[x264-devel] Re: Rate control and scene change detection

[Måns Rullgård]

Felix Mellmann <felix.mellmann at web.de> writes:

> Måns Rullgård wrote:
>
>>I've been working on implementing proper rate control for x264.  The
>>rate control is working quite well, but gets confused by the oversized
>>P frames produced at scene changes, so a good scene change detection
>>could be useful.  Does anyone have one in the works, or know a good
>>algorithm?
>>
> I think you answered your question basically already - P frames on
> screen changes are obviously greater than the average, so wouldn't
> that be a sufficient criterion for a scene change. Let's say you've
> rendered a P-frame which is 20 percent greater than the average - is
> it then possible to render the same frame as an I-frame and compare
> both if the I-frame would be smaller to store?

That would be likely to produce good enough results, but it takes more
time than is necessary.  I was hoping to find some method that could
do a quick estimate before actually encoding the frame.  Maybe that's
not possible, though.

-- 
Måns Rullgård
mru at mru.ath.cx

-- 
This is the x264-devel mailing-list
If you are in trouble, please contact <postmaster at videolan.org>