CloverBootloader/CloverApp/libimagequant-2.12.6/CHANGELOG
vectorsigma72 286f7c643f Optimize Theme ability for Clover.app
When this option is enabled all png images will be converted to PNG-8 (8 bit per pixel).
2020-03-07 19:30:59 +01:00

66 lines
2.1 KiB
Plaintext

version 2.12
------------
- new liq_histogram_add_fixed_color()
- faster for large/complex images
- workarounds for Microsoft's outdated C compiler
version 2.11
------------
- new liq_image_set_background() for high-quality remapping of GIF frames
- new liq_image_set_importance_map() for controlling which parts of the image get more palette colors
- improved OpenMP support
version 2.10
-----------
- supports building with Rust/Cargo
version 2.9
-----------
- new liq_histogram_add_colors()
version 2.8
-----------
- standalone version
- added Java interface (JNI)
- new API for remapping multiple images to a single shared palette
version 2.7
-----------
- improved dithering of saturated and semitransparent colors
- libimagequant reports detailed progress and supports aborting of operations via callbacks
- fixed order of log output when using openmp
version 2.5
-----------
- replaced color search algorithm with vantage point tree, which is much faster and more reliable
- deprecated IE6 workaround
- warn when compiled without color profile support
- improved support for predefined colors in libimagequant
version 2.4
-----------
- fixed remapping of bright colors when dithering
- added libimagequant API to add fixed preset colors to the palette
version 2.3
-----------
- added ./configure script for better support of Intel C compiler and dependencies [thanks to pdknsk]
- tweaked quality metric to better estimate quality of images with large solid backgrounds [thanks to Rolf Timmermans]
- avoid applying quality setting to images that use palette already
version 2.2
-----------
- OpenMP acceleration
- improved support for Intel C Compiler, speedup in 32-bit GCC, and some workarounds for Visual Studio's incomplete C support
version 2.1
-----------
- option to generate posterized output (for use with 16-bit textures)
version 2.0
-----------
- refactored codebase into pngquant and standalone libimagequant
- reduced memory usage by further 30% (and more for very large images)
- less precise remapping improving speed by 25% in higher speed settings
- fixed regression in dithering of alpha channel