add 'audacity text' format ('aud-txt' ?)

[NickleDave]

as mentioned in #80

should add this first before worrying about getting LabelTrack annotations directly from the .aud xml

[NickleDave]

I think I might be able to use this dataset to get some example audacity.txt files:
https://figshare.com/articles/dataset/Wav_and_label_files_used_in_the_workshop/4714387

I can import the labels there, then export them again in the standard audacity .txt format

Looks like there might be some edge cases to be aware of, e.g. if someone tried to use an extended format with frequency ranges:
https://manual.audacityteam.org/man/importing_and_exporting_labels.html

Would like to handle that as well long term but not worry about it now

[NickleDave]

Looked all over for Audacity annotation datasets again. Didn't find any.

Then spent some time actually trying to annotate with Audacity.

It is not great for bioacoustics data, for a couple reasons:

• no segmentation algorithm
• not super-friendly to view spectrogram and audio simultaneously (just realized there is a "multi-view")
• both together -> you are finding segments "by eye"
• if you're in spectrogram mode you get frequency ranges added to labels
  • these get added to exported label tracks as separate rows with a backslash, meaning it now cannot be easily loaded with e.g. pandas, esp since they come after a line break, so you can't just, say, tell pandas they are "comments"

Given all this, I think it's best to close this and #80 and if someone really wants Audacity we can open them up again

[NickleDave]

Nevermind!

@nTrouvain and @neuronalX have released a new version of https://zenodo.org/record/6521932 that includes exported Audacity txt labels!

Clearly this shows someone can use Audacity to annotate! 🙂

• add 'aud-txt' format
• test it with this dataset
• include one file along with citation to dataset in crowsetta.data (branch in progress: https://github.com/vocalpy/crowsetta/tree/add-example-annot-files-fix-%2390)