Feature request : Optimising (MPEG) video streams

[Psykauze]

Hello,

We all know jpegoptim by tjko for optimizing huffman table on jpeg files.
This is useful for pictures made by most of camera that make "non-optimized"
pictures.

Some of video files are made by those camera and make "non-optimized"
frames too.

I know than mjpeg, mpeg 1, 2 and probably 4 use jpeg algorythm
(DCT+Quant+Huffman) for I-frames and use Huffman table too for other frames.

My request is to make an option on ffmpeg for just optimizing Huffman table
in video stream.

Regards.