v0.4.3

Animated GIF Encoding with Partial Frames and Transparency

Fixed animated GIF encoding to properly handle partial frames (frames smaller than canvas at offset positions).

Changes:

• Frame positions (left, top) are now correctly preserved
• Frame dimensions use actual frame size instead of canvas size
• Transparency is properly handled in partial frames
• Disposal methods (none, background, previous) are preserved on round-trip

Backward Compatible - No breaking changes. Frame metadata is optional with sensible defaults.

v0.4.2

Added

• JPEG coefficient extraction and encoding API for steganography support:
  • Image.extractCoefficients() - Extract quantized DCT coefficients from JPEG images
  • Image.encodeFromCoefficients() - Re-encode JPEG from modified coefficients
  • JPEGFormat.extractCoefficients() - Format-specific coefficient extraction
  • JPEGFormat.encodeFromCoefficients() - Format-specific coefficient encoding
  • JPEGDecoder now supports extractCoefficients option to store quantized values
  • JPEGEncoder.encodeFromCoefficients() - Encode pre-quantized DCT coefficients
  • New types: CoefficientData, JPEGQuantizedCoefficients, JPEGComponentCoefficients
  • Comprehensive tests for coefficient extraction and encoding in
    test/formats/jpeg_coefficients.test.ts

Fixed

• Documentation: Added browser support notes to README.md and docs/src/index.md to clarify
  in-browser support.

0.4.1

Fixed

• Documentation: Corrected format support tables to accurately reflect TIFF compression support
  (PackBits and Deflate are pure-JS, not native-only)
• Documentation: Added missing ICO format to all format support tables in docs/src/formats.md
• Documentation: Added missing GIF implementation documentation link

Added

• CMYK color space support with conversion utilities:
  • rgbToCmyk() and cmykToRgb() functions for converting individual color values
  • rgbaToCmyk() and cmykToRgba() functions for converting entire image data arrays
  • Image.toCMYK() method to convert an image to CMYK representation
  • Image.fromCMYK() static method to create an image from CMYK data
• TIFF CMYK support:
  • Automatic decoding of CMYK TIFFs (photometric interpretation 5) with conversion to RGBA
  • cmyk: true encoding option in TIFFEncoderOptions to save images in CMYK color space
  • Works with all TIFF compression methods (uncompressed, LZW, PackBits, Deflate)
• PNG encoder now supports compressionLevel option (0-9) to control encoding speed vs. file size
  through adaptive PNG filter selection
• PNGEncoderOptions, GIFEncoderOptions, and APNGEncoderOptions types exported from mod.ts
  for API completeness and consistency
• GIF encoder now supports loop option to control animation loop count (0 = infinite loop, 1+ =
  specific count)
• Base64 utilities (encodeBase64, decodeBase64) and data URL helpers (toDataUrl,
  parseDataUrl) exported from mod.ts

Full Changelog: v0.4.0...v.0.4.1

0.4.0

Fixed

• Progressive JPEG decoding now correctly accumulates DCT coefficients across multiple scans by
  deferring IDCT until all scans complete, fixing issues with half-decoded or corrupted-looking
  progressive images
• Successive approximation in progressive JPEGs now properly handled: high-bit scans (Ah=0, Al>0)
  and refinement scans (Ah>0) correctly accumulate bit-precision, eliminating blocky/low-resolution
  appearance in images with successive approximation encoding
• DC predictor now correctly resets at the start of each scan per JPEG specification, preventing
  coefficient accumulation artifacts across scans

Added

• Progressive JPEG support in pure JavaScript decoder - can now decode both baseline (SOF0) and
  progressive (SOF2) JPEGs without requiring runtime APIs, including full successive approximation
  bit-refinement support
• createImageBitmap + OffscreenCanvas fallback for JPEG decoding in environments with canvas
  APIs available
• Comprehensive tests for progressive JPEG decoding in test/formats/jpeg_progressive.test.ts

Changed

• Removed all console.log, console.warn, and console.error statements from library code;
  non-fatal warnings are reported through ImageDecoderOptions.onWarning
• Fallback behavior (pure-JS to runtime APIs) is silent by default
• Image.decode and Image.decodeFrames now accept ImageDecoderOptions (tolerantDecoding,
  runtimeDecoding, onWarning)
• Image.encode("jpeg", ...) now honors quality and progressive via JPEGEncoderOptions
  (exported from mod.ts)
• TIFFEncoderOptions now exported from mod.ts (no longer re-exported from the TIFF format
  module)

Breaking

• Renamed *EncodeOptions types to *EncoderOptions for consistency with ImageDecoderOptions
• ASCIIOptions removed; use ASCIIEncoderOptions
• Encoding APIs (Image#encode, Image.encodeFrames, and ImageFormat.encode*) no longer accept
  decode options
• Per-decoder option types (JPEGDecoderOptions, GIFDecoderOptions, WebPDecoderOptions) removed
  in favor of ImageDecoderOptions

Full Changelog: v0.3.0...v0.4.0

0.3.0

[0.3.0] - 2025-12-13

Added

• Fault-tolerant decoding modes for GIF, WebP, and JPEG formats
• GIFDecoderOptions interface with tolerantDecoding option for frame-level
  error recovery
• WebPDecoderOptions interface with tolerantDecoding option for pixel-level
  error recovery
• JPEGDecoderOptions now exported from main module for advanced usage
• Comprehensive documentation for fault-tolerant decoding in README.md
• Tests for fault-tolerant modes in GIF and WebP decoders

Fixed

• JPEG roundtrip encoding now clamps DCT coefficients to valid Huffman table
  ranges, preventing "Invalid Huffman code" errors when re-encoding decoded
  JPEGs

Performance

• Optimized JPEG encoder/decoder with typed arrays (Float32Array, Int32Array,
  Uint8Array) for better memory efficiency and performance
• Optimized WebP encoder/decoder with typed arrays (Uint8Array, Uint32Array) for
  improved performance
• Reduced memory allocations in hot paths for DCT/IDCT operations
• Optimized image processing operations with lookup tables and reduced Math
  function calls
• Optimized resize operations (nearest, bilinear) with bitwise operations and
  pre-computed values
• Optimized rotation and flip operations using Uint32Array views and batch
  copying
• Optimized crop operation with row-based batch copying
• Optimized composite operation by reducing redundant calculations in inner
  loops
• Optimized Gaussian blur with pre-computed offsets and reduced Math function
  calls

Full Changelog: v0.2.4...v0.3.0

0.2.4

What's Changed

• [WIP] Add metadata extraction for non-decodable images in #60
• Add HEIC and AVIF format support with metadata extraction in #61
• Add format, compression, and frame count to metadata extraction in #62
• [WIP] Ensure HEIC and AVIF have consistent metadata extraction byin #63
• docs: align documentation with implemented API in #64
• Add repository improvements: CONTRIBUTING, CHANGELOG, .editorconfig, and enhanced npm package in #65
• Add getSupportedMetadata to all format handlers in #67

Full Changelog: v0.2.3...v0.2.4

v0.2.3

0.2.3