INTERACT FORUM
More => Old Versions => JRiver Media Center 24 for Windows => Topic started by: gheinzel on March 10, 2018, 06:18:41 am
-
I have a huge music only collection.
I like to discover or rediscover music from my own collection based on music I like. Not just like Genre or artist kind of similar.
I mean acoustic fingerprinting similar.
Why is it, after 24 paid software iterations not possible to have this simple but very useful feature?
I see many feature requests that are very geeky and are most likely used by 0,01% of all users, but this feature would be used by 100% of all users all the time. I feel I ask for that feature now for the 10th time.
-
Is that not what Play Doctor does?
-
Is that not what Play Doctor does?
No, unfortunately, not at all. I talking, take a fingerprint of a song, listen to it, and play a similar sounding songs based on the first or previous fingerprint. Play Doctor is OK, but is feels as if it uses Genre and Artist, and not truly musical fingerprints, how a human perceives music.
-
Play Doctor uses Genre, Beats per Minute, Favorites, and others. It also learns from what you skip.
-
Play Doctor uses Genre, Beats per Minute, Favorites, and others. It also learns from what you skip.
Thanks, to state what it does. I just know, it does not what I expect from a mature music server. Do you think this feature can be added sometime soon?
-
Interesting.
MC doesn't calculate or store an Acoustic Fingerprint at the moment. I use MusicBrainz Picard to check the identity of tracks, but that is all I use them for at the moment. I also use Shazam when Picard can't make an identification, which works well on most popular or common music.
Can software select similar tracks from the Acoustic Fingerprint? What other software is doing that at the moment, and not just using metadata? I'd like to look into that.
-
MC can do an audio fingerprint. It's been used for track lookup for many versions. It isn't otherwise exposed to users.
-
So it is calculated but not stored in a tag, but it could be, as part of Audio Analysis?
The interesting bit for me is, could similar music be selected based on the Acoustic Fingerprint?
Wikipedia mentions it. I think I did read of someone claiming they could do that, but there were sceptics. After all, an Acoustic Fingerprint can easily be used to recognise a track by comparing it to known fingerprints, but pulling out features of a track and finding other tracks with similar features is a whole new challenge.
-
So it is calculated but not stored in a tag, but it could be, as part of Audio Analysis?
The interesting bit for me is, could similar music be selected based on the Acoustic Fingerprint?
It could be used to select other tunes that matched it, but that's not very useful.
Wikipedia mentions it. I think I did read of someone claiming they could do that, but there were sceptics. After all, an Acoustic Fingerprint can easily be used to recognise a track by comparing it to known fingerprints, but pulling out features of a track and finding other tracks with similar features is a whole new challenge.
And seems pointless to me.
-
The process of Discovery, with the enormous range of media available these days, is becoming a very big part of media consumption, and very big business because of it. The best or at least easiest way to discover new material that you like is to find something that is similar to media you already like.
After all, Google makes a fortune out of presenting stuff that we spent time on previously, the presumption being that if we spent time on it, we liked it. Amazon is doing the same thing.
I hardly think that is pointless.
I'd still like to hear from gheinzel about software that he thinks does it now.
-
I didn't say that discovery is pointless or that finding similar music is pointless. Play Doctor has been around a while.
I do think that trying to use audio fingerprints to do it is pointless.