projects / openjpeg.git / commit
? search:
summary | shortlog | log | commit | commitdiff | tree
(parent: 7f8f475) | patch
1) quantization stepsizes stored as float instead of shifted integers -> fixes a...
authorAntonin Descampe <antonin@gmail.com>
Thu, 27 Oct 2005 07:49:29 +0000 (07:49 +0000)
committerAntonin Descampe <antonin@gmail.com>
Thu, 27 Oct 2005 07:49:29 +0000 (07:49 +0000)
commit7ee36c3a4c37d9c8cee30ac19fd5532a3863285b
tree77a7f1ec3e1b356a3ac4f2cb666e21af37a70728tree | snapshot
parent7f8f47566fc1e98ce1aad3ff1a1621bf8b454527commit | diff
1) quantization stepsizes stored as float instead of shifted integers -> fixes a pb of precision when using very small stepsizes. 2)  bug fixed when decoding until bitplane 0 -> r-value (1/2) added to the coefficient.
codec/image_to_j2k.c diff | blob | history
libopenjpeg/t1.c diff | blob | history
libopenjpeg/tcd.c diff | blob | history
libopenjpeg/tcd.h diff | blob | history
Fork of openjpeg with some changes used by DCP-o-matic.