Comments on: Did Apple advance music discovery?
Apple's Genius, a recommendation engine that relies heavily on a user's listening habits to build playlists, could help iPod and iPhone owners squeeze more enjoyment out of their devices.
Apple's Genius, a recommendation engine that relies heavily on a user's listening habits to build playlists, could help iPod and iPhone owners squeeze more enjoyment out of their devices.
The world may have thrilled to the potential for a Google Phone, but what Google actually unveiled is its plan for a new smartphone world order.
Photos: Unboxing Nexus One
faq Worms, Trojans, and SMS attacks are risks for mobile phones, but the biggest practical threat to users is losing the device.
The Web is now the place to go for news and entertainment. Look here for the latest on blogs, music, video, virtual worlds, social networking and more.
Add this feed to your online news reader
After going through the process of collecting library data and sending it to Apple your able to select a track and the system generates a playlist. The playlist scales from 25 to 100 tracks at a rate which implies that the list is pre-generated when you upload your data to Apple. There is also a Refresh button which appears to just shuffle the tracks already in the playlist. Next to the Refresh button is a save playlist button. The saved playlist is a special class of Genius playlists- not modifiable.
The way the system appears to work seems to imply that what's generated has nothing to do with order of play- just some sort of matched qualities. Is it using a Pandora like strategy- just generating a list of similar tracks- or actually creating a mix which is the sum of its parts?
An annoying fact is that the usability of the Genius system is limited by which tracks iTunes is able to comprehend for use at all. The listing appears to be limited to tracks to iTunes. This leads to limited options for your use as well as limited components for the playlist as a whole.
This might be the cause of commonalities amongst various Genius playlists I've generated- or it might just be that I like very similar music.I've tested what I would call a wide range of tracks- and I've received mixed but redundant bags every time.
A large flaw might lie in the possible need to completely reset your Genius information when you add a new track to your library in order to use it for a Genius mix. The process took about 10 minutes with my 846 item music library- and would be quite the hassle if it had to occur regularly.
The way iTunes handles file data- both the location of the file and the file's other details is probably my greatest issue with the program- and it appears that Genius is another flaw- at least at this point.
Next to the new interface options (which have been available in Windows Media player for a long time)- I find no particular ups for this build of iTunes.
If I don't own it, either I don't care to, don't like it, or haven't gotten around to purchasing it yet. So you keep renting, and pray that Rhapsody doesn't go under or you don't actually get tired of their service because all that music you're so fond of is gone down the tubes.
And in reference to Genius, I don't have an opinion of it's effectiveness yet, though I do think the concept is a cool one.
i may be biased because i own one, but i can't help feeling that the new zune features beat the new iPod features, especially for people with a subscription, but even for people without a subscription who are just limited to three plays per song (at least that's how i read that it will work)
the genius playlists do seem pretty cool as well, but i think it would be much more useful if you could generate playlists on your ipod instead of having to do it through the software
What i really want though, is something that accurately sorts my songs into genres, as opposed to having tons of differently spelled genres (pop/rock, pop-rock, rock/pop) and having half my music classified very generally as "alternative"
Let's contrast this with a service that I've been using for awhile but only recently started to leverage; last.fm. I have a Last.fm plug-in for iTunes that tracks all of the music I listen to on iTunes. Because iTunes tracks everything I listen to on my iPhone Last.FM can also gather all of my "mobile" music listening as well. In addition, Last.FM has established "web services" relationships with other music sites like the Hype Machine such that they can also gather any music you listen to on those sites.
For awhile now, its been tracking *everything* that I listen to and I recently started to leverage that by going into Last.FM (they have desktop & mobile apps as well as a great website) and checking out their recommendations. Shockingly they are very very good + they have a ton of music you can listen to for free as well as videos, information, and its very easy to navigate to related artists once you find one you like. I couldn't believe how useful and interesting it was to spend literally hours going through their site learning about new music. Some based on recommendations and some based on the new music I was exploring in real-time. But again, the killer difference here is that I can listen to full tracks, and lots of them, for free. iTunes is 100% geared towards getting me to purchase tracks.
Again, the clear advantage you get with iTunes is the integration and ease of use. But value? At this point the recommendations I'm getting with iTunes are pretty bad. I expect that part of it will get better, but unless they have a change of mind around how they are monetizing their service, I suspect it will not be a very useful tool for me... and as long as I'm feeding the genius sidebar free info on my musical preferences, they are getting something from me without any compensation and I'm not sure how I feel about that
bsilverstein.info
But if the system "learns" over time by drawing inferences from millions of people's music libraries and purchasing behaviors, then it could get much better. I personally think last.fm is excellent with its recommendation engine and its social networking features, and even Amazon's recommendations (over time) have gotten pretty decent (I'm less impressed with Pandora, though I've been told that it can be "trained" to be good or at least acceptable).
Genius has a little ways to go before it can even approach what the others do in making recommendations, but I'd give it some time...the information they have access to is an impressive foundation.
Logically, this model can be extended to other media (TV, movies, audiobooks, applications), and over time, to potentially recommend non digital products, taking advantage of the fact that there are now 65M iTunes account holders with a credit card on file. Can you say e-wallet with product recommendations? Amazon, are you listening?
Btw, some further analysis on today's 'Let's Rock' event can be found here:
65 Million Reasons to be bullish on Apple
http://thenetworkgarden.com/weblog/2008/09/65-million-reas.html
Check out the full post if interested.
Mark
"Nine comments so far and not a single one positive about the feature"
Because this isn't the real world: This... is... CNET!!!
Click on the names. More than half joined today. Shills for the competition. Most of the rest are die-hard PC gamer / Microsofties.
- by A. Viner Seiler September 11, 2008 9:53 AM PDT
- EXCERPT: "The only thing they're missing is a way to slip a song they want you to buy into the playlist," McQuivey said. "If they could get the rights to do that, they would have an amazing music recommendation machine."
- Like this Reply to this comment
-
(17 Comments)RESPONSE: But that's exactly what Genius does, McQuivey. It tells you that you are "missing" some songs -- then it tries to get you to buy a few MORE songs by the same artist! It doesn't take a genius to name more songs by the same artist. And guess what? I already OWN the songs it recommended -- but Genius didn't notice. Some genius! Maybe I picked too obscure a song. Maybe it will get better when more people participate. Maybe I should remove it!