Export detection as a video

[stevedee]

Some apps like Facebook make it difficult to attach audio files.
I normally use Shotcut to create an mp4 video from an .mp3 audio detection file + its associated .png spectrogram image.
I've no desire to 'bloat' BirdNet-Pi with unnecessary code, but wondered if it would be straightforward to add this kind of export feature to the Recordings screen, and whether it would be useful to others.

[blogdemoi]

Yes, I think it could be very useful, at least for me.

I am using ffmpeg on my laptop to create such video files from the .mp3 and an image, adding a spectrogram/waveform (I don't know the exact term), and increasing audio:

ffmpeg -i 1.mp3 -i 1.jpg -filter_complex "[0:a]showwaves=s=1920x1080:mode=cline:colors=white|white|white:rate=24:scale=lin,colorkey=0x000000:0.01:0.001,format=yuva420p[v];[1:v]scale=1920:1080[bg];[bg][v]overlay[outv]" -map "[outv]" -map 0:a output.mp4

ffmpeg -i 1.mp3 -filter:a "volume=10" output.mp3

And then I upload to twitter or other social media.
E.g: https://x.com/arnauld/status/1725078040638533977

Maybe it could be implemented directly in BirdnetPi, so one can download it and publish it on social media.

[stevedee]

Thanks blogdemoi, I've now taken it one step further by running ffmpeg via a button on my BirdNetPi.

Details here: http://captainbodgit.blogspot.com/2025/01/birdnet-pi-export-audio-recordings-as.html