mirror of
https://github.com/a2-4am/4cade.git
synced 2024-08-28 14:29:22 +00:00
74 lines
2.8 KiB
Python
Executable File
74 lines
2.8 KiB
Python
Executable File
#!/usr/bin/env python3
|
|
|
|
import PIL.Image # https://pillow.readthedocs.io/
|
|
import util
|
|
|
|
# mandelbrot.png is the source image. The source image MUST have a white background,
|
|
# but other colors and pixel depth are irrelevant. This one is black & white.
|
|
# Due to the Apple II pixel aspect ratio, we do a 1-time aspect-ratio-losing resize
|
|
# to squash the image to 87% height.
|
|
#
|
|
# $ gm convert mandelbrot.png -resize "100%x87%!" squash.png
|
|
# (Depending on your shell, you may need to escape the exclamation point. Grr.)
|
|
#
|
|
# Now we can create individual images for each "frame" of the animation, by
|
|
# resizing the (squashed) source image and putting it in a 280x192 frame.
|
|
#
|
|
# $ for w in `seq 1 1500`; do \
|
|
# gm convert -size 280x192 squash.png \
|
|
# -resize "$w" \
|
|
# -background white \
|
|
# -compose Copy \
|
|
# -gravity center \
|
|
# -extent 280x192 thumb"$w".png; \
|
|
# done
|
|
#
|
|
# Depending on the source image, you may need more or fewer than 500 frames. This
|
|
# number is duplimandelbroted below in the |frames| variable. Sorry.
|
|
#
|
|
# Now we have 1500 (or so) PNG images of what the HGR screen should look like at
|
|
# each stage. Despite each frame being 280x192 and in the correct aspect ratio,
|
|
# only coordinates
|
|
# - on every 3rd row
|
|
# - on even columns
|
|
# are included in the final data set.
|
|
#
|
|
# X coordinates are converted to byte+bitmask (but see notes below).
|
|
# Y coordinates are flipped (so 0,0 ends up on the bottom left) then
|
|
# divided by 3.
|
|
#
|
|
# 6502 code will be responsible for plotting each of these coordinates
|
|
# in a 2x3 block. The bitmask usually includes 2 adjacent pixels;
|
|
# the code will also plot the same 2 adjacent pixels in the next two rows.
|
|
#
|
|
# Unfortunately, since bytes are 7 pixels across, some of the 2-pixel-wide
|
|
# blocks will cross a byte boundary. To simplify the 6502 code, these are
|
|
# simply listed as separate coordinate pairs, each with a bitmask that
|
|
# includes 1 pixel instead of 2.
|
|
|
|
frames = 1500 # number of "thumbN.png" files
|
|
|
|
coords = []
|
|
for i in range(5, frames, 5):
|
|
p = PIL.Image.open("mandelbrot/thumb%s.png" % i)
|
|
for x in range(0, 280, 2):
|
|
for y in range(0, 192, 3):
|
|
if p.getpixel((x,191-y)) == (0,0,0):
|
|
coords.append((x,y))
|
|
|
|
unique_coords = util.unique(coords)
|
|
unique_vals = util.vals_3bit(unique_coords)
|
|
with open("../../../src/fx/fx.hgr.mandelbrot.data.a", "w") as f:
|
|
for aval, bval in unique_vals:
|
|
f.write(" !byte %s,%s\n" % (aval, bval))
|
|
|
|
ripple_vals = util.ripple(unique_vals)
|
|
with open("../../../src/fx/fx.hgr.mandelbrot.ripple.data.a", "w") as f:
|
|
for aval, bval in ripple_vals:
|
|
f.write(" !byte %s,%s\n" % (aval, bval))
|
|
|
|
unique_vals.reverse()
|
|
with open("../../../src/fx/fx.hgr.mandelbrot.in.data.a", "w") as f:
|
|
for aval, bval in unique_vals:
|
|
f.write(" !byte %s,%s\n" % (aval, bval))
|