Hi Gilles,

On Wed, Feb 16, 2011 at 2:25 PM, Gilles Maire <[email protected]> wrote:
> Hello
>
> Do you know how to merge two Yuv Images from their AVPicture Pointer ?
>
>
> For example :
>
> Image 1
>
>    __ __
> / \~~~/ \
> ,----( .. )
>
>    /      \__     __/
>   /|         (\  |(
>  ^ \   /___\  /\ |
>     |__|   |__|-"
>
>
>  Image 2
>
>
>
>    ~~~~~~~
>
>  ( °   ° )
>
>  \  L  /
>
>     =
>
> result
>
>
>     __ __ ~~~~~~~
> / \~~~/ \
> ,----( .. ) ( ° ° )
>
>    /      \__     __/    \   L  /
>   /|         (\  |(
>  ^ \   /___\  /\ |           =
>     |__|   |__|-"
>
>
>
>
>  Thanks :)

Might seem incredibly simple, but it's actually not, because how do
you decide which pixels to keep and which to add? Simple filters do
overlay-add, which means black pixels are considered "zero" and white
"one", and then you "add" the signal of each color component in each
pixel up. Obviously this only works in lab situations (e.g. think of
microscope images in green/red/blue channel, adding them up gives a
color image, as in Nature papers etc.), it would double the brightness
of the background.

More subtle filters set a background detection in one image and make
that transparent and then add only the opaque pixels up into the other
image. That's very complex however. This kind of stuff is what
Photoshop (and the GIMP) do for you, I don't think FFmpeg has such
filters yet.

Ronald
_______________________________________________
libav-user mailing list
[email protected]
https://lists.mplayerhq.hu/mailman/listinfo/libav-user

Reply via email to