music and more time on video -  expanding its iTunes empire to include TV and movies. This focus on video leaves few cycles left in Cupertino to devote to improving the music experience. This year Apple gave us music listeners a few crumbs: gapless playback, a slightly smarter shuffle, an album art browser and a minimal text search capability on its latest iPod.   There are lots of things that Apple could have done but didn't.  Apple didn't offer an 'all-you-can-eat'  music subscription service. Apple didn't add music or taste sharing to iTunes or the iPod. Apple didn't improve iTunes music discovery tools  to help get listeners beyond the short head - much less into the long tail. Apple didn't do much to help us find music on our own iPods - an iPod becomes the place where a song goes to die - 60% of songs on a typical iPod have never been listened to by its owner.

Not much has changed in a year.  Apple gave us music listeners a few more crumbs: a touch screen interface, coverflow on the iPod, the ability to buy songs on the iPod, and some crazy Starbucks connection that we will never see in New Hampshire.  Probably the biggest thing Apple did this year was to promote DRM-free music - they managed to sign up EMI (I still haven't found in DRM-free tracks in iTunes), but quickly lost the DRM-free race to AmazonMp3.com (which is now selling DRM-Free tracks from all of the major labels).

Apple is still foregoing incredible opportunities.  Their recommendations really kind of suck , they don't try to do any of the wonderful things with social music that companies like Last.fm are doing.  They still  don't give us any tools to help us explore our music collection or to find new music in the iTunes store.

I'm writing this the evening before the big 2008 MacWorld keynote.  I'd really like to see Steve make some big announcements around music.   An all-you-can-eat music subscription service that works on the iPod is at the top of my list of things to hope for.  I suspect, however,  that I will be disappointed.  Steve conquered the music world 5 years ago, and he's moved on to bigger things.  My only real hope is that when the iPhone SDK is released, we'll be able to write our own music apps.

I've updated our tastebroker to include a webservice that will generate music recommendations from APML.  It uses the 'tagomendations' recommender to generate artist-level recommendations based upon your music profile in your attention data.  It takes a few seconds to create the recommendations for an APML, which is really quite impressive considering all of the recommendation work is done on-the-fly, and is not pre-calculated.  (And I'm not bragging,  because I didn't write that bit of the code - it was developed by the advanced search technology group here in the labs - Thx Steve & Jeff!).

This weekened, Digg pointed me to a new flash game called 'Filler'.  I was kind of fun, and I risked a bit of carpal tunnel playing a few rounds.   This morning, I signed onto facebook and was greeted with this notification:


This made me feel a bit queasy.  I didn't tell 'kongregate' who I was, nor did I tell facebook that I played the 'filler' game.  They just worked it out for themselves.  At least facebook is asking me whether or not it should post the story to my mini-feed... but I really think kongregate should have asked me before it told facebook what I was doing.

Wednesday Jan 09, 2008

As a brass player  I can relate and empathize with these music clips: The Hilarious Trumpet Blooper Page.  (and really this post is just an excuse to fool around with the new embedded media player from Lucas and Co. over at Yahoo).

(the embedded player is really quite nice. Well done Lucas!)

 

Tuesday Jan 08, 2008

I saw over at the Read/Write Web that Idiomag is now supporting APML.  In theory, you can give Idiomag a URL to your APML file and it will bulid a music magazine for you with articles, reviews, galleries and interactive features.   Since I happen to have created a web service that will generate an APML file based upon my music listening behavior (see tastebroker.org) , I thought I give it a try.   I plugged in the URL (  http://TasteBroker.org/apml/last.fm/lamere ). But alas, it didn't work - I received the Idiomag error page (Oops! Something seems to have gone wrong!).  

Perhaps the folks over at IdioMag can help drop me a line so we can figure out where things are going awry - what good is a standard for taste portability if it isn't really portable!

Update:  Ah, I forgot that 'tastebroker.org' was just a redirect to http://research.sun.com:8080/AttentionProfile.  If I use the true APML link of http://research.sun.com:8080/AttentionProfile/apml/last.fm/lamere everything works fine.  (and Ed Barrow from IdioMag comments that they've updated Idiomag to deal with the redirect as well.  Now that is nimble!)
 

Over on BoingBoing, there's a pointer to this YouTube video that shows Sgt. Pepper broken in to 4 distinct tracks.  The source separation is uncanny - and with anything that appears to be too good to be true, a bit of skepticism is warranted.  This video is not really the result of source separation - but instead it is from the studio itself - apparently you can get them from here: Sgt Pepper's Multitracks. Still, it is pretty neat to hear  Sgt Pepper broken down into  its constituent parts.   (also liner notes )



Monday Jan 07, 2008

Tis the season for 'best of 2007' lists - best new artist, best album, best track etc.  Taking  a cue from Stephan Bauman and his Emotional Music Recommendation work, I've decided to skip the usual categories and give my top music goosebump moments of 2007.  These are the moments where I had extreme emotional reaction to the music - where the hair on the back of my neck would stand on end.  These moments are few and far between, only 3 of them for 2007:

 

 What was your top musical goosebump moments of 2007?

Saturday Jan 05, 2008

This morning I took 3 of the kids to the local high school to see and hear Presidential candidate Bararck Obama.  While waiting for the candidate to arrive, the 3,000 of us were treated to the Obama playlist.  It was  a pretty good set of songs with very appropriate themes for the Obama candidacy.  Here's the playlist (as reported in the Boston Globe):

1. Marvin Gaye, "Ain't No Mountain High Enough''
2. John Parr, "St. Elmo's Fire (Man in Motion)''
3. Tina Turner, "The Best''
4. The Doobie Brothers, "Takin' It To The Streets''
5. Earth, Wind & Fire, "Shining Star''
6. O'Jays, "Give The People What They Want''
7. Sam and Dave, "Hold On I'm Coming''
8. Kool & the Gang, "Celebration''
9. Natasha Bedingfield, "Unwritten''
10. The Isley Brothers, "Shout''
11. The Temptations, "Get Ready''
12. India.aire, "There's Hope''
13. McFadden and Whitehead, "Ain't No Stoppin' Us Now''
14. Staples Singers, "I'll Take You There''
15. Orleans, "Still The One''
16. Sly and the Family Stone, "Everyday People''
17. The Doobie Bros., "Long Train Running''
18. Stevie Wonder, "Sir Duke''
19. John Fogarty, "Centerfield''

These songs  have strong  themes of hope and change - hallmarks of the Obama campaign. 

(And if you are a Spotify user you can listen to Obama's Playlist online.

 

Thursday Jan 03, 2008

Music 2.0 companies that rely on web APIs of other companies are always at risk that the APIs that they depend on will be turned off.  If you interface with iTunes, Amazon, Yahoo, Rhapsody or many of the other music providers your business could be shut down if that company decides to alter or disable the API.   This has just happened to YottaMusic.  YottaMusic provided a superior interface to the Rhapsody music service that added significant value for Rhapsody subscribers.  YottaMusic made Rhapsody a better place.  However, according to TechCrunch, Rhapsody has shut down the non-public APIs that YottaMusic has relied upon.  Without these APIs, YottaMusic can't provide any extra value, and so it has shut its doors.   On the face of it, it seems crazy that Rhapsody would treat YottaMusic so poorly -  I can only surmise that Rhapsody wants to retain total control of the user experience or they want to be able to license access (i.e. for money) to 3rd parties.

This should serve as a warning to all of those music 2.0 companies that rely on the APIs of others.  Your livelihood could be taken away overnight when a critical API is eliminated.
 

Wednesday Jan 02, 2008

I was spending a quiet evening browsing through last.fm social tags when I stumbled across the tag: 742617000027.  Clearly this was some sort of bug. Some underlying coding constant poking through the web veneer.  Something akin to GLaDOS saying "The device is now more valuable than the organs and combined income of everyone in [Subject Hometown Here]" .     But this tag is not a bug.  It is the name of a song by Slipknot.  According to songfacts, 742617000027 is the intro track to Slipknot's self-titled CD. It contains some guitar scratches and weird sound samples.  " 742617000027 was the shipping code on their 1996 self-released album, Mate, Feed, Kill, Repeat. All the band members wear that number on their jumpsuits.

There's just no end to the wide variety of music metadata one can encounter.  But I must wonder what it is like for a hardcore Slipknot fan.  In my day, we could shout out 'freebird' or 'stairway' - but imagine trying to shout out '742617000027 in a concert.


Thursday Dec 27, 2007

Lucky me. My lovely wife gave me an iPhone for Xmas. So while I wait for new tires to be put on my car I can blog about it. Good for me. Not so good for you, dear reader.

Wednesday Dec 19, 2007

If you go to one of the many social music sites out there and get 'similar artists' recommendations for Jimi Hendrix.  You are likely to get a list such as the one you get from Last.fm:

 

There's no arguing that this is a good list - but it is also a rather diverse list.  Eric Clapton's blues guitar is quite different from the psychedelic acid rock of the Doors. I'd really like to know a bit more about the recommendations - in particular I'd like to know why a particular artist was recommended. This can help me gain trust in the recommender as well as help direct me to artists based on criteria that  are most relevant to me.  Unfortunately, most recommenders are based strictly on consumption habits, so the only recommendation explanation available is the Amazonian "People who listened to Jimi Hendrix also listened to Eric Clapton and the Doors"  - which is not too helpful for me.

We want to make recommendations transparent - so that you can ask 'why did you recommend this' and get a useful answer beyond the typical 'people who bought X also bought Y'.  In order to generate transparent  recommendations you need to have some understanding of the content - there has to be some way of knowing that Hendix tracks typically contain distorted guitar, for example.  Companies like Pandora, with their music genome project have spent years analyzing to hundreds of thousands of tracks, assigning 400+ attributes to each track. This lets Pandora give those great explanation that makes them so popular with their users:

Pandora can generate these excellent explanations because they've taken the time (and spent lots of money) to listen to and extensively label all of their music.  Most companies won't have the time, money or patience to do this - and even Pandora that is committed to this approach, can't keep up with the volume of new music that is generated every year.  Luckily, there are other sources of content description that we can use to generate recommendations.

One such source are social tags.  Social tags have been all of the rage in the web 2.0 world.  Sites like del.icio.us, Flickr and Last.fm demonstrate how the such tags can be extremely useful for searching and organizing content, especially non-text content.  Social tags can also be used to give us good transparent recommendations.  

Here are the most frequent social tags applied to Jimi Hendrix at Last.fm:

 

Now tags like 'rock' and 'blues' occur rather frequently for many artists, so are less descriptive than tags like 'guitar god' which occur infrequently across artists.  So we can take this into account and generate a list of the most distinctive tags for Jimi Hendrix.  These are tags that occur rarely across the entire set of artists, but occur frequently for Hendrix.  (This is the classic TF-IDF term weighting technique).  Here are the distinctive tags for Hendrix:

The less descriptive tags such as 'rock' have fallen off the list while extremely targeted tags like 'jimi hendix' and 'acid rock' have risen to the top.  This list is a much better description of Hendrix than the 'Frequent Tags' list and can serve as the basis for our transparent recommendations.

With these distinctive tags, we can generate recommendations based upon the cosine distance of these distinctive tags to the distinctive tag sets of other artists.  This type of recommendation, based upon the similarity of distinctive tags, gives us surprisingly good results.  My colleague, and resident neologist, Steve, coined the term 'tagomendations' for recommendations based on the social tags.  Here are the tagomendations for Jimi Hendrix (ordered by artist popularity):

Interestingly (and surprisingly) our tagomendations compared favorably to recommendations generated with the more traditional collaborative filtering techniques when we evaluated them in a survey.  And of course, with these tagomendations based upoon distinctive tags we can now explain why a recommendation was made, based upon how the distinctive tags for two artists overlap.  

For instance, for the Clapton recommedation, we can look at how the distinctive tags for Clapton and Hendrix overlap:


Clearly if you like Hendrix because of his blues guitar playing, you might want to give Clapton a listen. Compare this to the overlapping tags of The Door and Hendrix:

This is quite a different vibe - with the focus on psychedelia of the 60s.

These tag clouds showing how the distinctive tags for recommended artists overlap with the seed artist give me the opportunity to explore the recommendations based on my taste.  If I like Hendrix because I am a fan of face-melting guitar, I will quickly find artists, like Joe Satriani, Gary Moore, or Steve Vai - but if the reason I like Hendrix is because of the 60s, psychedelic vibe, I may find Jefferson Airplane or Steppenwolf more to my taste.

That's it in a nutshell - how we are using social tags to generate transparent,  explainable  recommendations.   And by the way, to do the heavy lifting for our Tagomendations  we are using the text search engine called Minion, developed by the Advanced Search Technology group here at Sun Labs.  Minion is a high quality, highly configurable search engine that is perfect for doing these types of experiments.  Look for Minion, It's coming to an open source repository near you very soon.

Likewise, we hope to release our web-based music explorer that can generate transparent tagomendations soon.  Here's a little bit of what it looks like:

 


This blog copyright 2010 by plamere

Yahoo! Media Player

Playback Controls

Previous Play Next
Vol
Testing
Hahahaha
Yahoo! Media Player
EngineersDesignProduct
Mike DavisLino WiehenLucas Gonze
William KhoeDouglas KimDave Warmerdam
Amit BehereSuman Nichani