Comment 19 for bug 237995

On Thu, Aug 20, 2009 at 10:55 AM, Alan Pope <email address hidden> wrote:

> To be honest I don't use xvidcap anymore, it's just too unreliable. I
> tend to use recordmydesktop. The reason for doing the separate audio and
> video was (2 years ago) mostly as a result of a limitation in the
> software. In that recordmydesktop would (and still does sometimes) lose
> sync between audio and video during the recording.

I just bought a Snowflake mic, and I think sync may now be even more of an
issue for me since RMD will not use it properly and I will have to record on
a separate app at the same time as RMD. (Oh lord how long will it be until
the 2.6.31 kernel becomes available - apparently that fixes it ;) ). But
like you said, even up to now, we've had sync issues, but they are
relatively slight over the course of even a 30min screencast.

> The second point you make was pretty much the way I'd liked it to go.
> Personally I have a dual boot system with one environment pretty much
> pristine, just for screencasting.

Yes, running it in a VM session always seemed like an invitation to
performance issues for me before I got my laptop (my desktop is a 6 yr old
P4 and I'm not sure it handle the recording in a VM that well). Dual booting
I think would be the optimum way.

>
> Ideally I'd like to scrap the pages that we have about "how-to" and
> start a fresh. I'd value your input for the way you do your screencasts.

While that seems to be changing lately (with html5, ogg and blender all
coming more and more onto our radar) for us, up to very recently it's been
RecordMyDesktop and a series of mencoder scripts (and for me, a run through
Audacity too). For sure it could be simplified and standardized though. The
software has matured substantially since we started IMO. Right now we are
offering downloadable ogg, streaming ogg (if you're using FF3.5), and
streaming flv as a fallback for other browsers. We're really hoping for more
streaming ogg support since that takes away our need to transcode to flash
(for time reasons along with the freedom aspect).

And while I've still got bugs to work out, I'm hoping to move to Blender for
the post production of my screencasts (editing, transcoding etc.). Right now
I can import the ogv's that RMD generates, edit everything up, but I'm not
entirely happy with the output files.. it needs some tweaking ;)

RQ