Improving the NTSC Colour Model
This article is a continuation of the series discussing the Atari 2600 colour generation model. In the first article I described how each colour in the 2600 palette can be thought of as a position on a colour wheel. In particular, I focused on how each hue in the palette is positioned at a fixed location relative to the colourburst and according to a phase. I described how the phase can be adjusted by the colour delay potentiometer and there was a discussion about what the “correct” phase value should be.
In that first article I also described how using a preset palette of RGB values was unsatisfactory because the phase of the palette could not be altered. I now realise that this is a false conclusion and the phase of a preset palette can be altered and indeed, can produce superior results. This article describes a new model that makes use of a standard RGB palette while maintaining the ability to adjust the palette’s phase.
Before looking at how phase adjustment of an RGB palette is possible, let us first examine the problems created by what I will refer to in this article as the mathematical model.