xvidcap thread

[Horst JENS]

here is my wiki entry for making screencasts on linux (ubuntu) using
xvidcap. Any other xvidcap-users out there ?

http://wiki.showmedo.com/index.php/Video_editing_Ubuntu

[guy...@gmail.com]

I use xvidcap, but I don't really have what to screencast.
Do you need some kind of help?

> signature.asc
> 1KDownload

[dai]

In the files section I've put a quick guide to setting up
recordmydesktop if you want to give that a go as a comparisson to
xvidcap

[Ian Ozsvald]

Dai, thanks for that. The file is 'RMD.pdf' which stands for RecordMyDesktop. It is a Linux program which rivals xVidCap and the pdf documents the settings that Dai used.

Dai used it to record part 8 of his 7zip series:
http://showmedo.com/videos/video?name=1070070&fromSeriesID=107

The result was an 11mb ogg file for about 8 minutes of video - an excellent compression rate for a standard ogg file. Other Linux users might want to take a look, it seems to be a very sensible way of recording screencasts.

Ian.

--
http://ShowMeDo.com
http://ShowMeDo.com/about (our pictures)
http://blog.ShowMeDo.com
http://groups.google.com/group/showmedo
I...@ShowMeDo.com

[Xanalogica]

I use ffmpeg for all my screencasts. I originally tried xvidcap (which is
just a GUI wrapper over ffmpeg) but hit audio incompabilities. xvidcap/ffmpeg
wants audio over OSS, not ALSA, which makes it complicated when you have
sounds used for other things as well and need to share the audio hardware.
But the biggest problem is that OSS does not support 24-bit audio sources and
the hiqh-quality USB headset/mic I bought _only_ supports 24-bit audio output.
Seems that's what professional folk use, although I didn't know that.

So I tried running arecord in parallel with xvidcap, as some have suggested
but hit audio sync problems. Then I noticed that the GStreamer system has a
window-record feature and tried with a complex GStreamer setup, first in C++
and then later recoded into Python. It was unreliable in quality, with
GStreamer crashing sometimes for no reason. So I tried getting lower-level
and wrote my own ALSA application in C++, and then again with Python, running
in parallel with xvidcap/ffmpeg but again the audio and video were not synced.
In frustration I finally broke down and just hacked ALSA support directly
into ffmpeg. And all my audio sync problems went away. It timestamps the
audio samples against the video frames.

I also tried pyvnc2swf in the very beginning but found it too limiting, in
that some viewers complain about flash and want ogg theora and other formats.
ffmpeg let's me capture in .avi and convert to all the other formats.

-Jeff

[Horst JENS]

> In frustration I finally broke down and just hacked ALSA support
> directly
> into ffmpeg. And all my audio sync problems went away. It timestamps the
> audio samples against the video frames.

Hi Jeff,
that sounds very interesting,
do you have a package of your hack to download (.deb or .rpm) ?
-Horst

[Xanalogica]

Not a package, no, but a tarball is online. You should only need to do the
usual "./configure" and "make" to obtain the binary.

https://www.dfwpython.org/repo/Projects/ffmpeg-w-alsa.tgz

I hope to find time to submit it as a patch upstream but it needs work first.
Right now it is hard-coded for what I needed, 24-bit audio samples and an
amplification of 8x. It needs to be hooked into ffmpeg's various command-line
options for sample size.

To use it, specify an arg of "-f alsa_device" instead of "-f audio_device".

-Jeff