Log In  



Cart #34411 | 2016-12-29 | Code ▽ | Embed ▽ | License: CC4-BY-NC-SA
17

Here's a toy I have been playing around with. Perhaps somebody can use these functions for fancy shaded 3D rendering or something like that.

I have a few functions that draw "RGB" pixels to the screen using pattern dither that are approximated using the rgb values of the pico8 palette.

 
color_bayer_plot_8(x,y,r,g,b) -- writes a given r,g,b at the x,y (r,g,b are from 0 to 1)

fast_rgb_create() -- pre-renders patterns using a 4X4 bayer dither representing 8 levels of R and B and 16 levels of G.
fast_rgb(x,y,r,g,b) -- writes a given r,g,b at the x,y (r,g,b are from 0 to 1)

function solid_trifill( x1,y1,x2,y2,x3,y3, r0,g0,b0,r1,g1,b1,r2,g2,b2)
Renders a triangle with the three corners colored at rgb and shades smoothly between them. Right now this uses the slower color_bayer_plot_8.

If you move the cursor around with the arrow keys you can see the triangles. The vertices can be grabbed with the "z" key and the color of the vertices can be changed with the "x" key.

Originally I had been thinking about making a gradient image compression algorithm, but it seemed less interesting after getting half-way through.

Here's a spinning R G B triangle for fun.

Cart #34413 | 2016-12-29 | Code ▽ | Embed ▽ | License: CC4-BY-NC-SA
17

17

You never cease to amaze me, dude.

Brilliant! Thank you for doing this.

suggestion: you can save a lot of tokens by directly importing the look-up table generated by your cart (attached below) into the spritesheet, then eliminating every method except for fast_rgb().

Or if you'd rather trade tokens for keeping your spritesheet intact, you could replace fast_rgb_create() and fast_rgb() with these modified functions, which precalculate and pack the LUT into 8kb of RAM instead:

 
rbitmasks={{shr,12},
          {shr,8},
          {shr,4},
          {shr,0},
          {shl,4},
          {shl,8},
          {shl,12},
          {shl,16}}
wbitmasks={{shl,12},
          {shl,8},
          {shl,4},
          {shl,0},
          {shr,4},
          {shr,8},
          {shr,12},
          {shr,16}}       

bayer_lut={}
function	fast_rgb_create()
	local sx=0
	local sy=0

	local bx=0
	local by=0

 local bit,mask,addr

	--init lut
	for j=0,127 do
	 for i=1,16 do
	  bayer_lut[j*16+i]=0
	 end
	end

	for g=0,15 do
		for r=0,7 do
			for b=0,7 do			
				for i=0,3 do
					for j=0,3 do
						x=(g%4)*32+(r%8)*4+i
						y=flr(g/4)*32+(b%8)*4+j
						c=color_bayer_8( x, y,r/7,g/15,b/7)
						bit=x%8+1
						x=flr(x/8)
						addr=y*16+x+1
						mask=wbitmasks[bit]
						bayer_lut[addr]+=mask[1](c,mask[2])
						--sset(x,y,c)
					end
				end
				sx+=4

			end
			sx=0
			sy+=4
		end
	sx=0
	sy=0
	end
end

function	fast_rgb(x,y,r,g,b)

		local x=flr(x)
		local y=flr(y)

  local bit,mask

		local g=flr(g*15)
		local r=flr(r*7)
		local b=flr(b*7)

		local sx=(g%4)*32+(r%8)*4+x%4
		local sy=flr(g/4)*32+(b%8)*4+y%4
	 bit=sx%8+1
		sx=flr(sx/8)
		mask=rbitmasks[bit]
		pset(x,y,band(mask[1](bayer_lut[sy*16+sx+1],mask[2]),0xf))
end



[Please log in to post a comment]