🎉 Celebrating 25 Years of GameDev.net! 🎉

Not many can claim 25 years on the Internet! Join us in celebrating this milestone. Learn more about our history, and thank you for being a part of our community!

Back to Math and Physics

Gaussian based Weights

Math and Physics Programming LinearAlgebra

Started by akaitora January 06, 2018 05:22 AM

2 comments, last by alvaro 6 years, 6 months ago

akaitora

399

Author

January 06, 2018 05:22 AM

Hi. I was studying the Bloom tutorial on learnopengl.com.
https://learnopengl.com/#!Advanced-Lighting/Bloom

I understand the benefits of using a Gaussian blur for achieving a Bloom effect. I understand that the author of this tutorial is using the gaussian values as weights which are used to determine how much neighboring pixels should affect the current pixel when blurred. In this tutorial, the author chooses to use the following weights,

0.227027, 0.1945946, 0.1216216, 0.054054, 0.016216

Where 0.227027 + (0.1945946*2.0) + (0.1216216 * 2.0) + (0.054054 * 2.0) + (0.016216 *2.0) comes out to about 1. The author chooses to same 4 samples to the left and right of the current pixel for a total sample count of 9 per axis. All that makes sense but what I don't understand is how to calculate the weights myself. For example, lets say I wanted a sample count of some arbitrary number like 5.

With a sample count of 5, the formula would then need to be A + (B * 2.0) + (C * 2.0) + (D * 2.0) + (E * 2.0) = 1
What I want to do is figure out how to calculate the wights, A,B,C,D, and E given the sample count. Thanks!

xdpixel.com - Practical Computer Graphics

0r0d

2,059

January 06, 2018 06:50 AM

This might help:

http://dev.theomader.com/gaussian-kernel-calculator/

alvaro

21,604

January 07, 2018 08:56 AM

Here's a short C++11 program that roughly produces the same numbers:


	#include <cmath>
	#include <iostream>
	 
	int main() {
	  double sigma = 1.75;
	  int N = 4;
	  
	  double total = std::erf((N+.5) / (std::sqrt(2.0) * sigma)) - std::erf((-N-.5) / (std::sqrt(2.0) * sigma));
	  for (int i = -N; i <= +N; ++i) {
	    std::cout << (std::erf((i+.5) / (std::sqrt(2.0) * sigma)) - std::erf((i-.5) / (std::sqrt(2.0) * sigma))) / total << '\n';
	  }
	}

🎉 Celebrating 25 Years of GameDev.net! 🎉

Gaussian based Weights

This topic is closed to new replies.

Popular Topics

Recommended Tutorials

🎉 Celebrating 25 Years of GameDev.net! 🎉

Gaussian based Weights

This topic is closed to new replies.

Popular Topics

Recommended Tutorials

Reticulating splines