Digital Images Part 3: Color TVs

Remember that term color-space? We used RGB (Red, Green, Blue) last time. Now we'll talk about the color space your old TVs used (without any of the physics involved). 

Transient

We're going to put the values of pixels the Red, Green, and Blue color channels into a neat little list (some may recognize it as a vector…you know who you are):

purplevector

Two representations

purple

of the same information.

And we're going to multiply it by this matrix:

Cue handwaving about how to multiply matrices. ( Source,  and for the  curious )

Cue handwaving about how to multiply matrices. (Source, and for the curious)

To get this:

But what does that even mean?

The top value is referring to brightness on a scale from 0 to 1.


The bottom two values are giving x- and y- coordinates for this graph:

…Except that x is I and y is Q

…Except that x is I and y is Q

Try matching our color to a point on the map. See how close you get to the coordinates above. The color is going to look quite different in the crowd of all its friends.

So that's just one pixel. Imagine streams of those coming in as electrical signal through the cable that runs into your wall, and you've got the basic idea of how color TVs used to work (back in the analog days). That information and that conversion matrix has a lot of complicated physics behind it regarding color and wavelength: that matrix and the whole mechanism of color television is the work of years of laboratory research. Further reading: YIQ and color TV

So when you're out this weekend, raise a glass to the unsung heroes of color television.