DitherIt 🎨

DitherIt is a comprehensive Dart library that implements various dithering algorithms for image processing. Transform your images with professional-grade dithering techniques to reduce color depth while maintaining visual quality.

📸 Visual Examples

See the difference dithering makes! Below are examples of each algorithm applied to the same source image:

Original Image

Original image (full color)

Floyd-Steinberg Dithering

Floyd-Steinberg Result
Floyd-Steinberg algorithm - Classic error diffusion

Ordered Dithering (Bayer Matrix)

Ordered Dithering Result
Ordered dithering - Regular pattern-based approach

Riemersma Dithering

Riemersma Result
Riemersma algorithm - Hilbert curve-based natural diffusion

📷 Image Credits

Example image: "Small Copper Butterfly" by Illuvis from Pixabay Used under the Pixabay Content License

See example/IMAGE_CREDITS.md for full attribution details.

✨ Features

Floyd-Steinberg Dithering: Classic error diffusion algorithm with fine-grained results
Ordered Dithering: Bayer matrix-based dithering with configurable matrix sizes (2x2, 4x4, 8x8)
Riemersma Dithering: Hilbert curve-based error diffusion for natural-looking results
High Performance: Optimized algorithms for fast processing
Easy to Use: Simple API with comprehensive documentation
Well Tested: Extensive test coverage ensuring reliability
Pure Dart: No native dependencies, works on all platforms

🚀 Getting Started

Installation

Add dither_it to your pubspec.yaml:

dependencies:
  image: ^4.2.0  # Required for image processing
  dither_it:
    git:
      url: https://github.com/Penfore/dither_it.git
      ref: main

Then run flutter pub get or dart pub get to install the package.

Run:

dart pub get

Quick Example

import 'package:dither_it/dither_it.dart';
import 'package:image/image.dart';
import 'dart:io';

void main() async {
  // Load an image
  final bytes = await File('input.jpg').readAsBytes();
  final originalImage = decodeImage(bytes)!;

  // Apply Floyd-Steinberg dithering
  final ditheredImage = DitherIt.floydSteinberg(image: originalImage);

  // Save the result
  await File('output_floyd.png').writeAsBytes(encodePng(ditheredImage));

  print('Dithering complete!');
}

🎯 Algorithms

Floyd-Steinberg Dithering

The Floyd-Steinberg algorithm distributes quantization error to neighboring pixels using a specific pattern:

final result = DitherIt.floydSteinberg(image: myImage);

Best for: General purpose dithering, photographs, smooth gradients

Ordered Dithering (Bayer Matrix)

Uses predefined threshold matrices for consistent, pattern-based dithering:

// 2x2 matrix (fastest, more visible pattern)
final result2x2 = DitherIt.ordered(image: myImage, matrixSize: 2);

// 4x4 matrix (balanced quality/performance)
final result4x4 = DitherIt.ordered(image: myImage, matrixSize: 4);

// 8x8 matrix (finest pattern, slower)
final result8x8 = DitherIt.ordered(image: myImage, matrixSize: 8);

Best for: Animations, real-time processing, consistent patterns

Riemersma Dithering

Uses a Hilbert curve pattern for more natural-looking error distribution:

// Default history size (16)
final result = DitherIt.riemersma(image: myImage);

// Custom history size for fine-tuning
final resultCustom = DitherIt.riemersma(image: myImage, historySize: 32);

Best for: Natural textures, organic images, artistic effects

📊 Algorithm Comparison

Algorithm	Speed	Quality	Pattern	Best Use Case
Floyd-Steinberg	Medium	High	Irregular	General purpose
Ordered (2x2)	Fast	Medium	Regular	Real-time, animations
Ordered (4x4)	Fast	Good	Regular	Balanced performance
Ordered (8x8)	Medium	High	Fine	High-quality output
Riemersma	Slow	Very High	Organic	Artistic, natural images

🔧 Advanced Usage

Processing Multiple Images

Future<void> processImageBatch(List<String> imagePaths) async {
  for (final path in imagePaths) {
    final bytes = await File(path).readAsBytes();
    final image = decodeImage(bytes)!;

    // Apply different algorithms
    final floyd = DitherIt.floydSteinberg(image: image);
    final ordered = DitherIt.ordered(image: image, matrixSize: 4);
    final riemersma = DitherIt.riemersma(image: image, historySize: 24);

    // Save results
    await File('${path}_floyd.png').writeAsBytes(encodePng(floyd));
    await File('${path}_ordered.png').writeAsBytes(encodePng(ordered));
    await File('${path}_riemersma.png').writeAsBytes(encodePng(riemersma));
  }
}

🎨 Examples

Check out our example code to get started quickly:

Basic Example

See example/dither_it_example.dart for a complete working example that demonstrates all three algorithms:

import 'dart:io';
import 'package:dither_it/dither_it.dart';
import 'package:image/image.dart';

void main() async {
  // Load an image
  final bytes = await File('example/input.png').readAsBytes();
  final image = decodeImage(bytes)!;

  // Apply Floyd-Steinberg dithering
  final floyd = DitherIt.floydSteinberg(image: image);
  await File('example/output_floyd_steinberg.png').writeAsBytes(encodePng(floyd));

  // Apply Ordered dithering
  final ordered = DitherIt.ordered(image: image, matrixSize: 4);
  await File('example/output_ordered.png').writeAsBytes(encodePng(ordered));

  // Apply Riemersma dithering
  final riemersma = DitherIt.riemersma(image: image);
  await File('example/output_riemersma.png').writeAsBytes(encodePng(riemersma));

  print('✅ Dithering complete! Check the output files.');
}

Visual Results

The visual examples above show the results of running this code. Compare how each algorithm handles the same input image differently!

Coming Soon

Additional examples in development:

comparison_example.dart - Side-by-side algorithm comparison with metrics (planned)
batch_processing.dart - Process multiple images efficiently (planned)
flutter_app_example/ - Complete Flutter app with UI (planned)

🔬 Technical Details

Algorithm Performance

Understanding the performance characteristics can help you choose the right algorithm for your use case:

Floyd-Steinberg

Processing Speed: Medium - processes each pixel sequentially
Memory Usage: Low - only needs to store error values for current and next row
Quality: High - excellent error diffusion creates smooth gradients
Best for: High-quality results when processing time is not critical

Ordered Dithering

Processing Speed: Fast - each pixel is processed independently
Memory Usage: Very Low - uses pre-computed threshold matrices
Quality: Good - creates consistent patterns, quality depends on matrix size
Best for: Real-time applications, animations, or when consistent patterns are desired

Riemersma

Processing Speed: Slower - maintains error history for each color channel
Memory Usage: Medium - stores error history (default: 16 values × 3 colors = 48 values per pixel)
Quality: Very High - natural-looking results with organic error distribution
Best for: Artistic applications where quality is more important than speed

Memory Requirements

Here's what each algorithm needs in terms of additional memory:

Floyd-Steinberg: ~4KB extra for a 1920×1080 image (stores one row of error values)
Ordered: No extra memory (uses static matrices)
Riemersma: ~300KB extra for a 1920×1080 image with default history size

Processing Time Examples

Approximate processing times for a 1920×1080 image on a modern CPU:

Ordered (2×2): ~50ms
Ordered (4×4): ~60ms
Ordered (8×8): ~80ms
Floyd-Steinberg: ~200ms
Riemersma: ~400ms

Note: Times vary based on hardware and image complexity

Parallelization Support

Ordered Dithering: ✅ Fully parallelizable - each pixel can be processed independently
Floyd-Steinberg: ❌ Sequential processing required due to error propagation
Riemersma: ❌ Sequential processing required due to error history dependencies

🤝 Contributing

We welcome contributions! Here's how you can help:

Bug Reports: Open an issue with detailed reproduction steps
Feature Requests: Suggest new algorithms or improvements
Code Contributions: Fork the repo and submit a pull request
Documentation: Help improve our docs and examples

Development Setup

# Clone the repository
git clone https://github.com/Penfore/dither_it.git
cd dither_it

# Install dependencies
dart pub get

# Run tests
dart test

# Run example (coming soon)
# cd example
# dart run basic_example.dart

Coding Standards

Follow Dart style guide
Maintain 100% test coverage for new features
Document public APIs with DartDoc
Use meaningful commit messages

🗺️ Roadmap

Upcoming Features

Additional error diffusion algorithms (Jarvis-Judice-Ninke, Stucki, etc.)
Blue noise dithering
Custom color palette support
Performance optimizations with isolates
Web demo application
CLI tool for batch processing

Version History

See CHANGELOG.md for detailed version history.

📝 License

This project is licensed under the BSD-3-Clause License - see the LICENSE file for details.

🙏 Acknowledgments

Robert Floyd and Louis Steinberg for the pioneering Floyd-Steinberg algorithm
Nico Riemersma for the Riemersma dithering technique
The Dart team for the excellent image processing foundation
Contributors who help improve this library

📞 Support

🐛 Issues: GitHub Issues
💬 Discussions: GitHub Discussions
📧 Contact: Create an issue for any questions

Made with ❤️ by Penfore

Star ⭐ this repository if you find it helpful!

DitherIt 🎨

📸 Visual Examples

Original Image

Floyd-Steinberg Dithering

Ordered Dithering (Bayer Matrix)

Riemersma Dithering

✨ Features

🚀 Getting Started

Installation

Quick Example

🎯 Algorithms

Floyd-Steinberg Dithering

Ordered Dithering (Bayer Matrix)

Riemersma Dithering

📊 Algorithm Comparison

🔧 Advanced Usage

Processing Multiple Images

🎨 Examples

Basic Example

Visual Results

Coming Soon

🔬 Technical Details

Algorithm Performance

Floyd-Steinberg

Ordered Dithering

Riemersma

Memory Requirements

Processing Time Examples

Parallelization Support

🤝 Contributing

Development Setup

Coding Standards

🗺️ Roadmap

Upcoming Features

Version History

📝 License

🙏 Acknowledgments

📞 Support

Libraries

dither_it package