I promised to create some screen casts for the autocomplete service with GWT based on neo4j. After I have created all the screencasts I had to go through quite some hassle in order to do so.
So let me share my experience and toolchain in order to produce a somewhat accaptable screen cast. I am still not happy with the quality of the results and if you have any suggestions feel free to tell me.
First of all you download and install recordmydesktop. Once this is done you can just do a screencast by calling the following line.
recordmydesktop --quick-subsampling --full-shots --no-shared --v_bitrate 2000000 --on-the-fly-encoding
bear in mind that you should have at least a dual core processor if you really want to do video encoding on the fly.
After creating this screencast I realized that my audio and video tracks where not in sync. Even worse. I realized that the video track was always a few seconds shorter than the audio track. At the end of the video when the audio wasn’t finnished yet the last video frame would just be displayed. This might be alright on a 20 second screencast but some of my screencasts where up to 14 minutes and that really sucked! because speaking and video would just not be synchrone anymore.
In another screen cast that I did today I had the problems with the sound not being loud enough so Here is the stack of software that works fine with corrupt .ogv files produced by recordmydesktop.
After searching quite a while and trying several programms (pitivi, Avidemux,oggtools(!),…) that all crashed on the .ogv file that came out of recordmydesktop I reallized that with ffmpeg2theora I could actually extract the audio and video files seperately. to do so just enter:
ffmpeg2theora video.ogv --noaudio
ffmpeg2theora video.ogv --novideo

The problem was the Audiofile. I was able to play it with my Audioplayer but again not able to open it with any editing software but oggconvert!
With this little tool I was able to open the corrupt .oga audio file and save it to an .ogg file (with vorbis audio codec!).
Now after I had a non corrupted audiofile I was able to adjusting the speed of the audiofile adjusting it to the length of the recoreded video with audacity. Easy instructions how to do this are available at http://wiki.audacityteam.org/wiki/Change_Speed Audacity also provided the options for me to amplify the volume of the sound.
Finally I had two files of the same length just need to merge them again. The oggTools again failed to help me on this but there is pitivi which is able to do the job.
So you see if one keeps patient one will reach his goal. Some of the screen casts are already published and more will come. So far I hope you enjoyed reading my experiences and solutions
by the way on wikipedia there are quite some tutorials on how to edit ogg videos.
http://commons.wikimedia.org/wiki/Help:Converting_video