MusicIP plug-in deejays your iTunes

Review of music recommendation engine MusicIP's beta iTunes plug-in.

Michelle Thatcher Former Senior Associate Editor, Laptops
Tech expert Michelle Thatcher grew up surrounded by gadgets and sustained by Tex-Mex cuisine. Life in two major cities--first Chicago, then San Francisco--broadened her culinary horizons beyond meat and cheese, and she's since enjoyed nearly a decade of wining, dining, and cooking up and down the California coast. Though her gadget lust remains, the practicalities of her small kitchen dictate that single-function geegaws never stay around for long.
Michelle Thatcher
3 min read
MusicIP MyDJ for iTunes
The MyDJ sidebar lets you autogenerate playlists without leaving iTunes.

Late Monday afternoon, music recommendation engine MusicIP released a beta of its MyDJ iTunes plug-in for Windows (available as a free download from the company site). I've spent the last few hours playing with the plug-in, and thus far I'm impressed. Like the company's standalone program, MusicIP Mixer 1.6, the MyDJ iTunes plug-in scans and analyzes your music library to ferret out similar songs and artists. But the standalone software requires you to organize music in MusicIP, then export the playlist to your media player software. The plug-in lets you automatically generate playlists without leaving iTunes.

Downloading and installing MyDJ was easy, and after the program conducted a basic analysis of my library I was ready to build my first playlist. (The "extended" analysis takes a bit longer; more on that below.) In iTunes I clicked on a track from my current musical obsession, The Helio Sequence, then moved to the MusicIP sidebar on the right-hand side of the screen. From there I could designate the length of my playlist--based on number of songs, length of time, or file size--and decide whether I wanted the list to favor the style of the song, the style of the artist, or a balance of the two. The plug-in also let me dictate the degree to which the mix matched my selected song, on a scale from zero ("tightly focused") to nine ("very diverse"). Clicking on "Make Playlist" almost instantly generated a playlist inside of iTunes. As an avid consumer of the free mp3s from CNET Download.com Music, I was thrilled to see that MyDJ had called up some really great tracks that were hidden deep in my iTunes library.

The sidebar also includes a window labeled "Similar to the selected track," which lists a handful of free tunes that match your song's profile. You can click on a song to give it a listen and, if you like it, click one button to download it and add it to your library. This feature has potential to be amazing, but at this point the catalog of discoverable music seems a bit thin. It offered up the same recommendations for several dissimilar songs, and sometimes none of the recommended tracks seemed like a decent match. The more distinctive the song, of course, the more relevant the recommendations: my southern-fried alt-country tunes achieved more appealing matches than generic-sounding indie rock.

Being a beta release, MyDJ for iTunes does have a few quirks. My biggest complaint lies with the user interface, which right now just sits on top of the right side of your iTunes window. There's currently no option to independently move or even resize the sidebar; instead you're forced to resize your iTunes window to leave room for MusicIP. I'd prefer to have a few display options, including placing the control panel at the bottom of the window, in lieu of the iTunes MiniStore. Also, Mac users are left out for now, though the company does plan to release an Mac OS X version "in the near future."

A final note of caution: the program's "extended analysis" of music files, which presumably would bring even more accurate recommendations, definitely eats up some CPU cycles. I was glad the program let me choose among Low, Normal, or High CPU usage in the Preferences, but even the Normal level completely monopolized the ancient 1.5GHz Pentium 4 machine that I use as my music server. Those with poky processors or large music collections will likely want to run the analysis overnight. I plan to do just that and will update this post in the morning with any new insights.