Apparently, PulseAudio has the option to redirect an audio stream to a device, though it is not installed by default by my distro (Linux Mint). I'm now trying to setup JackAudio which seems to have the option I need:
https://mintguide.org/musician/552-jack-aud...-for-linux.htmlUPDATE:
I have now finally managed to redirect the audio of the podcast to my laptop's microphone. I have given up with JackAudio, as it's very tricky to setup, and I went back to PulseAudio, which is actually already installed on Linux Mint. Only the PulseAudio's pavucontrol applet needs to be installed, as described here:
https://www.kirsle.net/blog/entry/redirect-...to-mic-in-linuxI have downloaded the podcast as an mp3 file, replayed it with VLC and redirected the audio output to Google's speech-to-text demo. You can see a screenshot of the process here:
Unfortunately, the demo is limited to only 5 minutes of speech, after that you need to refresh the page if you want to transcript another part of the audio. Also, while the transcripted text gives a general idea of what's being said, it has to be manually reviewed by listening to the whole podcast, so that all the parts make sense.
This has been anyway an interesting experiment, and thanks speculatrix for your offer to give the Amazon's transcription service a go. It might be interesting to compare the two, to see which one does the job better, I have attached here a text file with the transcription of the first 5 minutes, in case anyone is interested.
Varti
