Help creating screencasts on Ubuntu. keeping audio and video in sync

[tchomby tchomby]

I'm creating a series of screencasts to teach object-oriented programming in
java to complete beginners. This is for a university, the screencasts are to
replace 14 one-hour lectures, so they're pretty long screencasts and I have
a lot to make. When I'm done I want to make them all freely available.

I've been having a hellish time trying to figure out the process of creating
a screencast, and am at the point of desperation. I just really need to get
going on making the things, and stop battling with technical problems. If
anyone could help me it would be deeply appreciated.

So I thought I'd turn to this list for help, maybe some people could walk me
through the ways they do it, or explain to me why I'm having problems. I'm
looking for a simple approach, simple is the key, cause I just need to get
on with it.

What I tried to do was use gtk-recordmydesktop to record the full screen
(1280*1024) video and audio in one pass, then import the OGG file into Kino
(which uses ffmpeg to convert it to DV) and do some very simple editing,
editing out the bits where I cough or make mistakes, and adding some text,
then export it to MPEG.

The problem is that the audio and video will not stay in sync, no matter
what. I even installed a real-time kernel, and it didn't help. Either they
go out of sync in the original ogg file produced by recordmydeskop, or they
go out of sync after kino imports the file, or both. They get further and
further out of sync as time goes on, becoming completely seperated after a
few minutes, or sometimes they just go in and out of sync at different times
during the video.

What is causing this? Do I need to record a smaller area of the screen? Put
my screen in a smaller resolution? Install a virtual machine, run it in a
window smaller than the entire screen, and record that? Can this one-pass
approach with recordmydesktop and kino work at all?

Most screencasting on linux tutorials describe a two-pass approach, first
recording the video, then playing it back while recording the audio, then
combining the two. Does everyone do it this way? Isn't this awkward, having
to guess how fast to go when making the video? Are their techniques you can
use in a video editor, such as having the video freeze while the audio
continues on, to give the audio time to catch up? What editors do people use
and can they describe their techniques?

I know about this page:

https://wiki.ubuntu.com/ScreencastTeam/RecordingScreencasts

but it seems pretty old, xvidcap for one no longer works on ubuntu, and I
was hoping there was a simpler solution.

Maybe the best solution I have found is the one described here:

http://rhodesmill.org/brandon/2008/how-to-produce-a-linux-screencast/

basically, recording only a very small area of the screen using
recordmydesktop with on-the-fly-encoding, then using mencoder to transform
the ogg file into a dv file, then editing it in cinelerra, then exporting.
But recording only a small screen area is a limitation, I would need to
learn cinelerra which looks hellishly complicated (if you try to open the dv
file in kino it doesn't recognise it as a dv file and begins converting it
to dv), and from my tests I'm not confident that even this approach will
keep the audio and video in sync over many minutes.

I'm considering giving Wink a try.

Hell, I'm even considering going to Windows and buying some proprietary
solution.

Any help would be greatly appreciated.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: https://lists.ubuntu.com/archives/ubuntu-screencasts/attachments/20080915/3f422e68/attachment.htm