PNG

library written in ts to sample pixels from common png formats

support

chunks supported: IHDR, IDAT, IEND

algorithms supported: scanline filtering algorithm

compilation

if you want to try the test yourself, you can compile and run the project

to compile run:

1. install the initial libs if first run npm i .
2. compile the ts file, do tsc in the project's root folder.
3. run the compiled js file using node src/main.js.

some test images are in the files/ folder, the output image will be in the out folder.

i'll soon enough make it a library that can be included in other projects

usage

here is a use case example, of a function in the library this samples pixels for tests

async function sample_rectangle(x: number, y: number, width: number, height: number) {
  
  let sampler = new png_sampler();
  await sampler.init_sampler("files/minecraft_0.png");

  // preallocate the pixels
  let sampledPixels = Buffer.alloc(width * height * sampler.bytesPerPixel);

  console.time("sampling pixels");
  for (let j = 0; j < height; j++) {
    for (let i = 0; i < width; i++) {

      const pixel = sampler.sample_pixel(x + i, y + j);
      const offset = (j * width + i) * sampler.bytesPerPixel;
      sampledPixels.set(pixel, offset);
    }
  }
  console.timeEnd("sampling pixels");

  create_png(sampledPixels, width, height, "output.png");
}

what is a PNG?

for context,

a PNG file is made of chunks, a chunk is made up of a header, data, and a footer. header is 8 bytes, 4 bytes for chunk length, 4 bytes for chunk type. footer is 4 bytes of CRC32 checksum

the first chunk in a PNG file is the IHDR chunk. IHDR holds header info about the image

IHDR has fields

width, height 

depth - how many bits per channel, each pixel has channels. 

colorType - 0 is greyscale, 2 is RGB, 3 is palette, 4 is greyscale and alpha, 6 is RGBA. 

compression - 0 represents deflate / inflate, a lossless compression algorithm that uses LZ77 and huffman. 

filter - a filter method used before compression to make the compression more efficient 
         there are 5 filter methods. 

interlace - is the image loaded with interlace? format for faster image loading, but bigger IDAT size. 

next there may be the chunks sRGB, pHYs, PLTE, tRNS, IDAT, IEND, etc..

IDAT holds the pixels themselves and scanlines, note that each scanline is preceeded by a filter byte used in the filter algorithm

there may be many IDAT chunks that together form the image's data.

IEND holds no data, it just holds a header and footer.

IHDR, IDAT, IEND are the least you need, together with the magic number at the beginning.

here is the standard

http://www.libpng.org/pub/png/spec/1.2/PNG-Chunks.html#C.Summary-of-standard-chunks

example

here is an example of the format i wrote up in paint using ImHex.

for context, each two letters are two hex values, each represented by 4 bits. and together takeup 8 bits, a byte.