I didn't see any code in that link, but thanks for that!
(Edit) I spotted the code.
His example of 64 x 64 pixels is a total of 4096. My screen shows 160 x 480 pixels which is a total of 76,800 pixels.
So there is a trade-off. For much fewer pixels you probably have time to output colour information.
I'll be documenting the design decisions, and how it all works. I found that time was extremely critical in trying to output even 20 characters (160 pixels) horizontally. For example, I had to restructure the font table in such a way as to save a single divide by 8 (or shift right 3 bits).
It's not even as if assembler would help. It takes 16 clock cycles to output 8 pixels, so you have a limited amount of time to work out what the next 8 pixels are going to be.