LORES

Each dot-matrix pixel can be individually controlled. Each pixel can have 1 out of 2 colors. Each pixel is 2x1.

* X-Expanded multi color sprites. (4x1)
* Sprite colors controls the 2 colors.
* Chars over sprites
* ECM+MC = black

Memory waste:
   0-511 2pages

But only every 8th byts is used.

Resulting in heavy fragmentation of the 2 pages, but effectivly only 64 bytes(!!!) is wasted.

All in all, lores, no display bugs, low memory waste, easy timing. GOOD!!!

Problem not solved yet, how do we blank out every second line? One solution would be to bank switch but that
would effectivly double the amount of memory needed. Another solution is perhaps to actually change the sprite colors...

solution : put d027-d02e colors black, and d025/d026 yellow/brown, and switch multicolor/singlecolor each line (asl $d016, lsr $d016)

--------------------------------------------------------------------------------------


SEMI-HIRES

Each dot-matrix pixel can't be individually controlled, but 2 instead. Each 2 pixels can have 1 out of 2 colors.

* X-Expanded multi color sprites. (4x1)
* Sprite colors controls the 2 colors.
* Chars over sprites
* ECM-mode+Bitmap = black

Memory waste:
   0- 511 2pages
2048-2559 2pages
4096-4607 2pages
6144-6655 2pages
----------------
          8pages

But only every 8th byte is used => 8/8 = 1 page (256 bytes)

Resulting in modest fragmentation of the memory.
Bytes spreaded out over 32 blocks, but only blocks 0-1, 8-9, 16-17 and 24-25 are affected, and only every 8th byte in those.
Effective memory waste is 256 bytes. Realistic memory waste is 2048 bytes.

All in all, not perfect, but still good looking I think, no display bugs, modest memory waste, easy timing. GOOD!!!!

----------------------------------------------------------------------------------------


REAL-HIRES

Each dot-matrix pixel can be individually controlled with one out of 2 colors.

* X-Expanded single colors sprites. (2x1)
* $d020 to one of the colors.
* Sprites to the other color.
* Chars over sprites
* Bitmap mode

Memory waste:
0-8191 = 32 pages

But only every 8th byte is used => 32/8 = 4 pages (1024 bytes)
Resulting in heavy fragmentation of memory.
Every 8th bytes is used in an 32 block area.

Colors:
To guarantee correct colors badline FLI has to be inserted at the very last line of the display to fetch color data.
This needs an additional 1024 bytes of totally wasted memory, but will contain the FLI bug resulting in bad colors for the first three chars...
That FLI has also to be timed perfectly which will be very tricky since the ball, flippers etc actually can disrupt it.

Effective memory waste: 2048 bytes.

All in all, not even perferct (FLI-bug), total memory waste, tricky timing. BAD!!!

(Color bug might me solvable since the color depends on what opcode that comes after the $d011 write, we just want it black... not gray, not green... which is common)

------------------------------------------------------------------------------------------
