Log in

View Full Version : An unexpected property of b=0, c=1 cubics


Katie Boundary
3rd September 2017, 06:54
A few months or years ago, I was playing around with different settings for cubic filters and noted that despite the "B+2C=1" rule, the cubics that gave the best overall results seemed to be at about b=0, c=1.

Today, I was playing around with filter shapes in Desmos and noticed something very peculiar: at x=1, the tangent of a raw sinc function is -1, and the tangent of a b=0, c=1 cubic is ALSO -1. This is also where both filter shapes are at their steepest. That's the reason why cubics start to look like ass beyond c>1: the filter shape gets "too steep" (defined as steeper than raw sinc) at x=1.

Katie Boundary
6th September 2017, 04:54
To demonstrate, I made a pretty:

https://scontent-lax3-1.xx.fbcdn.net/v/t31.0-8/21427171_1912650838999369_1973563675983984927_o.jpg?oh=d8667dbd69ce3bd349c759dba9abeedd&oe=5A17F4B9