[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [MiNT] XaAES sources for FreeMiNT 1.16.3

on 11/29/09 9:25 AM, Helmut Karlowski wrote:

> This would be needed in XaAES. The XaAES snap, which you wrote, is able to
> store from a 16-bit-video, so there is a way (of course ...). Would you be
> so kind to let us know the algorithm you use? I'm sure this would save
> lots of time.

What I do is similar to ozk, but I test red, green, and blue separately.
Plot one red pixel, full brightness (1000) so you know it will be stored as
$ff. Examine the result and determine where exactly it placed the value and
if its intel versus motorola format.  Then plot a green pixel, note it's
placement, plot a blue pixel and note it's placement. I also generate bit
shift amounts for each component and shift direction flags for later use, to
speed up screen drawing.

You really only want to test values that produce $ff, because if you plot
VDI values that are under that (less than 1000) the VDI can do rounding and
the resulting examination may not yield the bit pattern expected.  Remember,
the VDI uses 0-1000 and not 0-255, so it very wise to plot only with 1000
values in all tests.

> BTW: XaAES-snap can't save 8-bit (which I can) nor 1-bit.

I think this predates my work on proper screen detection.  Have to look in
to this sometime.  

Lonny Pursell    http://www.bright.net/~gfabasic/