14 DAY TRIAL // Music Beta by Google is still only available to a select few users and, while its main features fail to impress, there is one thing that has been received very well, Instant Mix, a feature which generates smart playlists on the fly based on several factors.
The feature was created by Google Research, the group working on algorithm heavy-projects, and it's fairly impressive in the range of cues it takes to create a string of tracks that make sense played together.
"Instant Mix uses machine hearing to extract attributes from audio which can be used to answer questions such as 'Is there a Hammond B-3 organ?' (instrumentation / timbre), 'Is it angry?' (mood), 'Can I jog to it?' (tempo / meter) and so on," Douglas Eck, Research Scientist at Google, wrote.
Machine hearing (PDF) is somewhat of an emerging field and Google has actually done quite a bit of research on the subject, which aims to enable computers to make more sense of audio sources.
But that's just part of the equation, Instant Mix also uses data that is has been able to gather from several sources on which bands and artist go together well.
"Machine learning algorithms relate these audio features to what we know about music on the web, such as the fact that Jimmy Smith is a jazz organist or that Arcade Fire and Wolf Parade are similar artists," Eck explained.
"From this we can predict similar tracks for a seed track and, with some additional sequencing logic, generate Instant Mix playlists from songs in a user’s locker," he added.
All of this means that the suggestions will be more suited not only to the style of music you're starting to listen to, based on the artists playing, but also to the mood of the individual tracks, ensuring that tracks added to the song are similar, even if they're from wildly different artists.