[FFmpeg-soc] Optimal Huffman tables for (M)JPEG

[Deep Pal]

Hello,
       I have been working on the patch for the Huffman table.
It seems since I have not done much communication,
my name has been taken out from there. Please consider
me for that patch.

The call to ff_mjpeg_encode_picture_header takes place even before
the DCT encoding takes place. However, jpeg_table_header
includes the 4 Huffman tables in that function. So either

    a) I haveto defer the header call at a later stage
or b) insert the huffman tables and move all the subsequent
       header information as well for a later stage.

Any advise in this respect or any other aspects
pertaining to the huffman table generation is greatly
appreciated.

Also, I would like to pursue any one of the following
 a ) MVC decoder ( I would like to know the expectation of the mentor on this )
 b)  WMV encoder ( if a mentor comes forward, also some document material
      for writing the application will be helpful.)
 c)  Nellymoser Encoder
 d) The  Improve Ratecontrol project also seems interesting as there are bunch
     of algorithms , same with shot boundary detection. If the project is
     detailed a little bit more it would be extremely helpful.

Regards,
Deep