OLD | NEW |
1 1.0.2 | 1 1.1.90 (1.2 beta1) |
| 2 ================== |
| 3 |
| 4 [1] Added a JNI wrapper for TurboJPEG/OSS. See java/README for more details. |
| 5 |
| 6 [2] TurboJPEG/OSS can now scale down images during decompression. |
| 7 |
| 8 [3] Added SIMD routines for RGB-to-grayscale color conversion, which |
| 9 significantly improves the performance of grayscale JPEG compression from an |
| 10 RGB source image. |
| 11 |
| 12 [4] Improved performance for non-x86 machines. |
| 13 |
| 14 [5] Added a function to the TurboJPEG API which performs lossless transforms. |
| 15 This function uses the same back end as jpegtran, but it performs transcoding |
| 16 entirely in memory and allows multiple transforms and/or crop operations to be |
| 17 batched together, so the source coefficients only need to be read once. This |
| 18 is useful when generating image tiles from a single source JPEG. |
| 19 |
| 20 [6] Modified jpgtest to benchmark the new scaled decompression and lossless |
| 21 transform features in TurboJPEG/OSS. |
| 22 |
| 23 [7] Added support for 4:4:0 (transposed 4:2:2) subsampling in TurboJPEG, which |
| 24 was necessary in order for it to read 4:2:2 JPEG files that had been losslessly |
| 25 transposed or rotated 90 degrees. |
| 26 |
| 27 [8] All legacy VirtualGL code has been re-factored, and this has allowed |
| 28 libjpeg-turbo, in its entirety, to be re-licensed under a BSD-style license. |
| 29 |
| 30 [9] libjpeg-turbo can now be built with YASM. |
| 31 |
| 32 [10] Added SIMD-accelerated fast integer inverse DCT and YCbCr-to-RGB color |
| 33 conversion routines to accelerate JPEG decoding on ARM Linux and iOS platforms |
| 34 that have NEON instructions. |
| 35 |
| 36 [11] Refactored the TurboJPEG C API so that it uses pixel formats to define the |
| 37 size and component order of the uncompressed source/destination images as well |
| 38 as uses the libjpeg memory source and destination managers. The latter allows |
| 39 the TurboJPEG compressor to grow the JPEG buffer as necessary. |
| 40 |
| 41 [12] Eliminated errors in the output of jpegtran on Windows that occurred when |
| 42 the application was invoked using I/O redirection |
| 43 (jpegtran <input.jpg >output.jpg). |
| 44 |
| 45 [13] The inclusion of libjpeg v7 and v8 emulation as well as arithmetic coding |
| 46 support in libjpeg-turbo v1.1.0 introduced several new error constants in |
| 47 jerror.h, and these were mistakenly enabled for all emulation modes, causing |
| 48 the error enum in libjpeg-turbo to sometimes have different values than the |
| 49 same enum in libjpeg. This represents an ABI incompatibility, and it caused |
| 50 problems with rare applications that took specific action based on a particular |
| 51 error value. The fix was to include the new error constants conditionally |
| 52 based on whether libjpeg v7 or v8 emulation was enabled. |
| 53 |
| 54 [14] Fixed an issue whereby Windows applications that used libjpeg-turbo would |
| 55 fail to compile if the Windows system headers were included before jpeglib.h. |
| 56 This issue was caused by a conflict in the definition of the INT32 type. |
| 57 |
| 58 [15] Implemented a more efficient version of TJBUFSIZE() which computes a |
| 59 worst-case JPEG size based on the level of chrominance subsampling. |
| 60 |
| 61 [16] Fixed 32-bit supplementary package for amd64 Debian systems which was |
| 62 broken by enhancements to the packaging system in 1.1. |
| 63 |
| 64 |
| 65 1.1.1 |
2 ===== | 66 ===== |
3 | 67 |
4 [1] Added further protections against invalid Huffman codes. | 68 [1] Fixed a 1-pixel error in row 0, column 21 of the luminance plane generated |
| 69 by tjEncodeYUV(). |
| 70 |
| 71 [2] libjpeg-turbo's accelerated Huffman decoder previously ignored unexpected |
| 72 markers found in the middle of the JPEG data stream during decompression. It |
| 73 will now hand off decoding of a particular block to the unaccelerated Huffman |
| 74 decoder if an unexpected marker is found, so that the unaccelerated Huffman |
| 75 decoder can generate an appropriate warning. |
| 76 |
| 77 [3] Older versions of MinGW64 prefixed symbol names with underscores by |
| 78 default, which differed from the behavior of 64-bit Visual C++. MinGW64 1.0 |
| 79 has adopted the behavior of 64-bit Visual C++ as the default, so to accommodate |
| 80 this, the libjpeg-turbo SIMD function names are no longer prefixed with an |
| 81 underscore when building with MinGW64. This means that, when building |
| 82 libjpeg-turbo with older versions of MinGW64, you will now have to add |
| 83 -fno-leading-underscore to the CFLAGS. |
| 84 |
| 85 [4] Fixed a regression bug in the NSIS script that caused the Windows installer |
| 86 build to fail when using the Visual Studio IDE. |
| 87 |
| 88 [5] Fixed a bug in jpeg_read_coefficients() whereby it would not initialize |
| 89 cinfo->image_width and cinfo->image_height if libjpeg v7 or v8 emulation was |
| 90 enabled. This specifically caused the jpegoptim program to fail if it was |
| 91 linked against a version of libjpeg-turbo that was built with libjpeg v7 or v8 |
| 92 emulation. |
| 93 |
| 94 [6] Eliminated excessive I/O overhead that occurred when reading BMP files in |
| 95 cjpeg. |
| 96 |
| 97 [7] Eliminated errors in the output of cjpeg on Windows that occurred when the |
| 98 application was invoked using I/O redirection (cjpeg <inputfile >output.jpg). |
| 99 |
| 100 |
| 101 1.1.0 |
| 102 ===== |
| 103 |
| 104 [1] The algorithm used by the SIMD quantization function cannot produce correct |
| 105 results when the JPEG quality is >= 98 and the fast integer forward DCT is |
| 106 used. Thus, the non-SIMD quantization function is now used for those cases, |
| 107 and libjpeg-turbo should now produce identical output to libjpeg v6b in all |
| 108 cases. |
| 109 |
| 110 [2] Despite the above, the fast integer forward DCT still degrades somewhat for |
| 111 JPEG qualities greater than 95, so TurboJPEG/OSS will now automatically use the |
| 112 slow integer forward DCT when generating JPEG images of quality 96 or greater. |
| 113 This reduces compression performance by as much as 15% for these high-quality |
| 114 images but is necessary to ensure that the images are perceptually lossless. |
| 115 It also ensures that the library can avoid the performance pitfall created by |
| 116 [1]. |
| 117 |
| 118 [3] Ported jpgtest.cxx to pure C to avoid the need for a C++ compiler. |
| 119 |
| 120 [4] Fixed visual artifacts in grayscale JPEG compression caused by a typo in |
| 121 the RGB-to-luminance lookup tables. |
| 122 |
| 123 [5] The Windows distribution packages now include the libjpeg run-time programs |
| 124 (cjpeg, etc.) |
| 125 |
| 126 [6] All packages now include jpgtest. |
| 127 |
| 128 [7] The TurboJPEG dynamic library now uses versioned symbols. |
| 129 |
| 130 [8] Added two new TurboJPEG API functions, tjEncodeYUV() and |
| 131 tjDecompressToYUV(), to replace the somewhat hackish TJ_YUV flag. |
| 132 |
| 133 |
| 134 1.0.90 (1.1 beta1) |
| 135 ================== |
| 136 |
| 137 [1] Added emulation of the libjpeg v7 and v8 APIs and ABIs. See |
| 138 README-turbo.txt for more details. This feature was sponsored by CamTrace SAS. |
| 139 |
| 140 [2] Created a new CMake-based build system for the Visual C++ and MinGW builds. |
| 141 |
| 142 [3] TurboJPEG/OSS can now compress from/decompress to grayscale bitmaps. |
| 143 |
| 144 [4] jpgtest can now be used to test decompression performance with existing |
| 145 JPEG images. |
| 146 |
| 147 [5] If the default install prefix (/opt/libjpeg-turbo) is used, then |
| 148 'make install' now creates /opt/libjpeg-turbo/lib32 and |
| 149 /opt/libjpeg-turbo/lib64 sym links to duplicate the behavior of the binary |
| 150 packages. |
| 151 |
| 152 [6] All symbols in the libjpeg-turbo dynamic library are now versioned, even |
| 153 when the library is built with libjpeg v6b emulation. |
| 154 |
| 155 [7] Added arithmetic encoding and decoding support (can be disabled with |
| 156 configure or CMake options) |
| 157 |
| 158 [8] Added a TJ_YUV flag to TurboJPEG/OSS which causes both the compressor and |
| 159 decompressor to output planar YUV images. |
| 160 |
| 161 [9] Added an extended version of tjDecompressHeader() to TurboJPEG/OSS which |
| 162 allows the caller to determine the type of subsampling used in a JPEG image. |
| 163 |
| 164 [10] Added further protections against invalid Huffman codes. |
5 | 165 |
6 | 166 |
7 1.0.1 | 167 1.0.1 |
8 ===== | 168 ===== |
9 | 169 |
10 [1] The Huffman decoder will now handle erroneous Huffman codes (for instance, | 170 [1] The Huffman decoder will now handle erroneous Huffman codes (for instance, |
11 from a corrupt JPEG image.) Previously, these would cause libjpeg-turbo to | 171 from a corrupt JPEG image.) Previously, these would cause libjpeg-turbo to |
12 crash under certain circumstances. | 172 crash under certain circumstances. |
13 | 173 |
14 [2] Fixed typo in SIMD dispatch routines which was causing 4:2:2 upsampling to | 174 [2] Fixed typo in SIMD dispatch routines which was causing 4:2:2 upsampling to |
(...skipping 42 matching lines...) Expand 10 before | Expand all | Expand 10 after Loading... |
57 [1] Added documentation to .deb packages | 217 [1] Added documentation to .deb packages |
58 | 218 |
59 [2] 2968313: Fixed data corruption issues when decompressing large JPEG images | 219 [2] 2968313: Fixed data corruption issues when decompressing large JPEG images |
60 and/or using buffered I/O with the libjpeg-turbo decompressor | 220 and/or using buffered I/O with the libjpeg-turbo decompressor |
61 | 221 |
62 | 222 |
63 0.0.90 | 223 0.0.90 |
64 ====== | 224 ====== |
65 | 225 |
66 Initial release | 226 Initial release |
OLD | NEW |