[chrisdew]

The current encoding is complete, so no future extensions are possible.

A fix would be to change:

  110rrrrr ggggbbbb - copy the last pixel and adjust RGB by (r-15, g-7, b-7)

to:

  1100rrrr ggggbbbb - copy the last pixel and adjust RGB by (r-7, g-7, b-7)

This leaves:

  1101???? ...

available for future extensions.

[HanClinto]

So I implemented this change and tested it out. Turns out, in my simple test suite, this actually changes the size of images. Not drastically. Often just a 5-10 kb difference (on a 500-600kb file), but that's still more than I expected for changing r-15 to r-7.

[phire]

Suggests the DIFF16 delta-color mode is responsible for quite a bit of the savings. Maybe it would be worth experimenting the exact encoding.

One idea would be to start a predicted color (calculate the delta of the previous two pixels, apply that, then specify a delta to that). Another would be to encode the delta in YUV or some other color space, and then experiment with the best balance of bits between those channels.

Perhaps it would be better to steal bits from RUN16 instead, I somewhat doubt it's usefulness.

[HanClinto]

I really like this.