On Sun, Jan 21, 2007 at 07:44:07PM +0100, Benjamin Larsson wrote: > > And new patch attached. .. some nitpicks .. > --- libavcodec/flashsvenc.c (revision 0) > +++ libavcodec/flashsvenc.c (revision 0) > @@ -0,0 +1,340 @@ > + > +/* Encoding development sponsored by http://fh-campuswien.ac.at > + */ You could save a line here by putting */ at the end of the line ;) > + * The picture is divided into blocks that are zlib compressed. zlib-compressed > + * have their compressed size represented with 16bits in the beginnig. begining > + * If the size = 0 then the block is unchanged from the previous frame. > + * All blocks are decompressed until the buffer is consumed. > + * > + * Encoding ideas, a basic encoder would just use a fixed block size. > + * Block sizes can be multipels of 16, from 16 to 256. The blocks don't > + * have to be quadratic. A brute force search with a set of diffrent different > + * block sizes should give a better result then to just use a fixed size. thAn > +/* TODO: > + * Don't reencode the frame in bruteforece mode if the frame is a dupe. Speed up. brute force > + av_log(avctx, AV_LOG_ERROR, "Input dimensions to large, input must be max 4096x4096 !\n"); too > + //copy the block to the temp buffer before compression (if it differs from the previous frames block) frame's > + av_log(s->avctx, AV_LOG_ERROR, "error while compressing of block %dx%d\n", i, j); compressing block > + //Search for the smallest framesize and encode the frame with those parameters frame size > + av_log(avctx, AV_LOG_ERROR, "[%d][%d]optimsize = %d, res = %d|\n", opt_w, opt_h, smallest_size, res); optimize? optim_size? > + //mark the frame so the muxer can set correct frametype frame type Diego
RetroSearch is an open source project built by @garambo | Open a GitHub Issue
Search and Browse the WWW like it's 1997 | Search results from DuckDuckGo
HTML:
3.2
| Encoding:
UTF-8
| Version:
0.7.4