[Theora-dev] Re: PC-based video server

[Ralph Giles]

On Wed, Jan 26, 2005 at 01:20:26PM -0700, Andrey Filippov wrote:

> As I understand, in videosecurity in many cases most of the frame is
> background that is perfectly stable, that will give significant advantage.
> And this is why they usually use "motion detectors" - I don't think it is
> needed in the case of Theora - whith proper adjustments there should be
> very few data sent out ifnothing changes

Yes, this is true. I did some theora movies of a simulation 
visualization and the slow-moving parts compressed *very* well. You may 
however, need to do some kind of pre-processing if you want really 
low-bitrate to avoid encoding the camera noise every frame. q.v. vorbis 
and digital vs. analogue silence.

> >  It's great that the network cameras can output video in theora format.
> When I decided I'm ready to implement videocompression I was planning to
> use MPEG-2. What stopped me from that - I could not stand that Elphel name
> will appear in the list http://www.mpegla.com/m2/m2-licensees.cfm :-)

Amen :-)

> Initially that server should do 2 (or 3) parallel jobs:
> 1 - uninterruptedly store all the high-res video on disks (DVR);
> and
> 2 - serve lower resolution video from live data in parallel with recording
> with minimal delay.

So you want to re-encode smaller versions of the camera streams?

 -r