The device is built around ATTiny13A. It is powered by a single 3V coin battery and is quite trivial. It just might be the simplest, functional device we've ever built...
The Python script is pretty simple as well but it does need some other software like FFMPEG and SciPy. You can find the instructions on how to set up your system to work with our script on the project's GitHub page.
We are considering building a voice recorder with a tone generator, but this time the tone would be emitted via a speaker. If the tone will be loud enough, it should be picked up by microphones in all the cameras mounted in the cockpit and then we may have a chance to make a fully automatic solution. It would split the audio and video recordings and align them fully autonomously.
We'll see...
I know there are some smart tools in video editing software that can align audio with video, or align few video streams using their audio tracks but it seems that those tools fail when there's lot of noise in the audio tracks.
Nevertheless, our little gadget is not perfect. We should really build a custom voice recorder with a large button and some audio/visual feedback for marking important points in the recordings.