Re: NES - Apple Sprite card

[mjmahon@aol.com (Michael J. Mahon)]

Roy replied:

>Michael J. Mahon wrote:
>> Roy replied:
>> 
>>>Yves McDonald wrote:
>>>
>>>>Right on, I have the sprite card at home but never tried it.  It was 
>>>>designed by Steve Ciarcia and introduced in the late Byte magazine. 
>>>>Resolution was inferior to Apple's high res mode (256 horizontal pixels 
>>>>vs 280) but it definitely had more colors (up to 16).  All you could 
>>>>program were sprites of 8x8 or 16x16 pixel resolution.  To build an 
>>>>image such as game backdrop, you have to 'tile' sprites.
>>>
>>>I'm not sure exactly what you mean by "tile".
>>>The 9918/28 VDP had 32 different "layers" that were "stacked" one on top 
>>>of another. A single sprite only could be on a layer. The result was you 
>>>could "paint" a scene on the lowest "layer" and various sprites moving 
>>>about so that you had a 3D effect (a sprite shaped like a plane would 
>>>move "behind" a sprite shaped like a cloud, or in front of another "cloud
>> 
>> What Yves is referring to is that there is no "pixel-mapped" layer.
>> The background layer is made up of 8x8 or 16x16 "characters"
>> or "tiles" which are mapped into the display space (where visible).
>
>My manual on the 9918 is out on loan at the moment, but I thought the 
>lowest level was a pixel-mapped layer.
>
>Certainly, if you did a graphic image in BASIC on either the TI or the 
>Coleco, sprites were never used. The ADAM's BASIC had exactly the same 
>commands as AppleSoft, so that any graphics work done in AppleSoft (that 
>didn't use Peeks or Pokes of course) could be done on the ADAM.

It is possible to do general graphics with the 9918--it just is a
mess compared to "normal" Apple graphics.  (And the peculiar
mapping of the Apple hi-res screen is regarded by many machine-
language programmers as a problem in itself!  ;-)

When the 9918 is operated in Graphics II mode, there are enough
character patterns so that it is possible to control the state of each
of the 256 x 192 pixels on the pattern plane.  The pixel information
is divided up into two tables, the pattern (character or tile) data
and the color table.  With this mapping set up, it is possible to
treat the pattern plane as pixel-mapped, with a very strange relation
between x,y coordinates and the character memory.  The addresses
in video memory would not be simply related to the coordinates,
and a lookup table would be required for any speed (even worse than
the Apple hi-res screen).

Of course, video memory is not directly addressable, so after the
updates have been computed, and the addresses, then each video
memory byte must be updated by polling it for "not busy" status,
then writing two address bytes and one data byte to the 9918 chip.

All things considered, much more trouble than a memory-mapped,
semi-linear graphics system like the Apple II.

-michael

Check out 8-bit Apple sound that will amaze you on my
Home page:  http://members.aol.com/MJMahon/