This c++ class has been derived from Charlie Tangora's public domain "gif.h" project, which you can find here. In that spirit, this is also public domain. More of this, less of requiring people to do, or not do, this or that. :)
Basically, you prepare a series of 32-bit (4 RGBA bytes per pixel) images arranged as rows going from the top of the image to the bottom. The A (alpha) channel is ignored, but must be there anyway. Each of the RGBA channels is one byte, with a range of 0 (dark) to 255 (light.)
This class can do color selection for you, but there are better color selectors out there. If you have such a better color selector, then:
-
Pre-select a GifPalette for each image you provide and supply it to GIFWriteFrame along with the image
-
Pre-reduce each image to just those 24-bit colors that are contained in the palette you're supplying.
This will bypass all of the palette selection code in this class.
You can't mix the two approaches: either you supply a pre-selected palette for every frame where each image contains only those colors, or you don't supply a palette and the image can be any set of 24-bit colors.
After the setup as explained below, you pass each image to GifWriteFrame()
and when they've all been submitted, you call GifEnd()
.`
Skeletal usage example:
#include "gifanim.h"
{
GifAnim ganim;
GifWriter g;
// Set up filename,
// xw=N, yw=N,
// framedelay=1...1000 or so,
// loopcount =0 means infinite, otherwise as specified
// bitdepth=8,
// dither=false (most likely, but you pick)
// each image is an array of bytes 0-255, in order ((R G B A) x xw) x yw)
// ----------------------------------------------------------------------
// and then:
// ---------
ganim.GifBegin(&g, fileName, xw, yw, framedelay, loopcount, bitdepth, dither);
// For each frame of the animation, set up RGBA-formatted image, then:
// -------------------------------------------------------------------
ganim.GifWriteFrame(&g, image, xw, yw, framedelay, bitdepth, dither);
// After all frames, this:
// -----------------------
ganim.GifEnd(&g);
}
My fervent thanks to Charlie Tangora for his original public domain c implementation; he saved me a ton of work on my iToolBox Project. Converting his c to a solid c++ class wasn't bad at all; figuring out GIF... that would have been quite the undertaking!
For the record, I'm using this class exactly as provided here inside a
Qt c++ project for OS X/MacOS, and it works great there as-is,
although I do use a custom memory manager rather than the bare calloc()
and free()
used here. Nothing you have to worry about, I'm just into
firewalling and tracking memory during development so as to catch as many
problems as possible, as early as possible. But because I do use such a
memory manager, this code has been through rigorous memory behavior
validation inside my
iToolBox Project,
and so I have good reason to think you can confidently trust it to be
memory-safe.