Re: Double hires mode color artifacts

[Dirk Thierbach <dthierbach@usenet.arcornews.de>]

Michael J. Mahon <mjmahon@aol.com> wrote:
> Linards Ticmanis wrote:

>> As regards the actual Apple hardware, double hires simply drives out a
>> high voltage for every pixel that's set and a low voltage for every
>> pixel that's not set, using seven pixels from each byte, at a speed of
>> (theoretically) 14,318,181.8181... pixels per second.
>> That's all there is to it. The End.

I was beginning to suspect this, but wasn't sure. Thanks, that was the
information I needed.

>> I hope I have amused you, if not I hope I have confused you. ;-)

Neither, you were quite helpful :-)

> This analysis is what would happen in an NTSC monitor if the two chroma
> channels had infinite bandwidth.  In fact, they have quite limited
> bandwidths, on the order of 1MHz, and the two channels have different
> bandwidths!

> In addition, the bandwidth of the luminance channel is also limited, but
> to a larger value, on the order of 3MHz.

Is the limit implemented in the monitor? I tried to do some calculations
before going to bed, but must have confused something -- the standard DVD
format contains 720 pixels, so the TVs should be able to display this 
many different luminance samples. The output signal in the active area
is 52us long, makes 0.072us per pixel. If pixels alternate between
black and white, that's a bandwidth of about 7MHz for luminance. Or is
it just impossible to display alternating pixels at that resolution? I
must be making some mistake here...

> The result is that a larger window (more FIR samples) must be considered
> to adequately simulate the analog processing--in fact, closer to 14
> (which would correspond to about 1 microsecond of time).
[...]
> Since the effect of preceding and following dots has a rapidly
> diminishing effect, 

And that gets to the core of the problem: How strongly do the extra
samples influence the hue? Because...

> I'm not aware of anyone taking the time to do this, but it would be
> appropriate for a machine-independent NTSC monitor emulation.  At
> one time, the MAME project seemed to be heading in this direction.

... I'm just in the process if re-implementing double hires for the
"apple-emul" or "apple2" emulator on Linux (after being fed up with the
fact that it only run under svga, or under X with 8bpp). I've implemented
hires in a reasonably fast way by using tables which guarantee 4 bit
samples for every pixel, and at the moment I am wondering if this is
enough for double hires, or if I have to make the window a bit larger
to be able to display the color artifacts with reasanble accuracy.

- Dirk