[Theora] ffmpeg/ffmpeg2theora &vhooks

[Aaron Colwell]

Oops. It would help if I read the whole email before responding.

I don't think modifying ffmpeg2theora should be overly complicated. You just
need to find where the raw frames come out of the FFMPEG code  and modify
them before they get encoded by the theora encoder code. This appears to be
somewhere around line 615 in ffmpeg2theora.c. You should be able to modify
the buffers in yuv before oggmux_add_video() is called.


Aaron

On Tue, Feb 21, 2006 at 05:27:14PM -0800, Aphid wrote:
> Hi theora-list,
> 
> I'm involved in in a project to redistribute C-SPAN's floor footage  
> from the US house and senate that we're capturing and converting*.   
> This footage is in the public doman, however there's a C-SPAN overlay  
> which is a registered trademark - we're having to overlay a "public  
> domain" graphic on top of it.  We're currently doing this with  
> ffmpeg's vhook functionality.  We ultimately need this footage to be  
> in theora, because we're using annodex to index the footage. So - a  
> couple of questions:
> 
> I've noticed some activity on the ffmpeg changelog w/r/t decoding  
> theora content - is there any hope of ENcoding via ffmpeg?
> 	this would allow us to use our ffmpeg vhooks to get our theora 
> 	content.
> 
> Is it possible/feasible to roll vhook funcitonality into ffmpeg2theora?
> 
> Otherwise, it seems as though we're stuck rendering mp4 in ffmpeg  
> with our overlay, then ffmpeg2theoraing that mp4.  This is  
> problematic because we're also hoping to offer a live feed.
> 
> Any advice or suggestions would be very appreciated.
> 
> Peace
> A
> 
> *the footage is currently available digitally in DRM'd & proprietary  
> RealVideo; we think it's a better idea to have it available in an  
> open format.
> _______________________________________________
> Theora mailing list
> Theora at xiph.org
> http://lists.xiph.org/mailman/listinfo/theora
>