Apple released version 8.0 of iTunes and included a nifty little function called Genius. It takes your song list and stores it in a database. Then by simply clicking on almost any song in your iTunes, it will create a new playlist for you. The songs in the playlist consist of similarities to the song you clicked.
For example, if you click on “Same Old Drag” by Apples in Stereo Genius builds a playlist based on that song from all the other songs you have in iTunes. It also opens up a side window which displays more songs from the iTunes store. These songs are recommended purchases to compliment the songs in the playlist. Unfortunately, if you selected a song for Genius playlist creation that is not available in the iTunes store it won’t build a playlist for you.
This type of database mining is fascinating to me and is indicative of the intelligent systems that are appearing practically daily on the Web. Let’s take a look at some of the playlist selections from my library. It has over 10,000 songs ranging from rock to blues to progressive to alternative to everything else. (I still can’t handle most opera and bluegrass.) Here are some of the songs in the new playlist using Same Old Drag/The Apples in Stereo as the seed:
- Another Sunny Day/Belle and Sebastian
- Stuck Between Stations/The Hold Steady
- Nature of the Experiment/Tokyo Police Club
These three examples, all pretty solid alternative songs, hold up well to the seed song. The playlist extended out to 50 songs and it’s great for getting through the workday.
Next, I used the venerable, fuzz-guitar classic, “Spirit in the Sky” by Norman Greenbaum. Genius went right to work selecting mostly rock classics from the 60’s and 70’s. Here are some of the songs it attracted to my playlist:
- Should I Stay or Should I Go?/The Clash
- Reelin’ In the Years/Steely Dan
- No Rain/Blind Melon (interesting).
The Blind Melon track was the only song not from the era of the rest of the songs. That shows Genius may be making some rudimentary connections to similar artists from different eras. I wonder if it looks at the artist’s root influences to make these connections?
One way to find out is to use a jazz or blues track. Jazz and blues artists nearly always do covers of music they enjoy or were influential to them. So the best place to start is a playlist built from a Howlin’ Wolf track called Rockin’ Daddy. From that mid-1950’s song Genius pulled three blues artists from later eras. They are:
- Shake ‘em On Down/R.L. Burnside
- Cut It Out/Gary Moore
- It Hurt Me Too/Keb Mo
Though there were plenty of Albert King, Freddie King and B.B. King tunes in the playlist, Genius was actually able to pull artists from different eras.
If we rank the order in which Genius selects songs for the playlist I’d have to say:
- Genre - blues, rock, alternative, etc.
- Era - 50’s, 60’s, etc.
- Name - more than one song by Howlin’ Wolf or Apples in Stereo, etc. in the playlist
My guess is that rarely does Genius make choices other than those three main attributes. It doesn’t make many, if any, choices based on region or similar sound or band member connections. Maybe that’s coming. Other reviewers have actually had somewhat opposite results. One of the flakier playlist results come from TUAW.com who says, “My Genius is a Moron!”
How satisfying is the Genius experience?
- It’s a quick way to make an interesting, though not perfect, playlist.
- I don’t care for the iTunes store recommended song sidebar.
- The playlists, for the most part, make sense and are enjoyable to listen to.
- They put too many songs by the same artist in the playlist.
- You can’t create guidelines or preferences to the actions Genius takes such as limiting the amount of songs by the same artist or using only 4-5 star rated artists, etc.
It’s an interesting start. I’d liken it to the early days of Google when they didn’t always give you the great, practicially dead-on search hits they do today. I hope to see future improvements in Genius, otherwise it’s back to making my own playlists.
Would you like to learn more? You may:
- Fill out our information request form.
- Review information found on our Web site.
- Or contact me directly at 216-674-1600 x 118
Tags: database mining • genius • itunes • music • playlist
Copyright © 2008 by IdeaStar. All rights reserved.












































Add a Comment