Posted By
TLC
on 2010-12-15
12:26:13
|  Re: Vertical splitting
Guys, thanks for your kind words  , well... The example routine really wasn't meant to be something extraordinal . The part which skips badlines is a pretty old trick, people have been exploiting that (...if not just for the sake of skipping badlines, then for FLD effects, which is the same league; you just need to do some wiggling of the $ff1f register) for long. The other part (cycle exact timing + mode change at some predictive point of a line) has already been exploited by Luca in his original code, as far as I can see.
Luca: the routine only depends on your timing routine at its "start"; the cycletime of the rest of the code is constant, ie. doesn't depend on anything but itself. Just make sure that first write to $ff06 (in the IRQ routine, of course) happens two or three characters "before" the character area /under the left border of the first non-badline of the visible screen (ie. the second visible line; the first visible line of the screen is a badline). From that on, the routine is strongly supposed to work just as it is.
You can't do anything with the first pixels row, that row is a badline (the second badline of the first two badlines of the screen), ie. it is needed, and it can't be broken in half. You can possibly edit your gfx so that that row is masked (black, or anything). The colours should be normal as far as I can guess -- with the exception of the fact that you have to make it with the set of colours set by the first row (as a consequence of skipping badlines; colour ram is never read, as long as the skipping of badlines is in effect).
Litwr: the host has been down for possibly half an hour or so (it should be back now)... sorry for the inconvenience.
|
