projects / libav.git / blame
| Commit | Line | Data |
|---|---|---|
| 9aeeeb63 FB |
1 | Technical notes: |
| 2 | --------------- | |
| 3 | ||
| 4 | - To increase speed, only motion vectors (0,0) are tested | |
| 5 | ||
| 6 | - The decision intra/predicted macroblock is the algorithm suggested | |
| 7 | by the mpeg 1 specification. | |
| 8 | ||
| 9 | - only Huffman based H263 is supported, mainly because of patent | |
| 10 | issues. | |
| 11 | ||
| 12 | - Many options can be modified only in the source code. | |
| 13 | ||
| 14 | - I rewrote the mpeg audio layer 2 compatible encoder from scratch. It | |
| 15 | is one of the simplest encoder you can imagine (800 lines of C code | |
| 16 | !). It is also one of the fastest because of its simplicity. They | |
| 17 | are still some problems of overflow. A minimal psycho acoustic model | |
| 18 | could be added. Only mono stream can be generated. I may add stereo | |
| 19 | if needed. | |
| 20 | ||
| 21 | - I rewrote the AC3 audio encoder from scratch. It is fairly naive, | |
| 22 | but the result are quiet interesting at 64 kbit/s. It includes | |
| 23 | extensions for low sampling rates used in some Internet | |
| 24 | formats. Differential and coupled stereo is not handled. Stereo | |
| 25 | channels are simply handled as two mono channels. |