canvas_ity
fpng
canvas_ity | fpng | |
---|---|---|
7 | 10 | |
318 | 830 | |
- | - | |
2.5 | 4.1 | |
2 months ago | 5 months ago | |
C++ | C | |
ISC License | - |
Stars - the number of stars that a project has on GitHub. Growth - month over month growth in stars.
Activity is a relative number indicating how actively a project is being developed. Recent commits have higher weight than older ones.
For example, an activity of 9.0 indicates that a project is amongst the top 10% of the most actively developed projects that we are tracking.
canvas_ity
-
Writing a TrueType font renderer
I have a small TTF implementation that's in the neighborhood of that size and is open source. It's part of my canvas_ity single-header library [0] that's around 2300 LOC / 36 KB object size and implements a C++ version of most of the 2D HTML5 canvas spec [1].
The core implementation of the TTF parsing and drawing is in L1526-L1846 with another small bit at L3205-L3274 of src/canvas_ity.hpp.
It's something of a toy implementation that only supports western left-to-right text, and doesn't do any hinting at all, nor kerning, nor shaping. But it's enough to draw a basic "Hello world!" using any typical TTF file.
The test suite in test/test.cpp L84-304 embeds a few custom Base64-encoded TTF files. They're small and only have a few glyphs but they do exercise a number of interesting edge cases in the OpenType TTF spec [2]. Have a look at the HTML5 port of the test suite at test/test.html in different browsers to see how their canvas implementations render those fonts.
[0] https://github.com/a-e-k/canvas_ity
[1] https://www.w3.org/TR/2015/REC-2dcontext-20151119/
[2] https://standards.iso.org/ittf/PubliclyAvailableStandards/c0...
-
The Lone Developer Problem
Agreed, that sort of documentation is pure gold when done well.
It's something I always try to pay forward by doing in my own code. For example, one of my own solo projects was an STB-style single-header -like rasterizer library for C++. I started the implementation half of the library with a short outline of the rendering pipeline's dataflow and the top-level functions responsible for each stage:
https://github.com/a-e-k/canvas_ity/blob/f32fbb37e2fe7c0fcae...
-
Ask HN: What have you created that deserves a second chance on HN?
https://github.com/a-e-k/canvas_ity
This is an STB-style single-header C++ library with no dependencies beyond the standard C++ library. In about 2300 lines of 78-column code (not counting blanks or comments), or 1300 semicolons, it implements an API based on the basic W3C specification to draw 2D vector graphics into an image buffer:
- Strokes and fills (with antialiasing and gamma-correct blending)
- Linear and radial gradients
- Patterns (with repeat modes and bi-cubic resampling)
- Line caps and line joins (handling high curvature)
- Dash patterns and dash offsets
- Transforms
- Lines, quadratic and cubic Beziers, arcs, and rectangles
- Text (very basic, but does its own TTF font file parsing!)
- Raster images (i.e., sprites)
- Clipping (via masking)
- Compositing modes (Porter-Duff)
- Drop shadows with Gaussian blurs
I also uncovered a number of interesting browser quirks along the way with the HTML5 port of my testing suite.
- Hello, PNG
- A tiny, single-header -like 2D rasterizer for C++
-
canvas_ity - A tiny, single-header <canvas>-like 2D rasterizer
Repository: https://github.com/a-e-k/canvas_ity
- Show HN: Canvas_ity – A tiny, single-header -like 2D rasterizer for C++
fpng
-
png crate gets an ultrafast compression mode, up to 4x faster decompression
When the QOI format was first announced it wasn't clear that was even possible while keeping PNG format compatibility. But the fpng and fpnge C/C++ libraries showed it was, and today you can take advantage of those advances in a general purpose PNG library in Rust!
-
Hello, PNG
CRC is a table and 5 lines of code. That's trivial.
>zlib is 23k lines
It's not needed to make a PNG reader/writer. zlib is massive overkill for only making a PNG reader or writer. Here's a tiny deflate/inflate code [2] under 1k lines (and could be much smaller if needed).
stb[0] has single headers of ~7k lines total including all of the formats PNG, JPG, BMP,. PSD, GIF, HDR, and PIC. Here's [1] a 3k lines single file PNG version with tons if #ifdefs for all sorts of platforms. Removing those and I'd not be surprised if you could not do it in ~1k lines (which I'd consider quite simple compared to most of todays' media formats).
>Of course they're not common formats so you're stuck with complex formats like PNG
BMP is super common and easy to use anywhere.
I use flat image files all the time for quick and dirty stuff. They quickly saturate disk speeds and networking speeds (say recording a few decent speed cameras), and I've found PNG compression to alleviate those saturate CPU speeds (some libs are super slow, some are vastly faster). I've many times made custom compression formats to balance these for high performance tools when neither things like BMPs or things like PNG would suffice.
[0] https://github.com/nothings/stb
[1] https://github.com/richgel999/fpng/blob/main/src/fpng.cpp
[2] https://github.com/jibsen/tinf/tree/master/src
- Quite OK Image is now my favorite asset format
-
Computing Adler32 Checksums at 41 GB/s
This was actually considered, and other libraries do ignore checksums, or at least have options to:
https://github.com/richgel999/fpng/issues/9
-
QOI – The Quite OK Image Format
In the other direction, you can target a subset of PNG to get less optimized images but with QOI-like encode and decode speed: https://github.com/richgel999/fpng
-
ReShade 5.0 Released!
On specific operations like screenshots for example the new version is WAY faster though. We were using stb_image_write.h but switched to fpng which promised 12-19x faster compression at smaller sizes. That and the fact that screenshot saving have now been given it's own thread so it now longer causes a small stutter like when it was on the main thread, means that screenshot are now near instantaneous.
-
QOI – The “Quite OK Image Format” for fast, lossless image compression
I think QOI inspired the creation of https://github.com/richgel999/fpng which creates standard PNGs and compares itself directly to QOI.
-
Losslessly compresses RGB and RGBA images to a similar size of PNG, while offering a 20x-50x speedup in compression and 3x-4x speedup in decompression
BTW, today I found this fpng-fast PNG writer. There is a comparison with QOI in the readme.
What are some alternatives?
nanovgXC - Lightweight vector graphics library implementing exact-coverage antialiasing in OpenGL
qoi - The “Quite OK Image Format” for fast, lossless image compression
tinf - Tiny inflate library (inflate, gzip, zlib)
fpnge - Demo of a fast PNG encoder.
Tephra - A modern, high-performance C++17 graphics and compute library based on Vulkan
libdeflate - Heavily optimized library for DEFLATE/zlib/gzip compression and decompression
art - @Bigfan/art is a React custom renderer for HTML5 Canvas.
oss-nvjpg - Hardware-accelerated JPEG decoding on the Nvidia Tegra X1
osxphotos - Python app to work with pictures and associated metadata from Apple Photos on macOS. Also includes a package to provide programmatic access to the Photos library, pictures, and metadata.
PNG-spec - Maintenance of the PNG specification
figlet-fonts - my collection of figlet / toilet ascii art fonts
php-qoi - QOI image encoder and decoder written in pure PHP