-
Notifications
You must be signed in to change notification settings - Fork 17
/
CHANGES
103 lines (89 loc) · 5.96 KB
/
CHANGES
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
Version 0.6.1
- Update window less often in Windows version of viewer program to improve compression speed.
- Support for 8-bit and 16-bit images and texture formats with one or two components.
- Support for RGTC1/2 (BC4/BC5) unsigned and signed texture compression formats with one or two components.
- Fix the block compare functions for 16-bit unsigned and signed components.
- Allow conversion in viewer program between one and two-component 8-bit and 16-bit signed and unsigned image
formats.
- Optimize the bptc_float decode function so that bptc_float compression is 10% faster.
- Optimize the bptc decode function so that bptc compression is 10% faster.
- Support using half-float source image for RGB(A)8 compression formats, resulting in higher quality compression
if the source image truly has more precision that a regular RGB(A)8 image (i.e. is not just a converted regular
image).
- Similarly, support using a 16-bit one or two component source image for 8-bit component compression formats such
as unsigned RGTC1/2.
- Add more conversion options to the viewer program between 8-bit and 16-bit formats and add an option to extend
half-float images from one or two components to three.
- Allow mipmaps to be generated even when there is already more than one mipmap present.
- Support generation of mipmaps for 8-bit and 16-bit signed and unsigned formats, and for half-float formats.
- Improve support for converting images to uncompressed textures.
Version 0.6
- Debug support for calibration of genetic algorithm parameters.
- Fix difference mode for half-float images in viewer program.
- Change --fast genetic algorithm parameters: population size 64 instead of 128, 200 generations instead of 100.
- Implement mode statistics in verbose mode for ETC2 and BPTC_FLOAT.
- Implement modal optimization (specific modes tied to islands) for BPTC_FLOAT compression, resulting in quality
improvement.
- Implement adaptive optimization, the GA is run twice as long when the RMSE of a block is above a certain threshold.
- Add --hdr option for gamma corrected HDR half-float texture compression. Optimize half-float comparison functions
using tables.
- Change compare_images to use the same block comparison functions used during compression.
- Fix bptc_signed_float (bc6h_sf16) decoding and compression.
- Added menu items to remove or add alpha component in viewer program.
- Fix R11_EAC and RG11_EAC compression and provide menu items to convert to r16 or rg16 format in viewer program.
- Add HDR display settings dialog to viewer program.
Version 0.5.1
- Correct typo in glInternalFormat id for DXT1.
- Fix major bug in writing DDS files (Caps field wasn't properly written).
- Properly support uncompressed DDS files with a DX10 header.
- Fix loading of uncompressed DDS textures.
- Finally fix BPTC (BC7) compression and decompression.
- Fix BPTC_FLOAT (BC6H) compression and decompression.
Version 0.5
- Make distinction between encoding and decoding in decoding functions. This can be relevant for behaviour being exactly
according to specifications.
- Fix bug in half-float conversion.
- Port to Windows (Visual C++ or MinGW).
- Fix byte order check bug in dxtc.c.
- Fix alpha handling in BPTC compression.
Version 0.4
- Support for the high quality BPTC (BC7) 128-bit format.
- Support some more PC formats in .ktx files.
- Added a table of texture format information and use it in various places. This should allow more formats to be
read and written.
- Make the --decompress command more versatile by allowing decompression to uncompressed texture files with optional
mipmaps instead of images only.
- Split block comparison function into own source file.
- Support half-float internal image format for compression and comparison (previously the only internal image format was
32-bit RGBA).
- Work-in-progress on BPTC_FLOAT (BC6H).
- Implement mipmap generation for power-of-two half-float textures.
- Fix bug in --ultra mode when the texture width was not a multiple of the block size.
- Add simple texture file viewer program "texview" using GTK+ 3.
Version 0.3
- Tweaked the GA parameters (mutation rate) for 128-bit block compression, improving quality.
- Added DXT3 and DXT5 compression.
- Added support for loading/saving multiple mipmap levels for DDS files.
- Added manual optimizing of alpha values for texture types with a limited alpha range (DXT3 and ETC2 punchthrough).
- During mipmap generation no longer blend alpha values in 1-bit (on/off) alpha images.
- Add --generations and --islands options for custom speed settings.
- When comparing pixels where both alpha components are zero, disregard the difference in RGB colors. This improves the
quality of compressed textures with sharp alpha transitions.
- Preliminary support for one channel (R11_EAC) and dual channel (RG11_EAC) texture formats and their signed variants.
- Added --flip-vertical and --quiet options.
- Split some source files.
- Added support for sRGB ETC2 textures. They are treated as regular RGB textures by the compressor, except for mipmap
generation.
- Use uint64_t in <stdint.h> for 64-bit unsigned integers.
- Generalized block size (although it is currently 4x4 for all supported formats in the GA).
- Reported RMSE and PSNR was slightly off due to calculation error (the number of pixels was taken as the extended
dimensions, not the real dimensions), now fixed.
- Support for decompression of half-float RGB and RGBA textures as well packed-pixel uncompressed textures from .ktx
files.
- Support for conversion (via "compression") to uncompressed packed-pixel RGB and RGBA and half-float RGB and RGBA and
storing them in .ktx files.
- Support for compressing a texture file that already contains mipmaps.
- Fix bugs in .pkm file loading.
- Preliminary support for ASTC file conversion via ARM's encoder via system shell commands to and from .ktx files and
to and from ARM's atsc format.
- Support for some uncompressed .dds formats.