v2.0.0

Advisories and Breaking Changes

Major breaking changes

• Due to the new unmanaged memory engine, we had to redesign our pixel processing and low level memory management API-s. The most visible change is that the GetPixelRowSpan() methods have been replaced with ProcessPixelRows() variants. This was done to guarantee memory safety. For more details see:
  • 2.0 documentation on Working with Pixel Buffers
  • 2.0 documentation on Memory Management
  • Pixel processing breaking changes & API discussion - #1739
• Png and Jpeg now decode to pixels formats closer to the encoded format when using the non-generic API. (For example Jpeg now decodes to Image<Rgb24> instead of Image<Rgba32>.) This was done to reduce memory usage and remove some surprises for users.
• We've dropped the old NetStandard 1.3 target. This targeted frameworks not supported by Microsoft and was adding overhead to development.

Complete list of breaking PR-s

• Add Abgr32 pixel type, extend IPixel - #1886 via @ynse01, @antonfirsov, @JimBobSquarePants
• Remove netstandard 1.3 target - #1888 via @JimBobSquarePants
• ImageFormatDetector Configuration Cleanup - #1936 via @JimBobSquarePants
• Unmanaged pooling MemoryAllocator - #1730 via @antonfirsov
• Png - Preserve Pixel Format for Non-generic decode. - #1861 via @JimBobSquarePants
• Decode Jpeg into Rgb24 when pixel type not specified. - #1773

Release Notes

API

• Remove HistogramEqualizationOptions.Default - #1840 via @antonfirsov
• Added missing CancellationToken parameters to Image - #1856 via @justinbhopper
• Expose ImageFrame.PixelBuffer - #1906 via @JimBobSquarePants
• Change the constructor of PixelTypeInfo and ImageMetadata from internal to public so that a user can implement IImageInfo - #1929 via @jz5, @antonfirsov, @JimBobSquarePants

Area:Build

• Update codcov and config - #1820 via @JimBobSquarePants
• Use a scheduled job for codecov - #1934 via @JimBobSquarePants

Area:Formats

• Add support for Portable Bitmap images - #1851 via @ynse01, @antonfirsov, @brianpopow, @JimBobSquarePants

Area:Performance

• Webp: Use byte arrays instead of Dictionary's for lookups - #1800 via @brianpopow
• Add sse2 version of select - #1804 via @brianpopow
• Add SSE2 versions of ClampedAddSubtractFull and ClampedAddSubtractHalf - #1805 via @brianpopow, @JimBobSquarePants
• SSE41 version of Hadamard transform - #1810 via @brianpopow
• Add SSE41 version of Quantize block - #1811 via @brianpopow
• Add SSE2 version of Mean16x4 - #1814 via @brianpopow
• Add SSE2 version of Vp8Sse4X4 - #1817 via @brianpopow
• Hoist some of the calculations from loops of 3DMoments() - #1818 via @kunalspathak, @brianpopow
• Add sse2 version of inverse transform - #1819 via @brianpopow, @JimBobSquarePants
• Add AVX version of CollectColorBlueTransforms and CollectColorRedTransforms - #1824 via @brianpopow, @JimBobSquarePants
• Add AVX version of TransformColor and TransformColorInverse - #1830 via @brianpopow, @JimBobSquarePants
• Webp improvements - #1846 via @brianpopow
• Jpeg new IDCT & fused zigzag/transpose steps - #1847 via @br3aker, @JimBobSquarePants
• Add AVX2 version of AddVector - #1849 via @brianpopow
• Add SSE41 version of UpSample - #1836 via @brianpopow, @JimBobSquarePants
• Add AVX2 versions of CombinedShannonEntropy - #1848 via @brianpopow
• Add SSE2 version of FTransform - #1852 via @brianpopow, @JimBobSquarePants
• Add AVX2 version of Vp8_Sse4X4 - #1860 via @brianpopow
• Jpeg codebase cleanup - #1853 via @br3aker
• Add AVX2 version of QuantizeBlock - #1857 via @brianpopow, @JimBobSquarePants
• Add SSE2 version of TransformOne and TransformTwo - #1862 via @brianpopow, @JimBobSquarePants
• Add SSE2 version for webp Filter methods - #1871 via @brianpopow, @JimBobSquarePants
• Add AVX2 version of CollectHistogram - #1882 via @brianpopow
• Add SSE2 and AVX2 versions of Vp8_Sse16X16 and Vp8_Sse16X8 - #1881 via @brianpopow, @JimBobSquarePants
• Add AVX2 version of GetResidualCost and SSE2 version of SetCoeffs - #1902 via @brianpopow
• SSE2 / AVX2 of webp AlphaDecoding - #1930 via @brianpopow, @JimBobSquarePants
• Jpeg encoder optimization via removing second transpose call from FDCT - #1958 via @br3aker

Bug

• Fix gaussian blurring changing the blur colour when transperancy is used - #1834 via @Turnerj
• Make ColorDistanceCache non-readonly - #1841 via @antonfirsov, @JimBobSquarePants
• Use Rgb24 for Async Jpeg Decode - #1859 via @JimBobSquarePants
• Fix Color conversion for differing types - #1865 via @JimBobSquarePants
• Allocate clean buffer in AnalyzeEntropy - #1873 via @brianpopow
• Fixed jpeg encoder last bytes flushing bug - #1876 via @br3aker
• Tiff: Add maximum IFD threshold - #1892 via @brianpopow
• Fixed Jpeg ...

v1.0.4

Advisories and Breaking Changes

• None

Release Notes

Bug Fixes

• Fix a few uses of DeflateStream.Read - #1707 via @stephentoub

v1.0.3

Advisories and Breaking Changes

• None

Release Notes

API Enhancements

• Add new PixelAlphaRepresentation property and implement for TPixel types - #1420 via @JimBobSquarePants, @antonfirsov
• New Image.WrapMemory(void*) overloads - #1419 via @Sergio0694, @JimBobSquarePants
• Add PremultiplyAlpha to ResizeOptions - #1504 via @ptasev
• Change size guards for WrapMemory to allow Memory/IMemoryOwner blocks to be larger than the image size - #1526 via @jeffska, @JimBobSquarePants
• New overloads for binary threshold operations. The new argument Color… - #1555 via @Franz1960, @JimBobSquarePants
• Introduce Numerics class and Refactor Clamp Utils - #1436 via @JimBobSquarePants
• Pixel remapping/Swizzling - #1388 via @Evangelink, @JimBobSquarePants

Performance

• PixelOperations.From speedup - #1435 via @Sergio0694
• Implement PackFromRgbPlanes for Rgba32 and Rgb24 - #1462 via @antonfirsov, @JimBobSquarePants
• Optimize Convolution - #1465 via @JimBobSquarePants
• Optimize bokeh blur convolution - #1475 via @Sergio0694, @JimBobSquarePants
• 1D convolution optimization and general codegen tweaks - #1477 via @Sergio0694
• Much Faster sRGB Companding - #1481 via @JimBobSquarePants
• Speed improvements to resize kernel (w/ SIMD) - #1513 via @Sergio0694
• Use bulk conversion in Write8BitColor - #1428 via @brianpopow
• Vectorize Scale16X16To8X8 - #1517 via @tkp1n
• JpegEncoder - Optimize Some Low Hanging Fruit - #1533 via @JimBobSquarePants
• Vectorize Jpeg Encoder Color Conversion - #1508 via @tkp1n
• Fix vectorized 4:2:0 subsampling - #1551 via @tkp1n
• Reduce code duplication due to reified generics - #1444 via @Sergio0694
• Add check for Avx2 in ResizeKernel - #1548 via @brianpopow

Bug Fixes

• Sanitize Gif descriptor bounds. - #1532 via @JimBobSquarePants
• Use GetPixelRowSpan to access pixel data in Histogram equalization - #1431 via @brianpopow
• Ensure Span length of source and destination are equal during pixel conversion - #1443 via @brianpopow
• Fix for Issue #1505 - #1506 via @brianpopow, @JimBobSquarePants
• Filter processor should use scaled vectors. - #1463 via @maxvoxel8

Other

• Fix link + other docs tweaks. - #1478 via @JimBobSquarePants
• Small README.md edit - #1535 via @kendaleiv
• Update shared infra structure to treat image file types by git lfs - #1558 via @brianpopow
• Enable resize tests on MacOs - #1434 via @JimBobSquarePants
• Add xunit trait attributes to tests - #1455 via @brianpopow
• Benchmark Code Quality Cleanup - #1479 via @JimBobSquarePants
• Enable support to use libgdiplus library on macOS (splitted) - #1424 via @AlexNDRmac
• Add cache for Nuget packages - #1439 via @AlexNDRmac, @antonfirsov
• Add net5.0 targetframework - #1519 via @stefannikolei
• Better JpegEncoder profiling & benchmarks - #1534 via @antonfirsov
• Fix Code Coverage - #1554 via @JimBobSquarePants
• Update to use new props/targets plus cleanup. - #1451 via @JimBobSquarePants
• Add action to save Output Images for failed tests - #1452 via @AlexNDRmac
• Use correct shared ruleset path. - #1464 via @JimBobSquarePants

v1.0.2

Advisories and Breaking Changes

• None

Release Notes

Bug Fixes

• Use interest for target bounds. Fixes #1342 - #1343 via @JimBobSquarePants
• Fixed typo in AdaptiveThreshold - #1362 via @rold2007
• Add reference output for #1362 - #1363 via @brianpopow
• Fixed setting the file size when the BMP encoder has a color palette - #1384 via @dlemstra, @JimBobSquarePants
• Fix ARM Build for UWP Release - #1391 via @JimBobSquarePants

Area:Api

• Auto repair Png options to use Bit8. - #1395 via @JimBobSquarePants
• Add missing SaveAsync method - #1334 via @JimBobSquarePants
• Image.WrapMemory APIs wrapping Memory - #1314 via @Sergio0694, @JimBobSquarePants

Area:Performance

• Add Avx backed Block8x8F Transpose method - #1374 via @JimBobSquarePants
• Optimize Block8x8F low hanging fruit and fix naming - #1390 via @JimBobSquarePants
• Add SimdUtils.HwIntrinsics - #1398 via @JimBobSquarePants
• Add Avx2 Vector4 Span Premultiplication and Reverse - #1399 via @JimBobSquarePants
• Add AVX2 Vector4Octet.Pack implementation - #1402 via @JimBobSquarePants
• 3 <==> 4 Channel Shuffling with Hardware Intrinsics - #1409 via @JimBobSquarePants
• Vectorize (AVX2) JPEG Color Converter - #1411 via @tkp1n, @antonfirsov, @JimBobSquarePants
• Bokeh blur constructor tweaks (input validation, tiny speedup) - #1332 via @Sergio0694
• Improved codegen in ImageSharp.Guard - #1333 via @Sergio0694, @JimBobSquarePants
• Add 4 Channel Shuffle Intrinsics - #1404 via @JimBobSquarePants
• Optimize TryGetSinglePixelSpan - #1407 via @JimBobSquarePants
• Bump ChunkedMemoryStream buffer length - #1328 via @JimBobSquarePants

Area:Build

• Fix compatibility with NET 5 SDK InternalsVisibleTo - #1337 via @JimBobSquarePants
• Use Ubuntu for Code Coverage. - #1377 via @JimBobSquarePants
• Update Dependencies - #1378 via @JimBobSquarePants
• Upgraded Magick.NET. - #1383 via @dlemstra, @JimBobSquarePants
• Fix JpegDecoderTests.Decode_IsCancellable - #1415 via @antonfirsov
• Advanced Feature Testing - #1385 via @JimBobSquarePants

v1.0.1

Advisories and Breaking Changes

• None

Release Notes

Bug

• Simplify jpeg encoder and fix for Windows ARM - #1304 via @JimBobSquarePants

• Fix non-seekable stream reading - #1316 via @JimBobSquarePants

Other

• Add GitHub Sponsors - #1311 via @JimBobSquarePants

v1.0.0

Advisories and Breaking Changes

Area:Api

• Better Edge Detection API - #1299 via @JimBobSquarePants
• Cancellable codec API, and overloads for loading/saving - #1296 via @antonfirsov, @JimBobSquarePants

Release Notes

Area:Api

• LoadAsync methods accepting string path. - #1257 via @pekspro, @JimBobSquarePants
• Async version of DiscoverDecoder - #1260 via @pekspro, @JimBobSquarePants
• Save as extensions - #1256 via @pekspro, @antonfirsov, @JimBobSquarePants

Area:Formats

• Buffered Decoding. - #1269 via @JimBobSquarePants

Area:Performance

• Make stream buffer size configurable. - #1286 via @JimBobSquarePants

Bug

• Change position guard from MustBeLessThanTo to MustBeLessThanOrEqualTo - #1291 via @brianpopow

Area:Build

• Update all dependency references. - #1262 via @JimBobSquarePants

Area:Build

• Better Builds - #1245 via @JimBobSquarePants

Documentation

• Add submodule info to CONTRIBUTING.md - #1277 via @wilka, @JimBobSquarePants

1.0.0-rc.3

Area:Api

• Async APIs / Fake Async Codecs - #1196 via @tocsoft, @JimBobSquarePants

Bug

• Only throw for multi SOF when fully decoding. - #1226 via @JimBobSquarePants
• Update Adler32 to correctly filter intrinsics. - #1229 via @JimBobSquarePants
• Use ConcurrentDictionary for options properties. - #1234 via @JimBobSquarePants

Area:Build

• Bump Gitversion - #1232 via @JimBobSquarePants

Other

• Update README.md - #1216 via @shaggygi
• Use MemoryAllocator in FixedCapacityPooledMemoryStream - #1223 via @antonfirsov
• Tests for FixedCapacityPooledMemoryStream - #1225 via @antonfirsov
• Added Third Party Notices document - #1236 via @Turnerj, @JimBobSquarePants

1.0.0-rc.2

Release Notes

Reintroduces the Apache 2.0 License as the open source license.

Area:Performance

• Add Hardware Accelerated Checksums - #1201 via @JimBobSquarePants

Bug

• Return input checksum value on empty buffer - #1203 via @JimBobSquarePants
• Swap buffers when encoding PNG with Filtered Adam7 Palette mode - #1212 via @JimBobSquarePants
• Change Scanlines and SamplesPerLine from short to int - #1197 via @brianpopow
• Ensure min dimension of 1px following rounding. - #1198 via @JimBobSquarePants

Enhancement

• Added: ability to skip unneeded chunks for optimization mode - #1012 via @equinox2k, @brianpopow, @JimBobSquarePants

Area:Build

• Use official Codecov action - #1214 via @JimBobSquarePants

Formats:Tga

• Additionally checks in DetectFormat for TGA files - #1210 via @brianpopow

Other

• Minor readme typo. "commmon" -> "common" - #1200 via @Jaben, @JimBobSquarePants
• Add Discussions Links - #1213 via @JimBobSquarePants

1.0.0-rc.1

Release Notes

Milestone: 1.0.0-rc1

Area:Api

• ExifTagDescriptionAttribute doesnt need fields - #978 via @SimonCropp, @JimBobSquarePants
• Refactor IDisposable pattern - #1009 via @JimBobSquarePants
• Expose visitor through advanced namespace. - #1010 via @JimBobSquarePants
• Make processors public, refactor cloning. - #1011 via @JimBobSquarePants
• Expose pixelblenders. #967 - #1028 via @JimBobSquarePants
• Expose Buffer2D and ParallelHelper internals - #1035 via @antonfirsov, @JimBobSquarePants
• Expose final internals to allow decoupling Drawing. - #1046 via @JimBobSquarePants
• Allow returning the image format with the info. - #1059 via @JimBobSquarePants
• Add La16 and La32 IPixel formats. - #1062 via @JimBobSquarePants
• Public Vector4 to/from conversions - #1050 via @Sergio0694, @JimBobSquarePants
• Allow passing custom Configuration instances to processors - #1070 via @JimBobSquarePants
• Simplify and cleanup build constants - #1075 via @JimBobSquarePants
• Merge SixLabors.Core sources into ImageSharp - #1087 via @antonfirsov, @DaniaalNadir, @dlemstra, @iamcarbon, @JimBobSquarePants, @Mertsch, @synercoder, @tocsoft
• Expose Pixel Operations and Configuration - #1095 via @JimBobSquarePants
• Rename Pixel Shader Extensions and Delegates - #1096 via @JimBobSquarePants
• Add new Color parsing methods. - #1104 via @JimBobSquarePants
• Single row value delegates for parallel iterator - #1108 via @Sergio0694
• Remove Closure Allocations when Iterating Rows in Parallel - #1107 via @JimBobSquarePants, @Sergio0694
• Bokeh blur cache optimization - #1113 via @Sergio0694
• Fix quantization and dithering. - #1114 via @JimBobSquarePants, @brianpopow
• Implement discontiguous buffer handling - #1109 via @antonfirsov, @brianpopow, @JimBobSquarePants
• Generic Octet struct - #1131 via @Sergio0694
• Refactoring and tweaks (2) - #1122 via @Sergio0694, @JimBobSquarePants
• Single row value delegate option - #1132 via @Sergio0694
• Text segment ReadOnlySpan initialization - #1133 via @Sergio0694, @JimBobSquarePants
• Dither and Quantize API and Performance Improvements. - #1138 via @JimBobSquarePants, @antonfirsov
• Add overloads. Fix #1144 - #1146 via @JimBobSquarePants
• Add image extension method for save as tga - #1155 via @brianpopow
• Clean Up Namespaces and Visibility - #1161 via @JimBobSquarePants
• Improve accessability of Span methods. - #1172 via @JimBobSquarePants
• Update ToBase64String - #1175 via @JimBobSquarePants
• Allocation-free IMemoryGroup enumeration - #1173 via @Sergio0694, @JimBobSquarePants
• Add exception type and update xml docs. - #1179 via @JimBobSquarePants
• Add Property bag to Configuration and Image Processing Context - #1176 via @tocsoft, @JimBobSquarePants
• Change ImageFormatException to InvalidImageContentException - #1190 via @brianpopow
• Introduce PngCompressionLevel - #1192 via @JimBobSquarePants
• BufferRegion => Buffer2DRegion - #1194 via @JimBobSquarePants

Area:Bug

• Fix RGB to HSL conversion - #1001 via @JimBobSquarePants
• Fix #999 and add tests - #1003 via @JimBobSquarePants
• Fix #1004 - #1005 via @JimBobSquarePants, @antonfirsov
• Prevent PathGradientBrush from throwing an error with corner cases - #1007 via @mysticfall
• Fix horizontally out-of-bounds error when drawing text - #1008 via @Turnerj
• Change histogram clipLimit to be an absolute value - #1017 via @brianpopow, @JimBobSquarePants
• Png decoder fix for Issue #1014 - #1020 via @brianpopow
• Update the hex value for DarkSeaGreen color - #1021 via @zappyzhao, @JimBobSquarePants
• Fix reading of pngs with corrupt IEND chunks. - #1055 via @JimBobSquarePants
• Fix #1006 Resize from large source to small thumbnail produces deformed top pixel rows - #1060 via @JimBobSquarePants
• Js/fix namespace - #1066 via @JimBobSquarePants
• EXIF - Post review fixes - #1069 via @JimBobSquarePants
• Enumerate once on clone. Fix #1073 - #1074 via @JimBobSquarePants
• Remove invalid Subsampling checks. - #1078 via @JimBobSquarePants
• Use per-edge buffer + minor optimizations. - #1079 via @JimBobSquarePants
• Update Exif LensInfo - #1093 via @JimBobSquarePants
• Default Gif RepeatCount to 1 - #1100 via @JimBobSquarePants
• Break when minY >= bottom - [#1106](https://github...

1.0.0-beta7

• #816 - Update dependencies + fix stylecop enhancement contributed by JimBobSquarePants
• #819 - Respecting fileHeader Offset, skipping bytes if necessary bug formats:bmp contributed by brianpopow
• #820 - Track input stream position on load. bug formats:jpeg contributed by JimBobSquarePants
• #830 - Fix Decoding 8-Bit interlaced grayscale png's with alpha channel contributed by brianpopow
• #831 - Fix Decoding interlaced grayscale png's contributed by brianpopow
• #832 - Add support for Decoding BI_ALPHABITFIELDS contributed by brianpopow
• #833 - Allow Build with Visual Studio For Mac build contributed by JimBobSquarePants
• #834 - Remove hardcoded subsampling. bug formats:jpeg contributed by JimBobSquarePants
• #835 - GDI+ stability fixes for the benchmark and tests contributed by Lakritzator
• #836 - Jpeg Fuzz Fixes bug formats:jpeg contributed by JimBobSquarePants
• #841 - Lazily derive the huffman tables. Fix #839 bug formats:jpeg contributed by JimBobSquarePants
• #842 - Bokeh blur implementation contributed by Sergio0694
• #848 - Implement IEquatable<Block8x8F> and check when summing. bug formats:jpeg contributed by JimBobSquarePants
• #851 - Png Decoder Bounds check. Fix #849 bug formats:png contributed by JimBobSquarePants
• #853 - #852 Adding instructions to add submodules when cloning the repository documentation contributed by bulldetektor
• #855 - Updating some readonly static data in JpegEncoderCore to take advantage of compiler functionality. formats:jpeg performance contributed by tannergooding
• #863 - Remove Enum Comparison Boxing performance contributed by JimBobSquarePants
• #864 - Fix incorrect gradient color-stop calculation bug drawing contributed by Poyo-SSB
• #865 - Change existing gradient brushes to accept PointF drawing enhancement contributed by Poyo-SSB
• #869 - Introduce extended pixel conversion api enhancement performance contributed by antonfirsov
• #870 - Nits codequality contributed by iamcarbon
• #874 - Use more accuracy when calculating variance. Fix #866 bug quantizers contributed by JimBobSquarePants
• #877 - Now throws a better excpetion DrawImage source does not overlap target contributed by atruskie
• #879 - Nits - Metadata documentation metadata:exif metadata:icc metadata:iptc metadata:xmp contributed by iamcarbon
• #880 - Nits - Colorspaces bug colorspaces performance contributed by iamcarbon
• #883 - ImageBrush shouldn't Dispose of the image it is using. bug drawing contributed by BorisTheBrave
• #884 - Nits - Benchmarks codequality contributed by iamcarbon
• #885 - Handle incorrect colorspace metadata. bug formats:jpeg contributed by JimBobSquarePants
• #887 - Cleanup General Convolution api bug performance contributed by JimBobSquarePants
• #888 - Limit ResizeProcessor memory consumption enhancement memory performance contributed by antonfirsov
• #889 - Bitmap encoder writes V3 header as default bug formats:bmp contributed by brianpopow
• #891 - Fix travis builds build contributed by antonfirsov
• #893 - Pass correct output size source image is smaller than the Target in ResizeMode.Min [bug](https://api.github.com/repos/...