More > JRiver Media Center 24 for Windows

[Feature Request] Auto Playlist based on music fingerprint

<< < (2/3) > >>

RoderickGI:
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.

JimH:
MC can do an audio fingerprint.  It's been used for track lookup for many versions.  It isn't otherwise exposed to users.

RoderickGI:
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.

JimH:

--- Quote from: RoderickGI on March 10, 2018, 05:03:21 pm ---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?

--- End quote ---
It could be used to select other tunes that matched it, but that's not very useful.

--- Quote ---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.

--- End quote ---
And seems pointless to me.

RoderickGI:
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.

Navigation

[0] Message Index

[#] Next page

[*] Previous page