diff options
Diffstat (limited to 'converter/ppm/ppmtompeg/BUGS')
-rw-r--r-- | converter/ppm/ppmtompeg/BUGS | 43 |
1 files changed, 43 insertions, 0 deletions
diff --git a/converter/ppm/ppmtompeg/BUGS b/converter/ppm/ppmtompeg/BUGS new file mode 100644 index 00000000..64269dfb --- /dev/null +++ b/converter/ppm/ppmtompeg/BUGS @@ -0,0 +1,43 @@ +Known BUGS: + +1. <fixed> + +2. rate control and specifics files interact badly + +3. using REMOTE parallel execution: + `command` to find INPUT files does not work + SPECIFICS_FILE (CDL_FILE) + USER_DATA + dont work (should send the files over the socket, but do not + +4. using any sort of parallel code: + REFERENCE_FRAME DECODED + does not work + +5. Cannot use both STDIN and CDL_FILEs (ok, since it doesnt make sense...) + +6. <fixed> + +7. Patterns of BBI... seem to cause problems for parallel encoding. + I believe the bugs are fixed for sequential, but if you hit one, let us + know! + +8. Sometimes parallel encoding results in a zero size being written into + the stream. To fix, specify SIZE of the actual size. + +8.1 (related? minor, anyway) in parallel is also writes the Quant matrix + even when it is the default one + +9. It is unclear (to us anyway) if we should generate Y's from 0..255 or 16..224 + +10. <fixed> + +11. <fixed> + +12. reading JPEG from stdin, the encoder will sometimes stop before the end +of the stream (if the images are smaller than half the buffer size). + +13. Sometimes INPUT_CONVERT * is not good (requires "cat *"), in a parallel +case reported by a user it was needed. + +14. Parallel encoding often generates improper temporal_ref's... |