Allow importing wav/flac/m4a and initialize snapshot from it

[xelrix123]

I have recorded over the years some audio files of movement sounds before/after service. It would be very nice if I could load those recordings (assuming some reasonable minimum length) to tg and have it use the recorded audio as a source (instead of live capture from a mic) for a snapshot. Ideally so that in addition to having import in GUI, I could also do ./tg-timer --input-snapshot ./input.flac to load it directly to snapshot, with ability to give option multiple times so that 1..n snapshots could be initialized from audio files.

Attached a sample capture of recently serviced Valjoux 7734 movement (made using piezo/contact mic) for a reference. It's WAV (because github wouldn't let upload FLAC), but ffmpeg/sox etc can easily used to convert it to flac/raw etc. format if needed. Before importing to tg, or I guess tg could also try to use those internally.

PS. Thanks for your work on tg!!!

valjoux7734.wav

[xyzzy42]

It's already possible to do this by setting up a loopback audio device. Play the recorded audio file with anything, perhaps ffmpeg, and then tell tg to use the loopback monitor of the playback as its source.

On Linux with pulse audio, I do this:

pactl load-module module-null-sink sink_name="tg-loopback"
pactl set-default-source tg-loopback.monitor
ffmpeg -re -stream_loop -1 -i valjoux7734.wav -f pulse -device tg-loopback output

[xelrix123]

Thank You! Nice trick! 👍 Should I close the issue, or do you want to leave it open for now?

[xelrix123]

Couldn't resist creating a simple tUI to circle through audio files given as param 1..n to feed them into loopback audio sink for tg-timer to display. See below...

$ python3 ticktockplayer.py attic/*.flac *.wav

Python3:

ticktockplayer.py

--

notes: claude code used so treat accordingly, Linux only (aka needs pulse audio/pipewire), for experimentation / demoing only