openjpeg/libopenjpeg/t1.c, branch v2 Fork of openjpeg library for JPEG2000 encoding and decoding https://git.carlh.net/cgit/openjpeg/atom?h=v2 2010-10-07T17:45:04Z The original v1.3 branch had: 2010-10-07T17:45:04Z Mathieu Malaterre mathieu.malaterre@gmail.com 2010-10-07T17:45:04Z urn:sha1:cc01949777e56db1cb1bb2fd3809e2963da6efa3 ((int*)tiledp)[(j * tile_w) + i] = tmp / 2; while v2 had: ((OPJ_INT32*)tiledp)[(j * tile_w) + i] = tmp >> 1; Divide by two and a right shift operation are only equivalent when the data is unsigned. In this case the data is signed, so the right shift operation is incorrectly clearing the sign bit. Patch from: Sheet Spotter Deleting obsolete files and directories, adding v2-specific files and directories, updating existing files to v2. See README.v2 for more info 2008-07-31T18:47:41Z Antonin Descampe antonin@gmail.com 2008-07-31T18:47:41Z urn:sha1:7caaea18b728ead4e45226aafc09dba01e514a2d Patch by Callum Lerwick. The opj_tcd_cblk array is one of the largest allocations, because it contains a bunch of static buffers. This also makes it a major source of cache thrashing. This patch allocates the buffers from the heap, and dynamically sizes them in the decoder. I have not yet managed to dynamically size them in the encoder, getting the decoder to do it was tricky enough... I also split opj_tcd_cblk_t into separate encode and decode versions. A lot of fields were not used by both, so this cuts its size even further. 2007-11-14T10:52:02Z Francois-Olivier Devaux fodevaux@users.noreply.github.com 2007-11-14T10:52:02Z urn:sha1:010ae27471655e3fe05912bd9965245a4f2cd7f3 Patch by Callum Lerwick. This patch rearranges the largest memory allocations so they're allocated as late as possible, and freed as soon as possible. This cuts memory usage by about half on two large test images. 2007-11-14T08:45:00Z Francois-Olivier Devaux fodevaux@users.noreply.github.com 2007-11-14T08:45:00Z urn:sha1:24e189e4d8a231e7273d788b2bd77e2484cec761 Patch by Dzonatas and Callum Lerwick. Fp/vectorization patch which basically converts most of the irreversible decode codepath to floating point, eliminating a few rounds of int/fp conversion, resulting in a vast performance improvement, and an increase in accuracy. 2007-11-13T17:35:12Z Francois-Olivier Devaux fodevaux@users.noreply.github.com 2007-11-13T17:35:12Z urn:sha1:dbeebe72b9d35f6ff807c21c7f217b569fa894f6 In t1.c, small change to avoid calling twice t1_getwmsedec() 2007-11-08T14:26:03Z Francois-Olivier Devaux fodevaux@users.noreply.github.com 2007-11-08T14:26:03Z urn:sha1:c38de0728d481f1c821d9e86df960e1f1c36c04f Patch from Callum Lewick. Basic gcc optimization flags in cmake and makefile match. Patch from Callum Lewick. Memset patch. See ChangeLog for more details. Thanks Callum ! 2007-10-18T12:26:11Z Francois-Olivier Devaux fodevaux@users.noreply.github.com 2007-10-18T12:26:11Z urn:sha1:9e114bcaa47194a9f6af8eadf6d7cb9bb1ec277f Indexes can now be generated when decoding J2K codestreams. 2007-09-07T15:01:55Z Francois-Olivier Devaux fodevaux@users.noreply.github.com 2007-09-07T15:01:55Z urn:sha1:3816e0edf40b766ab78cdbf3ba5634b684630d0f Use <xmmintrin.h> for WIN32, else use <mm_malloc.h> 2007-08-21T12:57:39Z Francois-Olivier Devaux fodevaux@users.noreply.github.com 2007-08-21T12:57:39Z urn:sha1:1d02a8b59505e2faace66ad5cbb607e994cc46eb Moved the orient flipping into the ctxno_zc table. 2007-08-21T12:51:57Z Francois-Olivier Devaux fodevaux@users.noreply.github.com 2007-08-21T12:51:57Z urn:sha1:07be749bf2a8a1ae27f31b24b2eb132f5537b665 Thanks to Callum Lerwick who provided all these great patches !