mikepound / convolve Goto Github PK
View Code? Open in Web Editor NEWSimple demonstration of separable convolutions
License: The Unlicense
Simple demonstration of separable convolutions
License: The Unlicense
It’s an honor to learn from your convolve!I want to know that can lens be combined with coc?
Context: https://twitter.com/rickbrewPDN/status/1562499054952534018
The kernel parameters listed at https://github.com/mikepound/convolve/blob/master/complex_kernels.py are a set of 6 hardcoded 4-vector arrays. I've been having difficulty determining how these are actually generated, and how I might generate kernel parameters beyond the 6 that are listed here (as in, how to calculate for 7, 8, 9, etc.). The explanations I've found are not close enough to my comprehension of the math involved ("it's just a Gaussian kernel with a complex width" ... ??!! 😂 from: http://yehar.com/blog/?p=1495). Every example of Bokeh blur code I've found online lists the same set of pregenerated kernel values, with no code showing how to actually generate them.
My motivation is to experiment with higher "quality" settings in Paint.NET's new Bokeh blur filter. "Quality" is a slider in the UI that controls which set of kernel parameters are used, which is then tied to the number of rendering passes.
Ringing artifacts are present when there's an imbalance between quality, radius, and the gamma exponent (default is 3). I'm hoping to extend the range at which these values can be combined before artifacts are noticeable (or even eliminating them, if possible). My rendering pipeline is fully 32-bit floating point per pixel component (128-bits total), so there's a lot of room for extra precision vs. the typical 8-bit RGBA32.
My dream solution here would be a function that takes in an integer for the # of components, and returns an array of 4 vectors for the kernel parameters. For example, generate_kernel_parameters(2)
would return [[0.886528, 5.268909, 0.411259, -0.548794], [1.960518, 1.558213, 0.513282, 4.56111]]
(the numbers from https://github.com/mikepound/convolve/blob/master/complex_kernels.py). I'm not sure how feasible or computationally expensive that is though.
A declarative, efficient, and flexible JavaScript library for building user interfaces.
🖖 Vue.js is a progressive, incrementally-adoptable JavaScript framework for building UI on the web.
TypeScript is a superset of JavaScript that compiles to clean JavaScript output.
An Open Source Machine Learning Framework for Everyone
The Web framework for perfectionists with deadlines.
A PHP framework for web artisans
Bring data to life with SVG, Canvas and HTML. 📊📈🎉
JavaScript (JS) is a lightweight interpreted programming language with first-class functions.
Some thing interesting about web. New door for the world.
A server is a program made to process requests and deliver data to clients.
Machine learning is a way of modeling and interpreting data that allows a piece of software to respond intelligently.
Some thing interesting about visualization, use data art
Some thing interesting about game, make everyone happy.
We are working to build community through open source technology. NB: members must have two-factor auth.
Open source projects and samples from Microsoft.
Google ❤️ Open Source for everyone.
Alibaba Open Source for everyone
Data-Driven Documents codes.
China tencent open source team.