Git Product home page Git Product logo

tinydng's Introduction

Tiny DNG Loader and Writer library

Header-only simple&limited DNG(Digital NeGative, TIFF format + extension) loader and writer in C++11.

Currently TinyDNG only supports lossless RAW DNG and limited lossless JPEG DNG(no lossy compression support).

TinyDNG can also be used as an TIFF RGB image loader(8bit, 16bit and 32bit are supported).

TinyDNG loader module is being fuzz tested using LLVMFuzzer, and is enoughly secure(and no C++ exception and assert/abort code exists).

(NOTE: TinyDNG just loads DNG data as is, thus you'll need your own RAW processing code(e.g. debayer) to get a developed image as shown the above)

Features

Loading

  • RAW DNG data
  • Lossless JPEG
  • ZIP-compressed DNG
    • Use miniz or zlib
  • JPEG
    • Support JPEG image(e.g. thumbnail) through stb_image.h.
  • TIFF
    • 8bit uncompressed
    • 8bit LZW compressed(no preditor, horizontal diff predictor)
  • Experimental
    • Apple ProRAW(Lossless JPEG 12bit)
      • Lossless JPEG 12bit
      • Semantic map(8bit Standard JPEG)
    • Decode Canon RAW(CR2)
      • RAW
      • mRAW
      • sRAW
    • Decode Nikon RAW(NEF)
      • TODO
    • Reading custom TIFF tags.
  • Read DNG data from memory.

Writing

  • DNG and TIFF
    • LosslessJPEG compression

Supported DNG files

Here is the list of supported DNG files.

  • Sigma sd Quattro H
    • Uncompressed RGB 12bit image.
  • iPhone DNG
  • Apple ProRAW
    • Lossless JPEG 12bit
    • Semantic map
  • Black magic DNG
    • CinemaDNG(lossy compression) is not supported.
  • Canon CR2(experimental)
  • Magic lantern DNG
  • 8-bit TIFF image
    • LZW compressed 8-bit image.
  • 16-bit uncompressed TIFF image
  • 32-bit uncompressed TIFF image
  • OpCodeList
    • GainMap

Usage

Loading DNG

#include <cstdio>
#include <cstdlib>
#include <iostream>

// Define TINY_DNG_LOADER_IMPLEMENTATION and STB_IMAGE_IMPLEMENTATION in only one *.cc
#define TINY_DNG_LOADER_IMPLEMENTATION
#define STB_IMAGE_IMPLEMENTATION

// Enable ZIP compression(through miniz library)
// Please don't forget copying&adding `miniz.c` and `miniz.h` to your project.
// #define TINY_DNG_LOADER_ENABLE_ZIP

// Uncomment these two lines if you want to use system provided zlib library, not miniz
// #define TINY_DNG_LOADER_USE_SYSTEM_ZLIB
// #include <zlib.h>
#include "tiny_dng_loader.h"

int main(int argc, char **argv) {
  std::string input_filename = "colorchart.dng";

  if (argc > 1) {
    input_filename = std::string(argv[1]);
  }

  std::string warn, err;
  std::vector<tinydng::DNGImage> images;

  // List of custom field infos. This is optional and can be empty.
  std::vector<tinydng::FieldInfo> custom_field_lists;

  // Loads all images(IFD) in the DNG file to `images` array.
  // You can use `LoadDNGFromMemory` API to load DNG image from a memory.
  bool ret = tinydng::LoadDNG(input_filename.c_str(), custom_field_lists, &images, &warn, &err);


  if (!warn.empty()) {
    std::cout << "Warn: " << warn << std::endl;
  }

  if (!err.empty()) {
    std::cerr << "Err: " << err << std::endl;
  }

  if (ret) {
    for (size_t i = 0; i < images.size(); i++) {
      const tinydng::DNGImage &image = images[i];
;
      std::cout << "width = " << image.width << std::endl;
      std::cout << "height = " << image.height << std::endl;
      std::cout << "bits per piexl = " << image.bits_per_sample << std::endl;
      std::cout << "bits per piexl(original) = " << image.bits_per_sample_original << std::endl;
      std::cout << "samples per pixel = " << image.samples_per_pixel << std::endl;

    }
  }

  return EXIT_SUCCESS;
}

Writing DNG(and TIFF)

See examples/dngwriter and https://github.com/storyboardcreativity/zraw-decoder for more details.

#include <cassert>
#include <cstdio>
#include <cstdlib>
#include <iostream>

#define TINY_DNG_WRITER_IMPLEMENTATION
#include "tiny_dng_writer.h"

static void CreateRGBImage(tinydngwriter::DNGImage *dng_image,
                        const unsigned short basecol) {
  unsigned int image_width = 512;
  unsigned int image_height = 512;
  dng_image->SetSubfileType(false, false, false);
  dng_image->SetImageWidth(image_width);
  dng_image->SetImageLength(image_height);
  dng_image->SetRowsPerStrip(image_height);

  // SetSamplesPerPixel must be called before SetBitsPerSample()
  dng_image->SetSamplesPerPixel(3);
  uint16_t bps[3] = {16, 16, 16};
  dng_image->SetBitsPerSample(3, bps);
  dng_image->SetPlanarConfig(tinydngwriter::PLANARCONFIG_CONTIG);
  dng_image->SetCompression(tinydngwriter::COMPRESSION_NONE);
  dng_image->SetPhotometric(tinydngwriter::PHOTOMETRIC_RGB);
  dng_image->SetXResolution(1.0);
  dng_image->SetYResolution(1.2); // fractioal test
  dng_image->SetResolutionUnit(tinydngwriter::RESUNIT_NONE);
  dng_image->SetImageDescription("bora");

  std::vector<unsigned short> buf;
  buf.resize(image_width * image_height * 3);

  for (size_t y = 0; y < image_height; y++) {
    for (size_t x = 0; x < image_width; x++) {
      buf[3 * (y * image_width + x) + 0] = static_cast<unsigned short>(x % 512);
      buf[3 * (y * image_width + x) + 1] = static_cast<unsigned short>(y % 512);
      buf[3 * (y * image_width + x) + 2] = basecol;
    }
  }

  dng_image->SetImageData(reinterpret_cast<unsigned char *>(buf.data()),
                          buf.size() * sizeof(unsigned short));
}

int main(int argc, char **argv) {
  std::string output_filename = "output.dng";

  if (argc < 1) {
    std::cout << argv[0] << " <output.dng>" << std::endl;
  }

  if (argc > 1) {
    output_filename = std::string(argv[1]);
  }

  // TinyDNGWriter supports both BigEndian and LittleEndian TIFF.
  // Default = BigEndian.
  bool big_endian = false;

  if (argc > 2) {
    big_endian = bool(atoi(argv[2]));
  }

  {
    // DNGWriter supports multiple DNG images.
    // First create DNG image data, then pass it to DNGWriter with AddImage API.
    tinydngwriter::DNGImage dng_image0;
    dng_image0.SetBigEndian(big_endian);
    tinydngwriter::DNGImage dng_image1;
    dng_image1.SetBigEndian(big_endian);

    CreateRGBImage(&dng_image0, 12000);
    CreateRGBImage(&dng_image1, 42000);

    tinydngwriter::DNGWriter dng_writer(big_endian);
    bool ret = dng_writer.AddImage(&dng_image0);
    assert(ret);

    ret = dng_writer.AddImage(&dng_image1);
    assert(ret);

    std::string err;
    ret = dng_writer.WriteToFile(output_filename.c_str(), &err);

    if (!err.empty()) {
      std::cerr << err;
    }

    if (!ret) {
      return EXIT_FAILURE;
    }

    std::cout << "Wrote : " << output_filename << std::endl;
  }

  return EXIT_SUCCESS;
}

Customizations

  • TINY_DNG_LOADER_USE_THREAD : Enable threaded loading(requires C++11)
  • TINY_DNG_LOADER_ENABLE_ZIP : Enable decoding AdobeDeflate image(Currently, tiled RGB image only).
    • TINY_DNG_LOADER_USE_SYSTEM_ZLIB : Use system's zlib library instead of miniz.
  • TINY_DNG_LOADER_DEBUG : Enable debug printf(developer only!)
  • TINY_DNG_LOADER_NO_STB_IMAGE_INCLUDE : Do not include stb_image.h inside of tiny_dng_loader.h.
  • TINY_DNG_LOADER_NO_STDIO : Disable printf, cout/cerr.

Examples

Python binding(Experimental)

$ python -m pip install tinydng

Windows(including ARM), Linux(including aarch64) and macOS are supported.

See experimental/python for exsample python code.

Command line tools(Experimental)

When TinyDNG is installed from pip, CLI command tinydng is available.

$ tinydng input.dng

Fuzzing test

Resource

Here is the list of great articles on how to decode RAW file and how to develop RAW image.

TODO

  • Move to C++11.
    • Drop C++03 support.
  • Parse semantic map tags in Apple ProRAW.
  • Add DNG header load only mode
  • Parse more DNG headers
  • Parse more custom DNG(TIFF) tags
  • lossy DNG
  • Improve DNG writer
    • Support compression(LJPEG)
  • Support Big TIFF(4GB+)
  • Decode Nikon RAW(NEF)
  • Improve Canon RAW decoding
  • Optimimze lossless JPEG decoding
  • Delayed load of tiled image.

License

TinyDNG is licensed under MIT license.

TinyDNG uses the following third party libraries.

  • liblj92(Lossless JPEG library) : (c) Andrew Baldwin 2014. MIT license. https://bitbucket.org/baldand/mlrawviewer.git
  • stb_image : Public domain image loader.
  • lzw.hpp : Author: Guilherme R. Lampert. Public domain LZW decoder.
  • miniz : Copyright 2013-2014 RAD Game Tools and Valve Software. Copyright 2010-2014 Rich Geldreich and Tenacious Software LLC MIT license. See miniz.LICENSE

tinydng's People

Contributors

syoyo avatar storyboardcreativity avatar rajkundu avatar kuranes avatar wkjarosz avatar soma-arc avatar

Recommend Projects

  • React photo React

    A declarative, efficient, and flexible JavaScript library for building user interfaces.

  • Vue.js photo Vue.js

    ๐Ÿ–– Vue.js is a progressive, incrementally-adoptable JavaScript framework for building UI on the web.

  • Typescript photo Typescript

    TypeScript is a superset of JavaScript that compiles to clean JavaScript output.

  • TensorFlow photo TensorFlow

    An Open Source Machine Learning Framework for Everyone

  • Django photo Django

    The Web framework for perfectionists with deadlines.

  • D3 photo D3

    Bring data to life with SVG, Canvas and HTML. ๐Ÿ“Š๐Ÿ“ˆ๐ŸŽ‰

Recommend Topics

  • javascript

    JavaScript (JS) is a lightweight interpreted programming language with first-class functions.

  • web

    Some thing interesting about web. New door for the world.

  • server

    A server is a program made to process requests and deliver data to clients.

  • Machine learning

    Machine learning is a way of modeling and interpreting data that allows a piece of software to respond intelligently.

  • Game

    Some thing interesting about game, make everyone happy.

Recommend Org

  • Facebook photo Facebook

    We are working to build community through open source technology. NB: members must have two-factor auth.

  • Microsoft photo Microsoft

    Open source projects and samples from Microsoft.

  • Google photo Google

    Google โค๏ธ Open Source for everyone.

  • D3 photo D3

    Data-Driven Documents codes.