Programming:Convert CPC sprites to Plus hardware sprites

From CPCWiki - THE Amstrad CPC encyclopedia!
Revision as of 20:52, 1 February 2010 by Redbox (Talk | contribs) (Created page with 'The Plus hardware sprite format is different to the standard CPC sprite format, so sprites have to be converted if you want to use them as hardware sprites on the Plus. == The f…')

(diff) ← Older revision | Latest revision (diff) | Newer revision → (diff)
Jump to: navigation, search

The Plus hardware sprite format is different to the standard CPC sprite format, so sprites have to be converted if you want to use them as hardware sprites on the Plus.

The formats

Plus hardware sprites are 16x16 pixels in size and have 16 colours, with colour 0 always being transparent. Therefore, when converting standard CPC sprites it is easiest to convert them from MODE 0 because this mode supports a colour depth of 16 and two pixels are stored in one byte.

The pixel data on the standard CPC is stored in an awkward manner and a byte is organised as follows:

Pixels.png

The top row shows the two pixels and the bottom row shows how they are arranged in one byte on the standard CPC.

The pixel data on the Plus is organised in a straight forward manner being that each byte is a pixel and its value is the PEN colour. For example, the first two lines of a Plus sprite could be:

<geshi lang=Z80> 00,01,02,03,04,05,06,07,08,09,10,11,12,13,14,15 01,02,03,04,05,06,07,08,09,10,11,12,13,14,15,00 </geshi>

Conversion

To convert a sprite, you can use the converter utility on the DSK image (to load type RUN”DISC”) which can be downloaded below.

To use the converter utility, you must have first capture your standard CPC sprite (you can use the SPRITER utility on the disc) which must be 16 x 16 in size (8 bytes wide by 16 lines high). The converter utility will then display the sprites and convert them before saving the converted sprite to the disk.

Download

Media:Convspr.dsk - DSK image containing the converter utility

Source code

The source code for the converter (which will assemble in WinAPE and Maxam) is as follows:

<geshi lang=Z80>

Convert CPC sprite to Plus sprite
by Redbox, February 2010
Public Domain

org &8200

ld b,16 ;loop counter, sprite is 16 lines high

mainloop: push bc ;preserve loop counter

call doline ld hl,(scraddr) ;load HL with screen address call nextline ;work out next screen line down ld (scraddr),hl ;and store the screen address

pop bc ;restore loop counter djnz mainloop ;decrease B and jump back to loop if not 0

ret ;exit

doline: ld b,8 ;loop counter, sprite is 8 bytes wide = 16 pixels

lineloop: push bc ;store counter

ld hl,(scraddr) ;get screen location ld a,(hl) ;and load data from it into A

bit 7,a ;does bit 0 equal 1? if so, reset Z ld b,&1 ;pixel 0, bit 0 = &1 call nz,addbit ;and write it out to sprite memory

bit 3,a ;does bit 0 equal 1? if so, reset Z ld b,&2 ;pixel 0, bit 1 (%0010) = &2 call nz,addbit ;and write it out to sprite memory

bit 5,a ;does bit 0 equal 1? if so, reset Z ld b,&4 ;pixel 0, bit 2 (%0100) = &4 call nz,addbit ;and write it out to sprite memory

bit 1,a ;does bit 0 equal 1? if so, reset Z ld b,&8 ;pixel 0, bit 3 (%1000) = &8 call nz,addbit ;and write it out to sprite memory

call storeit ;store pixel at sprite address

ld a,&0 ;reset pixel ld (pixel),a ;and store reset

ld hl,(scraddr) ;get data from current screen location ld a,(hl) ;and load it into A

bit 6,a ;does bit 0 equal 1? if so, reset Z ld b,&1 ;pixel 1, bit 0 (%0001) = &1 call nz,addbit ;and write it out to sprite memory

bit 2,a ;does bit 0 equal 1? if so, reset Z ld b,&2 ;pixel 1, bit 1 (%0010) = &2 call nz,addbit ;and write it out to sprite memory

bit 4,a ;does bit 0 equal 1? if so, reset Z ld b,&4 ;pixel 1, bit 2 (%0100) = &4 call nz,addbit ;and write it out to sprite memory

bit 0,a ;does bit 0 equal 1? if so, reset Z ld b,&8 ;pixel 1, bit 3 (%1000) = &8 call nz,addbit ;and write it out to sprite memory

call storeit ;store pixel at sprite address

ld a,&0 ;reset pixel ld (pixel),a ;and store reset

ld hl,(scraddr) ;get screen address inc hl ;increase it ld (scraddr),hl ;store it for next time

pop bc ;restore loop counter djnz lineloop ;decrease B and jump back to loop if not 0

ret ;else return

addbit: push af ;preserve screen data ld a,(pixel) ;load A with pixel add b ;add what we've found to it ld (pixel),a ;store it again pop af ;restore screen data ret

storeit: ld hl,(spraddr) ;load HL with sprite address ld a,(pixel) ;load pixel into A ld (hl),a ;load sprite address with pixel value inc hl ;increase sprite address ld (spraddr),hl ;and store it for next time ret

nextline: ld a,l ;load A with L (low byte of screen address) sub 8 ;subtract the 8 bytes we've already done ld l,a ;and load it back into L ld a,8 ;load A with 8 add a,h ;which means we add &800 for the next pixel line down ld h,a ;write new address back to H ret nc ;and return if no overflow

ld de,&C050 ;else loop back round to top of screen plus one pixel line add hl,de ;by adding &C000+&50 to HL ret

scraddr: dw &C000

pixel: db &0

spraddr: dw &9000 </geshi>

See also