Visualizing Playlist Flow
Playlists are basically songs put into a specific order, sometimes for a particular occasion. So much is clear. But what makes a playlist good? A recent paper by Thor Kell and George Tzanetakis suggests, that good EDM playlists or DJ sets are usually made up of songs that match sonically: "Our transition analysis has shown timbre to be an important attribute used by DJs".
But while timbre (related to beaTunes color) apparently is a very important choice for EDM DJs, it may be less so for the madly in love teenager who works on his graduation mix, or the oldies show DJ, who is restricted to songs from the 50ies. But in any of these scenarios it can be helpful to get some visual feedback for how the list, not the song, sounds and what the transitions are like. beaTunes 4 addresses this by displaying a playlist image above the textual list and also showing a transition column in the main table.
The playlist image represents each song in three dimensions: height, width, and color. Each of these dimensions can be assigned to a song property, e.g. duration, BPM, or key. To change the current setting, simply right-click on the image and re-define what is being displayed. To listen to a given song, double-click on it to start playback.
While the playlist image is a positive, descriptive approach, i.e. it tells what you have, the transition column is a somewhat negative approach. Like a little brother it nags, it tells you with little icons what may be wrong with your playlist. Among the things it indicates are tempo, key, and language changes (all of these require prior analysis). If you are keen on avoiding certain transitions, like a harsh tempo change, the corresponding icon will make you aware of it (a tooltip shows more info), and might help you to avoid it.
Without a doubt, automatic playlist generation has come a long way and certainly has its (very convenient) place. But it does not replace manual set-building, as it cannot know your goals and interests or the given occasion. Automatic playlist algorithms simply cannot detect intent. What's left is trying to support playlist creators. That's where beaTunes shines. It's meant to support you in building excellent playlists—giving you a little more feedback than just the song names.