Split on Labels

New to Audacity (3.7.8.0 on macOS 26.6); may I ask for help, please?

Having produced the correct number of Tracks from an mp3 file using Analyze > Label Sounds, I go next to that dialogue’s Edit Metadata… box, where I see I can work on these parameters:

But when I load the resulting file (no errors etc) into Yate, they don’t appear as tags there for ‘Artist Name’, ‘Track Title’ etc.

Are they not ID3 tags, please?

Tia…!

Audacity has never delt with metadata very well. Maybe it will in version 4. I have not tried that yet.

We users have come to the conclusion that the only thing you can do right now is save the file with whatever tags is has, then after that edit it with a metadata editor like MP3tag.

1 Like

The procedure is as follows:

  1. Insert a label containing the song title/name at the beginning of the first song
  2. Do this for all other songs as well
  3. Select the command “File → Export Audio”
  4. In that dialog, select “Edit metadata” (bottom left) and just fill in the artist name and the album title
  5. click on “Set default” and then “OK”.
  6. Select the option “Multiple files”
  7. Split files based on - select “labels”
  8. select preferred naming scheme (I usually use “Track label name”)
  9. don’t forget to select your preferred audio format
  10. Then “Export”

This gives you a number of files which you can then edit using mp3tag (or in Music.app, if you prefer).

The Mac-version of mp3tag can be found here or in the Mac AppStore.

1 Like

Thanks, Rex! I can surely make that work, too :slight_smile:

Thanks, @romontschun ; will try that - provided it produces ID3 tags; is that the case, please?

And provided Audacity can automate adding track titles: I have 34 MP3 files which I need to split. That will produce over 500 separate tracks, each with different titles!

Your help appreciated :slight_smile: .

It is not a good idea to load mp3-files into Audacity and then export them again as mp3. Audacity converts an imported file into something it can edit / understand. And since mp3 is a compressedformat, it becomes de-compressed and then - during export - is compressed again, differently. This way you aredecreasing the quality of your files.

Better record them first as .wav, .aiff or any other uncompressed format.

Yes it does. If you follow my instructions, the content of each label is used not only for naming but also for the song title, and the album and artist names are also contained in the ID3 tag. Do not use the “year” field - it will come out double, for example, “1978” produces an ID3 tag of “1978 1978”. You really should fine-tune the ID3 tags with mp3tag (or another similar program).

There is also a quite detailed instruction in Audacity’s manual on the topic of splitting and exporting files. Worth being read.

Thanks so much for this helpful information and guidance, @romontschun !

I see. Yes. Thanks. So in my case infinitely preferable to batch convert the (as I say 34 each mp3 file with over 20 component tracks: that’s why I’m using Audacity to split them) into another format in, say, dBpoweramp or Permute and only then work on splitting each of those 34 mp3 files into the 550 tracks which they contain (ready to tag with Yate and finally import into Roon) in Audacity. Yes?

Thanks. I can’t easily see a way to do your Step 1, though. My apologies: as I say, new to Audacity.

When I load each of the 34 mp3 files (which - after your caveat above - might well be a lossless format*), only one track shows before I try to split it into the individual 20 or so tracks.

I’m splitting (= trying/planning to split) on Labels… I don’t have Cues or any other way of detecting breaks between tracks. I’m using Audacity to detect (presumably) silence.

Are you suggesting in your step 1 (thanks again!) that I should label each MP3 with something like ‘Volume I’ etc?

After splitting I have - roughly - 20 tracks. I don’t want to have to label them all (again, your Step 1) manually. 500+ ! I plan to use MusicBrainz or Discogs in Yate for that.

I’ve worked my way through the relevant part of the manual; thanks again for re-drawing my attention to it.

After all, I may well be missing some shortcut or trick which can make this whole process a lot more easy.

*These are harpsichord sonatas from the 1980s. Which lossless Mac file format would you advise? AIFF?

Thanks again!

You have recorded something. Say, a cassette tape or a vinyl record. Then, each side of the cassette or even the whole cassette is one single recording.

Then you need to determine where a new piece / song / whatever starts, click there and insert a label which contains the name of the song. Since the first song (probably) is at the beginning of your recording, you have to label it as well (#1 in my instructions).

If you export them as instructed, you’ll have as many files as you had songs in your recording. Properly named and with ID3 tags for artist / record / song name. You may then fine tune by adding publishing year, album cover, genre, etc if you wish.

AIFF is a lossless formar widely used on macOS, WAV would also work - it is more common on Windows. And there are also others. What you use is up to you. If you record using Audacity, you can save as .aup3 (Audacity 3 Project File). But I recommend that you additionally export into a lossless format as a backup in case the project files become damaged.

No - sou should give real names - not placeholders.

I have no idea what MusicBrainz or Yate are. You should really first try out my suggestions with one recording and you may discover how it works and how easy it is.

Thanks again, @romontschun; your patience much appreciated!

May I help you to help me by clarifying what I’ trying to do, please?

Yes. In this case I have 34 mp3 files. Each of these 34 files contains about 20 individual tracks.

I cannot tell where each track ends and the next one begins - either from the recording or by just looking at the display in Audacity. There are 500+ such ends/beginnings so isolating tracks and naming them manually is impractical.

I have discovered, though that I can get Audacity to do that for me by following these instructions. Namely (my first post explaining this was hidden because I linked to the other software I’m using in the hope that it would help anyone unfamiliar with it):

  1. import the first (of 34) mp3 (or another lossless format) files into Audacity
  2. select all
  3. Analyze > Label Sounds… > Apply. Audacity creates as many Labels (presumably based on the gap of silence between tracks) as there are tracks to be exported to Yate (my tagging software).
  4. File > Export Audio > On your computer.

It’s at that point that I’d like to add metadata - ID3 tags for Artist, Genre, Album title etc. But they don’t show up in Yate, my ID3 tagging software.

MusicBrainz is an online music database. With the correct metadata it will match each of my ‘Track 1’, ‘Track 2’, ‘Track 3’… … … ‘Track 20’ to the harpsichord sonatas’ correct names etc.

I can see no way of doing this (in Audacity) manually for 550 sonatas. Until, that is, I have already successfully split them!

I guess I may have to do without accurate meta tagging as generated by Audacity.

Thanks again for your help and suggestions, @romontschun !

I’m especially grateful for your caution against using mp3 files. dBpoweramp converts them to AIFF nicely :slight_smile: .

But I remain open to anyone correcting me on the process which I have arrived at so far.

There’s a free and open source MP3 editor called kid3.

https://kid3.kde.org/#download

The tagging part you’ve figured out — that should work fine once you’ve got the labels in place. But I think your real problem is finding where those 500+ split points actually are if you can’t hear or see the boundaries. Do the files have any silent gaps between tracks, or is it genuinely all one continuous mix? That’s what’s going to make or break the whole job.

1 Like

Thanks, Stacey!

I regret that my original post either wasn’t clear or has been corrupted/removed - because I believe I’ve also actually got the splits figured out as well :slight_smile: . This is what I’m doing:

  1. File > Open each individual file
  2. Skip to Start (Home key on numeric keypad)
  3. Select All (⌘A)
  4. Effect > Volume and Compression > Normalize > Apply
  5. Analyze > Label Sounds > ‘Track ##1’ > Apply
  6. File > Export Audio > On your computer

Does that look right to you, please… it seems to work?