A Brief Peek at Predictive Playlists

I took an extra 512MB stick from an unused laptop for temporary use with mine while I wait for mine and so, I had the chance to multitask again and played around with Winamp. I found this Predixis MusicMagic Mixer at the bottom of the Media Library. I tested how good it was with mixing songs of the same genre and mood. A little inspiration since my work involves intelligent systems (although I haven't contributed to that much to that part ...yet). First test was from the jazz oldies, Dizzy Gilespie's 'Dizzy the Duck'. The results were pretty impresive, similar artists that showed up were Stan Getz, Charles Mingus, Sergio Mendes, Miles Davis, Duke Ellington and Johny Hodges. My guess on the keywords are 'brass', 'trumpet', 'sax' and 'jazz'. Second test, easy/smooth jazz, Acoustic Alchemy's 'Say Yeah'. Wow, this is great. Very close, results included Fourplay, Marcus Miller, Sade, Lee Ritenour and D'Sound. Since Acoustic Alchemy was guitar instrumentals, Satch came out probably tagged with 'rock' and 'guitar' which caused Blink 182 and Finch to show up too. Predixis said it will work anywhere in the world in whatever language. I tried Razorback's 'Payaso'. Yngwie J. Malmsteen, Satch, Dream Theater, Guns N' Roses and Jimi Hendrix showed up. Very accurate, with all rock titles and considering that this is an OPM title except that no other OPM titles showed up. Here's the surprise, they support open source. Where they stand with the linked page's title, "Open Source, Open Data". Yes, the data used for analysis is submitted by the community that's why it'll work in any country. Amazing! Also, at the bottom of the generated playlist (back to Winamp), there are links to free tracks, most of them by indie artists if I'm not mistaken. Sweet!

iRant has moved. Click here.