It turns out, that in such a case, a more or less valid color burst signal was present after the horizontal blanking signal.
If you place the characters 147 and 161 into the first two columns of the video memory (the columns are usually not visible), you can create a 2,8 µsec color burst signal (see pictures). If you set the machine clock to NTSC 4.43Mhz, all characters would show up in a fixed color. Unfortunately, a NTSC monitor will not sync at this frequency any more, due to the SB600 video circuit divider setup.
Also, without the ability of a phase shift for individual pixels, no other colors can be generated. (This was solved in the Apple II)
Instead, if your machine clock is other than the NTSC 4.43Mhz clock, and the monitor is still able to sync to the video signal, color effects can be generated. This is due the shift in frequency causing a continuous phase shift over the horizontal trace.
The Color test below show, how this looks like. The effective color depends on the clock frequncy and horizontal position of the pixel and may change form blue to red within one character or over larger sections.
If you remove the two bytes form the first two columns of the screen (color burst), the video signal returns to black & white.
Is this useful?, I don't know, probably not but interesting to know, that a standard SB600 is capable to enable/disable colors by software.
- ColorBurst.png (62.65 KiB) Viewed 7692 times
