Hey Mr. DJ: A Wi-Fi Playlist Tailored to You
A group of enterprising young researchers in UCLA's graduate Computer Science department have discovered a fun new way of using your iPod playlist: Wi-Fi enabled software sends your music preferences to whatever computer is nearby. Then music tailored to your preferences pumps out of the speakers -- can you imagine walking into a coffee shop, and the Carrie Underwood track playing shuts off, and a Matthew Good album slides into the rotation instead? How awesome! What a great way to go about your day!
The research team is made up of UCLA Ph.D. students Kevin Eustice, V. Ramakrishna, and Nam Nguyen, and Dr. Peter Reiher is of the Computer Science Department is acting as advisor and co-author. Dr. Reiher gave the Smart Party presentation at CCNC (Consumer Communications & Networking Conference) last week, and the research has been funded by the NSF.
the UCLA team has set up a prototype Smart Party system in three offices of the university's computer science department. This system can respond to playlists stored on notebook computers and, in future, should work with portable music players. Since it can detect the proximity of people by triangulating wireless signals, when someone has left a room their playlist can be removed from the musical ballot to reflect the music of the remaining occupants.
The system is democratic too: "In our current implementation, all votes are equal – one device might propose heavy metal, another pop," Eustice says.
Since the blog reaction has been so mixed, we thought we'd catch up with the guys over at UCLA and get their explanation of why this might be such a cool technology.
LAist: The most immediate question is, of course, how would a computer access music that is not already on its own hard drive? Can you expand on how you would deal with DRM issues?
Eustice: DRM can be dealt with in a couple different ways. 1) We could only use unprotected content. Several studios have announced that they will release DRM free content. The Smart Party would still have to abide by licensing and public performance restrictions (where appropriate), but those are less thorny to deal with. 2) Instead of pulling content from party-goer devices, we could pull it from a home server, and just use suggestions and preferences from the attendee's devices. Less interesting, to be sure, but it does avoid the problem. 3) Temporary, i.e. one time, licensing of the audio playback hardware
via the user's device. Currently, software like iTunes allows you to license another device to play back protected content, and then later revoke that license. We could extend that model to support one-time authorization of party hardware to playback protected content.
There are likely other similar techniques we could use.
LAist: How do you envision this being used in a social situation? Is it specifically for parties, like some of the blogs have been saying?
Eustice: A number of blogs have railed on us for trying to replace human DJs at clubs--that's not really the point of the Smart Party. The Smart Party is more intended for homes, offices, gyms, restaurants, or other types of venues in which we'd like to have more democratic music selection. Home parties are definitely one type of event we're interested in, but we don't anticipate replacing brilliant DJs who are doing live mixing of tracks with a simple voting algorithm!
My ideal vision of the Smart Party goes beyond what we've currently done. Right now, the Smart Party is designed to select and play music based solely on guest preferences. We've talked about a number of improvements and different directions to take the party. One interesting aspect is exposing guests to new songs they might like--straight up preferential voting doesn't really let that happen.
So, we've been looking at forming "taste" groups, by letting my media device discover others nearby with some overlap in taste--presumably there will be some "surprises," as different users will have been exposed to different music. If we can draw out some of those interesting new songs, and users actually like them, we'll have done well.
I'm also really interested in the idea of music as an important element in the social & location equation. If we can watch users and preferences move around a group of physical places, perhaps we can actually influence users to move around by changing the types of music that play in nearby rooms, to draw them in. There's a lot to play with in this space.
LAist: How are the iPod and other portable music devices going to be integrated into this?
Eustice: A Wi-Fi enabled media player, or Wi-Fi-enabled cell phone that can
store media is the ideal device, since it's inobtrusive, and has reasonable battery life. We view that type of device as the user's avatar at the party--it visits virtual versions of the places that the
user visits, and represents the user and their interests.
LAist: What about security or hacking issues?
Eustice: Oh, there are a variety of security issues in our prototype that would
need to be ironed out before this could actually be a product. Right now, when the Party tries to find a song provider, we trust that media metadata are correct. A user could technically upload any song they desired with incorrect metadata--the Party would not detect this, and would play the song. This is resolvable, however, with some effort.
LAist: Why was this sort of technology a priority? Why do you think it's been catching on in the blogosphere?
Eustice: We're really interested with technology-mediated social interactions in physical spaces. The Smart Party is part of Panoply, an NSF-funded project looking at secure infrastructure for ubiquitous computing--a paradigm in which computation and network capabilities are embedded in devices all around us, and in which computers transparently and
flexibly support human interaction.
Really though, we needed applications to evaluate our infrastructure. The Smart Party was a fun idea, and it was easy to get people to work on it because of that. As to why it's catching on in the blogosphere, I think the time is ripe for this type of technology. The Zune and iPod Touch are almost there--I think either Apple or Microsoft could easily go in this direction with a little effort
You can read the research abstract here.
Photo by wili_hybrid via Flickr