while reviewing the video for my YunoHost talk, I realized an important issue with the video recordings from the devroom They only contain the speaker view and no slides …
I then realized that the http://control.video.fosdem.org/ interface (for which i was responsible for, at least during the first third of the devroom) was not just meant to validate that the recording was going well, but also to select what kind of view (only slide, only speaker, slide + speaker in the corner, or vice-versa) was to be recorded … My assumptions was just that both the speaker stream and slide stream were both recorded, it was not so obvious to me that I had to choose this .
So, quite sorry about that, I feel pretty guilty.
To try to fix my mistake, I think I can manage to create an ffmpeg-based script to reencode videos with slides added (e.g. slides in fullscreen + speaker in one of the corner ?). For this, I think “just” having a PDF version of the slide + the list of time at which to switch to next slide should be enough ?
So e.g. :
00:00 [slide 1] 00:24 [slide 2] 00:32 [slide 3] 1:13 [slide 4] ...
I know that sucks a bit, but at least we should be able to have a version with slide which is a bit more meaningful than just seeing the speaker :s