That pull request does that, but the subtitles are recorded live, it is not generated in post-processing. You could write a script to parse the .tlog or .bin file and generate it the same way.
It can be played live with some players such as VLC or burned into the video using ffmpeg or similar.
In fact there wasn’t any script involved, I did it all by hand, converting .tlog to CSV, then gathering the parameters I needed (date, time, depth and heading) and saving them to a subtitle file.
The final result was better than I expected, and the script is definitely in my “to do list”.
In this case, I had to process in two steps, one for the logo, and the other to hardcode the subtitle.
I did it using ffmpeg, following @williangalvani suggestion, the subtitles were edited manually with the data gathered from .tlog file and saved to a .SSA subtitle file.
As everything was post processed twice, it took a long time to reencode all the videos.
I managed to insert the logo with mmpeg and it’s fantastic. I used Mission Planner and converted .tlog to csv. But I have great difficulty in extracting the I needed parameters (date, time, depth and heading)
In this case, the first and second columns are date and time. For the parameters you need, just look for mavlink_vfr_hud_t line; the depth and heading will be in this row.