Today I'm going to do something a bit different. For once, the tables are turned and I'm asking a question, rather than providing an answer.
You may or may not know that since our 2.0 relaunch back in October last year, we've supported video to accompany posts. You might have seen this one I did about Moodle, or Jacob's excellent look at PiTiVi.
One of the things I've learnt is that producing videos on a Linux system is, well, tough. The set-up I personally use at the moment isn't a very good one, and (bad Peter) it involves some non-free software as part of the process too.
Let me explain. First of all, I use Istanbul to record the screen and dump it to an Ogg Theora video file. After that, it is transcoded to MPEG through ffmpeg and moved over to my Mac, where I edit and record the audio separately in iMovie 06. Once the editing and audio is done, I export it out through iMovie to H.264, and then use ffmpeg to transcode it back to the final versions - Ogg Theora and Flash Video for the embedded player.
As you can see, that workflow is seriously flawed in a number of ways. It involves too much needless transcoding, which hurts video quality badly. On top of that, it's a bit ridiculous to have to cart the files around between two different machines and two different platforms to get the process done.
I'd like to replace this way of doing things with a new workflow, based on free and open source software if at all possible, and ideally cutting the Mac out of the picture entirely so everything can be dealt with swiftly and easily on one Linux machine.
So here I throw it out to you. Has anyone out there had experience with FOSS-based Linux screen capture and does anyone have any ideas on how I could rework the way I do things to make this better? Even if it involves radically changing the set of tools I am using, I am open to any ideas that could solve the problem.
For a quick reference, all we need is the ability to:
- Record 640x480 screen capture at a reasonable frame rate (15 fps or so).
- Edit the video after production to cut out unnecessary parts etc.
- Record audio separately and splice in at the post-production stage.
- Export out to 640x480 Ogg Theora with Vorbis audio and 320x240 FLV/64kbps MP3 audio.
Thanks in advance for any comments and suggestions!