I have not used any features of AcousticID as of yet but am working on a complete re-encoding of my music after which I will run it through Picard to tag and do AcousticID submissions. I have gone through posts but still do not understand if it is Picard or the AcousticBrainz server that generates all the data. Is the new data (if it is new data) generated from existing data or does there need to be a new submission? I hope what I asked makes sense.
First, AcoustID is not the same as AcousticBrainz.
(There is no such thing as AcousticID.)
AcoustID.org is a 3rd party project providing complete audio identification service, working together with MusicBrainz.
AcousticBrainz is one of the many projects from MetaBrainz Foundation.
You can read about the process using Picard for fingerprinting here:
https://picard-docs.musicbrainz.org/en/tutorials/acoustid.html
It is fpcalc.exe (a commandline tool provided from Acoustid.org) which calculates a fingerprint from your song. This happens locally, without any external request. This fingerprint will be submitted by Picard to Acoustid.org.
Acoustid.org then compare this fingerprint to existing fingerprints. If found (exactly or with high similarity) it returns the so called AcoustID. With this ID Picard can find the additional metadata on MusicBrainz, if available.
AcousticBrainz is not involved in this fingerprinting process.
3 Likes
Thank you for the response and more importantly the link to the doc page. I am still a little confused about “acoustic similarity in AcousticBrainz”, is this information generated from existing fingerprints or does there need to be additional processing by fpcalc before submission?
fpcalc and fingerprints are used by AcoustID which is entirely separate project from AcousticBrainz and its Recording Similarity feature.
Recording similarity data is generated by AcousticBrainz server from the submissions provided by the users via AcousticBrainz Client software.
2 Likes
Thank you, I think I now understand, I will need to read more.
3 Likes