Just curious how many people use MoodLogic? I am not allowed to talk about whether or not I might be beta testing a third party development kit that might be a future connection into jEmplode. Just curious how many people would be interesting in this connectivity before I may or may not get too far into playing around with something like that...

ms

Interesting. Of course SONICblue have a strategic relationship with MoodLogic (and we own some stock I believe).

Rob

Hehe, that's a way of saying it.

Well, I tried Moodlogic once, not to imrpessive though, mainly because of their triallicensing.

TommyE

I used it for a bit and found it to be okay. The time it took to analyse a track and create the fingerprint took a little too long for me considering the number of CDs I've got though.

- Trevor

Is the SonicBlue relationship more of a "bundle MoodLogic onto the CD" or do the products actually integrated with MoodLogic? If the latter, do you know in what way they're integrated? Just trying to collect some ideas ...

Man, I could totally see the MoodLogic code, combined with the technology of MP3 players (tracking your listening behavior), to be the ultimate system for matching users' musical tastes to new artists.

MoodLogic have our Rio SDK and already support current portables natively. I can't comment on possible future directions or possible collaborations, but I can see the MoodLogic technology becoming more and more compelling as the database grows.

Rob

or as I like to call it -- "Project Serendipity"

I love MoodLogic! I use it all the time. It really helps simplify tagging and music organization. The only thing I would want them to fix is to allow other tags to be filled from their database, like the year. Anyway, I think MoodLogic would be a great addition to jEmplode!

I don't wanna naysay anybody, but I never found music suggestion algorithms very good. Maybe it's just me, but I love Liz Phair, but I hate Alanis Morissette and Sheryl Crow, for example. So, to me it's useless. My brain recognizes quality to be a much greater criterion than similarity (and even the similarity on those is stretching it, but I get those suggestions all the time).

But if other people like it, go for it. I have tried to find new stuff using those suggestions, and sometimes it pans out.

My brain recognizes quality to be a much greater criterion than similarity (and even the similarity on those is stretching it, but I get those suggestions all the time).

For me the dream isn't to have it exactly match my tastes. It's to have it suggest stuff that I might possibly like.

Okay, let's say for instance, I know I like Sheryl Crow (we'll just agree to disagree on that one ), I know that I only can barely tolerate Alanis Morisette, but I've never heard anything by Liz Phair, and perhaps I've never even heard the name (I actually have heard the name but let's just say I hadn't). I would still want it to suggest Liz Phair to me and offer to send me a sample MP3 and a link to where I can buy the album if I liked what I heard.

Whether or not I ended up liking Liz Phair is less relevant than having it say "maybe you'll like Liz Phair, try her out". At least that's better than having the record industry trying to cram Britney Spears or N'sync down my throat.

Imagine having artists suggested to you based on stylistic similarity, as opposed to having your music choices decided by how much cocaine was distributed to the radio DJ's by that particular record label.

The end result is that even the most obscure artists will get that chance to connect with fans who like their style of music. Artists who wouldn't otherwise have been able to make money would suddenly have actual FANS who happened to share their tastes.

This is definitely what I think too ... The intent is to introduce some sort of variety into the mix. It's so easy to stagnate with your personal collection because you don't even know a particular artist exists.

You're absolutely right, and I understand all of those points, and, in fact, I use that methodology on occasion, but it seems that every time that anyone suggests this, it's as if it's the answer to all of my problems. The problem as I see it is now you've transferred the ``power'' from the cocaine-swilling Los Angeles DJs to the money-hungry music retailers, and there's as much corruption in both places.

Oh, and as to Sheryl Crow and Liz Phair, if you like Sheryl Crow, I imagine that you'll like Liz Phair as well, since except for her (I assume) first album (the one with Santa Monica Boulevard or whatever that song was entitled), Sheryl pretty much ripped off Liz's sound and polished it up a little. Take a listen.

In reply to:

---

or as I like to call it -- "Project Serendipity"

---

So that's what those classes are for... I noticed them while doing some jEmplode work this weekend (the results of which should be in the next version, whenever that happens, and will likely make lots of people happy), but didn't examine them too closely...

The problem as I see it is now you've transferred the ``power'' from the cocaine-swilling Los Angeles DJs to the money-hungry music retailers, and there's as much corruption in both places.

My idea is that the database genuinely suggest artists based on stylistic similarity, without any weighting imposed by retailers, record labels, DJ's, anything else like that. It just does an impartial search on an entire database, and returns the syle-matching results no matter how obscure.

Of course, the database itself would have to be impartial in order for that to work. You'd have to start with something like the allmusic.com database and refine from there.

My initial work on this has been to rip data off of the AllMusic website pages... I emailed them about providing an API to their database, but given that all they have is a database, they were less than enthusiastic about that -- Which to me seems pretty silly since they're putting the data on public webpages. It strikes me that if you are counting on the fact that people won't parse HTML to protect your data, you have a pretty weak business model.