MusicIP..behind the idea

SXSW Music has a completely different vibe to the Interactive Conference. Many people don’t come out till dark, some streets are closed off, and music assaults your ears from all directions. My chance to stay for the extra time came through Rick Segal and MusicIP, who covered the cost of the music ticket for me.

I’ve being using the client software in its previous incarnation on and off for a while, and have being playing with the new version this week – I’ll post about that experience later. This week, they’ve had a large stall in the Trade Show and have launched a tie up with MusicBrainz and the release of musicdns.org

I caught up with Mathew Dunn, the President and CEO, and Stephanie Phillips, VP of Marketing, of MusicIP and had a chance to ask them a few questions about the other parts of the business.

The MusicIP team is 14 strong, located in various places across North America. As a virtual team, they rely heavily on Skype and iChat for communication; this will come more into play as they expand over the next year. Customers have already signed up from Latin America, UK, India and Singapore. Furthermore, they are working closely with a number of manufacturers, in China and other Far East countries, to implement their firmware on music players. As they expand, they will have to grow staff to handle the growth in other areas – so starting as a virtual team will have its advantages.

If you take a look at their website, MusicIP focuses on three markets – the listeners, the musicians and the industry. For the first 2, it is kind of like a dating service; a search engine tool to bring new music to listeners based on their preferences and a chance for musicians to get their music out to new ears. As with blogs, there can be an audience for everything but the challenge is discovering the new stuff – without having to rely on the confinement of words and metatags.

Anyone can load up a track; they need to register, pick a licence and up it goes. The track is analysed and fingerprinted and then goes into a the database, which is currently at 16 million tracks and growing at around 1 million/month. The fingerprinting is based on the digital file and sound analysis – not metadata. SO the service has the advannage of being language independent. I no longer need to search based on tags in the one language I know; using the service could find me music from multiple countries if they are sonically similar. This is the strength of the discovery part – it moves away from textual search engines.

The client software analyses the music on your hard-dirve. When I first looked at this, it took a loonnnggg time to chugg through the tracks. As the database has grown, there is less need to do that; the track can be looked up on the central database and the properties stored. It only needs to look deeply at tracks it has not seen before; there’s no record of who has what so privacy is maintained. the discovery happens when I decide to create a playlist from my tracks; it offers a selection of others from the larger database that fit the mood and tone.

The musicians they have spoken to this week have loved the idea, but they will be working on making the submission process more streamlined – it needs to be simple for even the complete technophobe. At the other end of the scale they’ll also be improving the the industrial scale registration; whilst maintaining the quality and attribution checks that they have. With the fingerprinting process, it is realtively easy to spot if a small artist has loaded up the latest Madonna record as their own!

I liked the licencsing approach, using Creative Commons. This means some music can be offered as downloads for discovery. For the website and literature, many of the images they have used were also CC licenced – they found the stock photography too rigid, not showing the passion that could be found in fan photos. This philosophy also reaches into the technology area – inthe week that the patent was granted on the fingerprinting sofware, they released it as open-source for people to use and build on.

For businesses they are offering a tool that will allow the labels, the shops, the bands to sell more music. Allowing the discovery and mining of the long tail. Currently, a lot of the personalised recommendations are driven from what you have purchased before (or explore) and are base don metadata, not sound. So you are restricted to a subset, where the tags have been set by someone else. By using this tool, the search can be broader, based on sound preferences, not just other peoples descriptions.

Both the client and hte web based tool that is out in a few weeks, will link to sources of music that is being discovered, giving opportunity for you to buy stuff that you may not have normally found. The toolset acts as an aggregator of sources, so whether a track is found on Amazon, CDBaby Waterloo records or the bands own site, it can direct you to the right place to obtain the track.

The web-based search engine looks fun…from a starting point you can continually explore and dive into genres and songs, play with the stuff to find things you like.

THe company was originally established in 2000, although R&D had started before then. During the 6 year journey, the technical offering has expanded at leat 4 fold, the application has become more accessible with registration now open and they’ve spent a lot of time understanding and explaining the offering – getting the branding right. SXSW is seen as a turning point and a kick start to a the next phase.

They believe that they are only service that offers the connection point between listeners, musicians and the industry; by providing the internet with ears and becoming a standard and a marketplace for expression. The complement services such aslast.fm, offering a different way to find music. So how will they know they have been successful? This time next year they want to be explaining less; have musicians know that MusicIP has bought them new listeners and have success stories; and be on the cover of Wired Magazine. We’ll just have to watch and wait.