What's Changing and Why

When using convolute() I'd expect the function to default to applying the given kernel to the whole image. However, the current function uses half the kernel size as default starting position, which I believe is a mistake.

Here's an example of current behavior when applying a 5x5 gaussian blur, zoomed in on the top-left:

I'd expect something like this:

Original image, zoomed out

What else might be affected

Shouldn't affect anything aside from fixing incorrect behavior

Tasks

• Add tests
• Update Documentation
• Update jimp.d.ts
• Add SemVer Label