I have gone through and updated most of my library using Picard. Now I am wanting to clean up the formats a bit. Everything to MP3 192k or lower if it already exists.
Looks like my files are getting thrown into the Unclustered folder again.
Is there a way to prevent this?
Using ffmpeg with:
-map_metadata 0 -id3v2_version 3 -c:v copy -b:a 192k
If you have MBIDs in there this should not happen.
In Picard, on OPTIONS \ GENERAL make sure there is no tick next to âIgnore MBIDs when loading new filesâ.
It is these MBIDs that means Picard knows what your tracks are.
Thanks Ivan.
That checkbox isnât ticked (I havenât touched it). Iâll do some more testing to see if I can track anything down.
Appreciate the idea and welcome anymore.
Do you files have the MusicBrainz IDs, specifically the recording and release ID, stored in the tags (using the standard tag names for it)?
Yeah, that looks good - I think the only settings I changed was the naming convention / move files.
I can âmapâ this specific file again, save, reload it, and see if any of the data changes from what it is now.
EDIT: I should have looked at a good file first.
Those tags arenât right⌠(this screenshot is me having mapped it and it showing the Original vs New tags)
The values are correct but the tags themselves arenât in the âhuman readableâ format for lack of better words. There are a lot like this (ORIGINALDATE vs Original Release Date).
Wonder how they ended up like that
If you are format shifting then you need to remap some of those. FLAC has different tags to MP3, etc. There is a chart in the help files below. And the best expert is also in this thread.
What format were these, and what are they now?
https://picard-docs.musicbrainz.org/en/appendices/tag_mapping.html
I bet someone somewhere has a clever ffmpeg script that would do this for you.
Thanks again. Half the battle is knowing what to search for 
Now that I know that, finding a few hits on this, like: MusicBrainz tags: MUSICBRAINZ_ALBUMID or MusicBrainz Album Id | Roger's Blog
I agree - there is likely a solution out there I just need to find.
I see whatâs happening. Picard has different tag names for different formats. When this got defined many years ago, the idea was to follow naming schemes that existing tags for each. So you get something like MUSICBRAINZ_ALBUMID in Vorbis/FLAC, TXXX:MusicBrainz Album Id in ID3, ----:com.apple.iTunes:MusicBrainz Album Id in MP4 etc.
Now when converting formats if the format support copying metadata at all it will often convert known properly (e.g. it will now about converting Vorbis ARTIST to ID3 TPE1), but all unknown / custom tags it will write to the target format as a freeform tag with the original name. So you end up with Vorbis MUSICBRAINZ_ALBUMID being written to ID3 TXXX:MUSICBRAINZ_ALBUMID.
What you could do in Picard is using a small script like this:
$if($not(%musicbrainz_albumid%),$set(musicbrainz_albumid,%MUSICBRAINZ_ALBUMID%)$delete(MUSICBRAINZ_ALBUMID))
$if($not(%musicbrainz_recordingid%),$set(musicbrainz_recordingid,%MUSICBRAINZ_TRACKID%)$delete(MUSICBRAINZ_TRACKID))
$if($not(%musicbrainz_trackid%),$set(musicbrainz_trackid,%MUSICBRAINZ_RELEASETRACKID%)$delete(MUSICBRAINZ_RELEASETRACKID))
Add this as a new script in Options > Scripting. Load your files, they end up in the unmatched pane. Right click them and select âRun Scriptsâ > Your script name. This should update the tags to use the standard tags.
Then save the files (I would suggest to only enable âSave tagsâ, not move or rename), remove the files from Picard and load them again. If everything went fine they now should load into the proper releases and tracks.
The above script is for the relevant tags needed for automatically matching loaded files, but you can of course extend it to all other mapped MBID fields.
That is fantastic! I like handling it in Picard. Thanks to all of you
The power of the Magic Scripts from the Prince of PicardâŚ
Once converted you may also want to add a step to delete the ALL_CAPS versions to clean up.
Iâm currently hitting the same problem and I found another solution for ffmpeg. I use:
ffmpeg -i $1 -f ffmetadata - | convert_metadata | ffmpeg -i $1 -i - -map_metadata 1 $2
where $1 is the input file and $2 the output file and convert_metadata is a bash function that reads from stdin and writes to stdout the converted metadata. Maybe I will post the full script when I am done.
Seems unnecessarily complicated.
How about a option in Picard to use Vorbis tags in all formats?
Adding another option and having two separate naming schemes does not seem to be less complicated. Especially as a lot of software supports the existing names.
Picard (and other software) need format specific mappings anyway because of the specific predefined tags for different formats.
Iâm maintaining my library in flac with Vorbis metadata. I transcode to mp3s. Adding the need to script the changes is the extra complexityâŚnot the software design part. I understand that, but itâs a different issue.
I don;t know the details of " TXXX:MusicBrainz Album Id in ID3". Canât the format handle uppercase? Is there really that much software that handles the MusicBrainz IDs? Couldnât we have upper case as well?
The program you use for encoding just needs to support translating the tags. FFmpeg doesnât but I just found out that fre:ac (freaccmd) does this correctly.