[x264-devel] Boundary constraints

[Loren Merritt]

On Mon, 19 Nov 2007, Etienne Bömcke wrote:

> I've got another question for you concerning the definition of the
> boundary constraints within the encoder. Here's the deal : I've got
> two video sequences, encoded separately. The first one is of classical
> dimension, 720x576 for example, and the second one is a banner, let's
> say 200x576. My goal is to incorporate the second sequence in the firt
> one, replacing the bottom 200 lines with the banner. In order to do
> that I need two things :
> - I have to manipulate the bitstream to replace the sequence bottom's
> macroblocks with the banner's macroblocks.
> - The sequence has to be encoded without any prediction from the
> bottom area, neither spatial nor temporal.
>
> As for now I'm focusing on the second part, ideally I'd like to be
> able to define a mask defining the sequence areas where prediction is
> allowed and feed it one way or another to the encoder. I suppose I'll
> have to modify the code where the boundary constraints are checked.
>
> I just wanted to have a quick evaluation from people who know this
> code better than I do. Do you think it could be done?

You have control over the encoding process, and want to encode two videos 
such that they can be either viewed separately or later losslessly merged?

Do your banners go only on the bottom of the frame, or would you ever want 
to replace e.g. the left side? Your example is inconsistent: 200x576 
would be a side. This is important bause you really don't want to 
implement it without slices, and slices normally only stack vertically.

--Loren Merritt