Michael J. Mahon wrote:
David Empson wrote:
Linards Ticmanis <ticmanis@gmx.de> wrote:
Michael J. Mahon wrote:
The answer is in the schematic--the NTSC color signal is created
not by the multiple-of-3.58MHz pulse scheme that is the Apple II
standard, but by a chrominance modulator driven by the RGB signals.
Therefore, its color rendition is not dependent on its pixel rate.
Does this apply only to SHR modes, or also to IIe compatible modes?
Probably all modes.
Agreed.
The IIgs RGB output emulates the colour behaviour of 8-bit Apple II
video modes, e.g. alternating green/violet pixels in HGR mode, with an
odd coloured pixel every 14th (yellow or brown).
The NTSC composite output can generate a proper colour burst derived
from the emulated colour, producing the right effect on NTSC, but with
proper colour encoding rather than a simulated colour burst and having
to rely on timing tricks.
Yes, that's my take on it, too.
Just for fun, I think I'll take a look at the IIgs composite
video with a scope...
Well, *that* was interesting!
I should have *started* by looking at the video, not the schematic!
First, the composite signal:
In SHR mode (the GSOS desktop), the chrominance signal is a phase-
modulated 3.58MHz sine wave--produced by the chroma modulator. For
the light blue background, the luminance signal is dominated by an
8MHz signal, the result of dither at a 16MHz pixel rate. White and
black are constant levels. The dark blue (the "standard" HGR color
6) shows no dither, and is a constant luminance level plus the 3.5MHz
chrominance subcarrier for "blue".
In HGR mode, the composite signal is *not* the standard digital
Apple II video, but is also luminance plus a 3.5MHz chrominance
subcarrier--so all composite video is generated in this way.
I also took a look at the analog RGB signals for the desktop, and
they are clearly dithered at a 16MHz pixel rate. (For colors that
have a constant component, like "traditional" HGR colors, there is no 8MHz
dither signal.)
So for any conversion of either digital or analog RGB signals to
another scan standard, it will be necessary to sense the presence
of the 8MHz dither signal and switch the pixel clock to 16MHz to
avoid a nasty 8/7 beat in the sampling that will produce 40 vertical
bars across the screen in regions of constant dithered color.
Apparently the VGC *does* contain a PLL to synthesize 16MHz from the
28MHz clock.
What an amazing and, yes, *alien* video generation the IIgs has!