Simplify
This commit is contained in:
kris
parent
20a73ffb8a
commit
e08f25e4cc
109
dither.pyx
109
dither.pyx
|
|
@ -55,8 +55,8 @@ cdef int dither_bounds_yb(Dither *dither, int y_res, int y) nogil:
|
|||
|
||||
@cython.boundscheck(False)
|
||||
@cython.wraparound(False)
|
||||
@functools.lru_cache(None)
|
||||
def lookahead_options(object screen, int lookahead, unsigned char last_pixel_nbit, int x):
|
||||
# @functools.lru_cache(None)
|
||||
cdef inline unsigned char lookahead_pixels(unsigned char last_pixel_nbit, unsigned int next_pixels, int lookahead):
|
||||
"""Compute all possible n-bit palette values for upcoming pixels, given x coord and state of n pixels to the left.
|
||||
|
||||
Args:
|
||||
|
|
@ -69,28 +69,25 @@ def lookahead_options(object screen, int lookahead, unsigned char last_pixel_nbi
|
|||
Returns: matrix of size (2**lookahead, lookahead) containing all 2**lookahead possible vectors of n-bit palette
|
||||
values accessible at positions x .. x + lookahead
|
||||
"""
|
||||
cdef unsigned char[:, ::1] options_nbit = np.empty((2 ** lookahead, lookahead), dtype=np.uint8)
|
||||
cdef int i, j, xx, p
|
||||
cdef unsigned char output_pixel_nbit
|
||||
cdef unsigned char[::1] palette_choices_nbit
|
||||
# XXX palette bit depth
|
||||
return (last_pixel_nbit >> (lookahead+1)) | (next_pixels << (8 - (lookahead + 1)))
|
||||
|
||||
cdef object palette = screen.palette
|
||||
cdef dict palette_rgb = palette.RGB
|
||||
for i in range(2 ** lookahead):
|
||||
output_pixel_nbit = last_pixel_nbit
|
||||
for j in range(lookahead):
|
||||
xx = x + j
|
||||
# Two possible n-bit palette choices at position xx, given state of n pixels to left.
|
||||
# TODO: port screen.py to pyx
|
||||
palette_choices_nbit = screen.pixel_palette_options(output_pixel_nbit, xx)
|
||||
next_bit_choice = (i & (1 << j)) >> j
|
||||
options_nbit[i, j] = palette_choices_nbit[next_bit_choice]
|
||||
output_pixel_nbit >>= 1
|
||||
output_pixel_nbit |= (next_bit_choice << 7)
|
||||
# print(bin(i),j,bin(last_pixel_nbit), bin(output_pixel_nbit), bin(options_nbit[i, j]))
|
||||
#print("Picking %s" % ((i & (1 << j)) >> j))
|
||||
# cdef unsigned char[:, ::1] options_nbit = np.empty((2 ** lookahead, lookahead), dtype=np.uint8)
|
||||
# cdef int i, j, p, k
|
||||
# cdef unsigned char output_pixel_nbit
|
||||
# cdef unsigned char[::1] palette_choices_nbit
|
||||
|
||||
return options_nbit
|
||||
# # cdef object palette = screen.palette
|
||||
# # cdef dict palette_rgb = palette.RGB
|
||||
|
||||
# cdef unsigned int[::1] lookahead_pixel_values = (np.arange(2**lookahead, dtype=np.uint32) << 8) | last_pixel_nbit
|
||||
# # XXX inline into dither_lookahead once working
|
||||
|
||||
# for i in range(2 ** lookahead):
|
||||
# for j in range(lookahead):
|
||||
# options_nbit[:, j] = (lookahead_pixel_values[i] >> (j+1)) & 0xff
|
||||
|
||||
# return options_nbit
|
||||
|
||||
|
||||
# Look ahead a number of pixels and compute choice for next pixel with lowest total squared error after dithering.
|
||||
|
|
@ -110,8 +107,8 @@ def lookahead_options(object screen, int lookahead, unsigned char last_pixel_nbi
|
|||
#
|
||||
@cython.boundscheck(False)
|
||||
@cython.wraparound(False)
|
||||
cdef int dither_lookahead(Dither* dither, float[:, ::1] palette_rgb,
|
||||
float[:, :, ::1] image_rgb, int x, int y, unsigned char[:, ::1] options_nbit, int lookahead,
|
||||
cdef int dither_lookahead(Dither* dither, float[:, :, ::1] palette_rgb,
|
||||
float[:, :, ::1] image_rgb, int x, int y, int lookahead, unsigned char last_pixels,
|
||||
int x_res):
|
||||
cdef int i, j, k, l
|
||||
cdef float[3] quant_error
|
||||
|
|
@ -130,6 +127,15 @@ cdef int dither_lookahead(Dither* dither, float[:, ::1] palette_rgb,
|
|||
|
||||
cdef unsigned char lookahead_bits
|
||||
|
||||
# def unsigned int[::1] lookahead_pixel_values = (np.arange(2**lookahead, dtype=np.uint32) << 8) | last_pixel_nbit
|
||||
# XXX inline into dither_lookahead once working
|
||||
|
||||
#for i in range(2 ** lookahead):
|
||||
# for j in range(lookahead):
|
||||
# options_nbit[:, j] = (lookahead_pixel_values[i] >> (j+1)) & 0xff
|
||||
|
||||
cdef unsigned char next_pixels
|
||||
|
||||
# For each 2**lookahead possibilities for the on/off state of the next lookahead pixels, apply error diffusion
|
||||
# and compute the total squared error to the source image. Since we only have two possible colours for each
|
||||
# given pixel (dependent on the state already chosen for pixels to the left), we need to look beyond local minima.
|
||||
|
|
@ -145,38 +151,26 @@ cdef int dither_lookahead(Dither* dither, float[:, ::1] palette_rgb,
|
|||
for j in range(xxr - x):
|
||||
xl = dither_bounds_xl(dither, j)
|
||||
xr = dither_bounds_xr(dither, xxr - x, j)
|
||||
|
||||
next_pixels = lookahead_pixels(last_pixels, i, j)
|
||||
|
||||
# We don't update the input at position x (since we've already chosen
|
||||
# fixed outputs), but we do propagate quantization errors to positions >x
|
||||
# so we can compensate for how good/bad these choices were. i.e. the
|
||||
# options_rgb choices are fixed, but we can still distribute quantization error
|
||||
# from having made these choices, in order to compute the total error.
|
||||
for k in range(3):
|
||||
quant_error[k] = lah_image_rgb[j * lah_shape2 + k] - palette_rgb[options_nbit[i,j], k]
|
||||
quant_error[k] = lah_image_rgb[j * lah_shape2 + k] - palette_rgb[next_pixels, (x+j) % 4, k]
|
||||
apply_one_line(dither, xl, xr, j, lah_image_rgb, lah_shape2, quant_error)
|
||||
|
||||
#r = <long>lah_image_rgb[j * lah_shape2 + 0]
|
||||
#g = <long>lah_image_rgb[j * lah_shape2 + 1]
|
||||
#b = <long>lah_image_rgb[j * lah_shape2 + 2]
|
||||
#flat = (r << 16) + (g << 8) + b
|
||||
# bit4 = options_nbit[i, j]
|
||||
total_error += colour_distance_squared(lah_image_rgb[j*lah_shape2], lah_image_rgb[j*lah_shape2+1], lah_image_rgb[j*lah_shape2+2], palette_rgb[next_pixels, (x+j)%4])
|
||||
|
||||
## XXX parametrize number of palette bits
|
||||
#if j < 8:
|
||||
# lookahead_bits = (last_bits >> (j+1))
|
||||
#else:
|
||||
# lookahead_bits = 0
|
||||
#lookahead_bits |= (options_nbit[i, j]) << (7-j)
|
||||
#lookahead_bits &= (1<<8)-1
|
||||
|
||||
total_error += colour_distance_squared(lah_image_rgb[j*lah_shape2], lah_image_rgb[j*lah_shape2+1], lah_image_rgb[j*lah_shape2+2], palette_rgb[options_nbit[i,j]])
|
||||
|
||||
#if x > (560*3/4) and y == 180:
|
||||
# print(x, bin(i), j, bin(options_nbit[i, j]), bin(best), best_error, total_error)
|
||||
#if y == 0:
|
||||
# print(x, bin(i), j, bin(next_pixels), bin(best), best_error, total_error, list(palette_rgb[next_pixels, (x+j)%4]))
|
||||
|
||||
if total_error >= best_error:
|
||||
break
|
||||
|
||||
|
||||
if total_error < best_error:
|
||||
best_error = total_error
|
||||
best = i
|
||||
|
|
@ -184,7 +178,6 @@ cdef int dither_lookahead(Dither* dither, float[:, ::1] palette_rgb,
|
|||
free(lah_image_rgb)
|
||||
return best
|
||||
|
||||
|
||||
@cython.boundscheck(False)
|
||||
@cython.wraparound(False)
|
||||
cdef inline float colour_distance_squared(float colour1_0, float colour1_1, float colour1_2, float[::1] colour2):
|
||||
|
|
@ -295,7 +288,7 @@ cdef unsigned char find_nearest_colour(float[::1] pixel_rgb, unsigned char[::1]
|
|||
@cython.boundscheck(False)
|
||||
@cython.wraparound(False)
|
||||
def dither_image(screen, float[:, :, ::1] image_rgb, dither, int lookahead, unsigned char verbose):
|
||||
cdef int y, x, i
|
||||
cdef int y, x, i, k
|
||||
cdef float[3] input_pixel_rgb
|
||||
cdef float[3] quant_error
|
||||
cdef unsigned char [:, ::1] options_nbit
|
||||
|
|
@ -311,10 +304,10 @@ def dither_image(screen, float[:, :, ::1] image_rgb, dither, int lookahead, unsi
|
|||
cdef int xres = screen.X_RES
|
||||
|
||||
# XXX not rgb any more
|
||||
cdef float[:, ::1] palette_rgb = np.zeros((len(screen.palette.CAM02UCS), 3), dtype=np.float32)
|
||||
for i in screen.palette.CAM02UCS.keys():
|
||||
cdef float[:, :, ::1] palette_rgb = np.zeros((len(screen.palette.CAM02UCS), 4, 3), dtype=np.float32)
|
||||
for i, k in screen.palette.CAM02UCS.keys():
|
||||
for j in range(3):
|
||||
palette_rgb[i, j] = screen.palette.CAM02UCS[i][j]
|
||||
palette_rgb[i, k, j] = screen.palette.CAM02UCS[i, k][j]
|
||||
|
||||
# cdef (unsigned char)[:, ::1] distances = screen.palette.distances
|
||||
|
||||
|
|
@ -332,8 +325,6 @@ def dither_image(screen, float[:, :, ::1] image_rgb, dither, int lookahead, unsi
|
|||
cdef (unsigned char)[:, ::1] image_nbit = np.empty(
|
||||
(image_rgb.shape[0], image_rgb.shape[1]), dtype=np.uint8)
|
||||
|
||||
# print(lookahead_options(screen, lookahead, 0b11111111, 0))
|
||||
|
||||
for y in range(yres):
|
||||
if verbose:
|
||||
print("%d/%d" % (y, yres))
|
||||
|
|
@ -343,16 +334,13 @@ def dither_image(screen, float[:, :, ::1] image_rgb, dither, int lookahead, unsi
|
|||
input_pixel_rgb[i] = image_rgb[y,x,i]
|
||||
if lookahead:
|
||||
# Compute all possible 2**N choices of n-bit pixel colours for positions x .. x + lookahead
|
||||
lookahead_palette_choices_nbit = lookahead_options(screen, lookahead, output_pixel_nbit, x % 4)
|
||||
|
||||
# lookahead_palette_choices_nbit = lookahead_options(lookahead, output_pixel_nbit)
|
||||
# Apply error diffusion for each of these 2**N choices, and compute which produces the closest match
|
||||
# to the source image over the succeeding N pixels
|
||||
best_idx = dither_lookahead(
|
||||
&cdither, palette_rgb, image_rgb, x, y, lookahead_palette_choices_nbit, lookahead,
|
||||
xres)
|
||||
output_pixel_nbit >>= 1
|
||||
output_pixel_nbit |= (best_idx & 0b1) << 7 # XXX n bit shift
|
||||
# print("Picked %d" % (best_idx & 0b1))
|
||||
next_pixels = dither_lookahead(
|
||||
&cdither, palette_rgb, image_rgb, x, y, lookahead, output_pixel_nbit, xres)
|
||||
output_pixel_nbit = lookahead_pixels(output_pixel_nbit, next_pixels, 0)
|
||||
# print("Picked %s" % bin(output_pixel_nbit)) # (best_idx & 0b1))
|
||||
# lookahead_palette_choices_nbit[best_idx, 0]
|
||||
#else:
|
||||
# # Choose the closest colour among the available n-bit palette options
|
||||
|
|
@ -360,11 +348,10 @@ def dither_image(screen, float[:, :, ::1] image_rgb, dither, int lookahead, unsi
|
|||
# output_pixel_nbit = find_nearest_colour(input_pixel_rgb, palette_choices_nbit, distances)
|
||||
|
||||
# Apply error diffusion from chosen output pixel value
|
||||
# print("picked %d, %s" % (lookahead_palette_choices_nbit[best_idx, 0], bin(lookahead_palette_choices_nbit[best_idx, 0])))
|
||||
output_pixel_rgb = palette_rgb[lookahead_palette_choices_nbit[best_idx, 0]]
|
||||
output_pixel_rgb = palette_rgb[output_pixel_nbit, x % 4]
|
||||
for i in range(3):
|
||||
quant_error[i] = input_pixel_rgb[i] - output_pixel_rgb[i]
|
||||
image_nbit[y, x] = lookahead_palette_choices_nbit[best_idx, 0] # output_pixel_nbit
|
||||
image_nbit[y, x] = output_pixel_nbit
|
||||
apply(&cdither, xres, yres, x, y, image_rgb, quant_error)
|
||||
|
||||
for i in range(3):
|
||||
|
|
|
|||
|
|
@ -8,36 +8,41 @@ import screen
|
|||
|
||||
def main():
|
||||
s = screen.DHGR560NTSCScreen(palette=None)
|
||||
bitmap = np.zeros((1, 8), dtype=np.bool)
|
||||
|
||||
colours = {}
|
||||
unique = set()
|
||||
|
||||
print("import numpy as np")
|
||||
print()
|
||||
print("SRGB = {")
|
||||
# For each sequence of 8 pixels, compute the RGB colour of the right-most
|
||||
# pixel, using NTSC emulation.
|
||||
for bits in range(256):
|
||||
bits8 = np.empty((8,), dtype=np.bool)
|
||||
for i in range(8):
|
||||
bits8[i] = bits & (1 << i)
|
||||
# Double Hi-Res has a timing shift that rotates the displayed bits one
|
||||
# position with respect to NTSC phase.
|
||||
bitmap[0, :] = bits8[[3, 0, 1, 2, 7, 4, 5, 6]]
|
||||
# Double Hi-Res has a timing shift that rotates the displayed bits one
|
||||
# position with respect to NTSC phase.
|
||||
# TODO: should be 3? Do I have a compensating off-by-one in bitmap_to_ntsc?
|
||||
ntsc_shift = 2
|
||||
for j in range(ntsc_shift, ntsc_shift+4):
|
||||
bitmap = np.zeros((1, 11+ntsc_shift), dtype=np.bool)
|
||||
for bits in range(256):
|
||||
bits8 = np.empty((8,), dtype=np.bool)
|
||||
for i in range(8):
|
||||
bits8[i] = bits & (1 << i)
|
||||
|
||||
ntsc = s.bitmap_to_ntsc(bitmap)
|
||||
last_colour = ntsc[0, -1, :]
|
||||
colours[bits] = last_colour
|
||||
unique.add(tuple(last_colour))
|
||||
print("%d: np.array((%d, %d, %d))," % (
|
||||
bits, last_colour[0], last_colour[1], last_colour[2]))
|
||||
bitmap[0, j:j+8] = bits8
|
||||
|
||||
ntsc = s.bitmap_to_ntsc(bitmap)
|
||||
last_colour = ntsc[0, 3*(j+8)-1, :]
|
||||
colours[(bits, j-ntsc_shift)] = last_colour
|
||||
unique.add(tuple(last_colour))
|
||||
print(" (%d, %d): np.array((%d, %d, %d))," % (
|
||||
bits, j-ntsc_shift, last_colour[0], last_colour[1], last_colour[2]))
|
||||
print("}")
|
||||
print("# %d unique colours" % len(unique))
|
||||
|
||||
# Show spectrum of available colours sorted by HSV hue value
|
||||
im = np.zeros((128, 256 * 16, 3), dtype=np.uint8)
|
||||
for x, hsv in enumerate(sorted([tuple(colour.RGB_to_HSV(c / 256)) for c in
|
||||
colours.values()])):
|
||||
im[0:128, x * 16: (x + 1) * 16, :] = colour.HSV_to_RGB(hsv) * 256
|
||||
im = np.zeros((128*4, 256 * 16, 3), dtype=np.uint8)
|
||||
for x, j in colours:
|
||||
im[128*j:128*(j+1), x * 16: (x + 1) * 16, :] = colours[x,j]
|
||||
|
||||
Image.fromarray(im).show()
|
||||
|
||||
|
|
|
|||
420
palette.py
420
palette.py
|
|
@ -3,6 +3,7 @@
|
|||
import colour
|
||||
import numpy as np
|
||||
import image
|
||||
import palette_ntsc
|
||||
|
||||
|
||||
class Palette:
|
||||
|
|
@ -46,81 +47,81 @@ class Palette:
|
|||
self.DOTS_TO_INDEX[v] = k
|
||||
|
||||
|
||||
class ToHgrPalette(Palette):
|
||||
"""4-bit palette used as default by other DHGR image converters."""
|
||||
DISTANCES_PATH = "data/distances_tohgr.data"
|
||||
PALETTE_DEPTH = 4
|
||||
|
||||
# Default tohgr/bmp2dhr palette
|
||||
SRGB = {
|
||||
0: np.array((0, 0, 0)), # Black
|
||||
8: np.array((148, 12, 125)), # Magenta
|
||||
4: np.array((99, 77, 0)), # Brown
|
||||
12: np.array((249, 86, 29)), # Orange
|
||||
2: np.array((51, 111, 0)), # Dark green
|
||||
10: np.array((126, 126, 126)), # Grey2
|
||||
6: np.array((67, 200, 0)), # Green
|
||||
14: np.array((221, 206, 23)), # Yellow
|
||||
1: np.array((32, 54, 212)), # Dark blue
|
||||
9: np.array((188, 55, 255)), # Violet
|
||||
5: np.array((126, 126, 126)), # Grey1
|
||||
13: np.array((255, 129, 236)), # Pink
|
||||
3: np.array((7, 168, 225)), # Med blue
|
||||
11: np.array((158, 172, 255)), # Light blue
|
||||
7: np.array((93, 248, 133)), # Aqua
|
||||
15: np.array((255, 255, 255)), # White
|
||||
}
|
||||
|
||||
|
||||
class OpenEmulatorPalette(Palette):
|
||||
"""4-bit palette chosen to approximately match OpenEmulator output."""
|
||||
DISTANCES_PATH = "data/distances_openemulator.data"
|
||||
PALETTE_DEPTH = 4
|
||||
|
||||
# OpenEmulator
|
||||
SRGB = {
|
||||
0: np.array((0, 0, 0)), # Black
|
||||
8: np.array((203, 0, 121)), # Magenta
|
||||
4: np.array((99, 103, 0)), # Brown
|
||||
12: np.array((244, 78, 0)), # Orange
|
||||
2: np.array((0, 150, 0)), # Dark green
|
||||
10: np.array((130, 130, 130)), # Grey2
|
||||
6: np.array((0, 235, 0)), # Green
|
||||
14: np.array((214, 218, 0)), # Yellow
|
||||
1: np.array((20, 0, 246)), # Dark blue
|
||||
9: np.array((230, 0, 244)), # Violet
|
||||
5: np.array((130, 130, 130)), # Grey1
|
||||
13: np.array((244, 105, 235)), # Pink
|
||||
3: np.array((0, 174, 243)), # Med blue
|
||||
11: np.array((160, 156, 244)), # Light blue
|
||||
7: np.array((25, 243, 136)), # Aqua
|
||||
15: np.array((244, 247, 244)), # White
|
||||
}
|
||||
|
||||
|
||||
class VirtualIIPalette(Palette):
|
||||
"""4-bit palette exactly matching Virtual II emulator output."""
|
||||
DISTANCES_PATH = "data/distances_virtualii.data"
|
||||
PALETTE_DEPTH = 4
|
||||
|
||||
SRGB = {
|
||||
0: np.array((0, 0, 0)), # Black
|
||||
8: np.array((231, 36, 66)), # Magenta
|
||||
4: np.array((154, 104, 0)), # Brown
|
||||
12: np.array((255, 124, 0)), # Orange
|
||||
2: np.array((0, 135, 45)), # Dark green
|
||||
10: np.array((104, 104, 104)), # Grey2
|
||||
6: np.array((0, 222, 0)), # Green
|
||||
14: np.array((255, 252, 0)), # Yellow
|
||||
1: np.array((1, 30, 169)), # Dark blue
|
||||
9: np.array((230, 73, 228)), # Violet
|
||||
5: np.array((185, 185, 185)), # Grey1
|
||||
13: np.array((255, 171, 153)), # Pink
|
||||
3: np.array((47, 69, 255)), # Med blue
|
||||
11: np.array((120, 187, 255)), # Light blue
|
||||
7: np.array((83, 250, 208)), # Aqua
|
||||
15: np.array((255, 255, 255)), # White
|
||||
}
|
||||
# class ToHgrPalette(Palette):
|
||||
# """4-bit palette used as default by other DHGR image converters."""
|
||||
# DISTANCES_PATH = "data/distances_tohgr.data"
|
||||
# PALETTE_DEPTH = 4
|
||||
#
|
||||
# # Default tohgr/bmp2dhr palette
|
||||
# SRGB = {
|
||||
# 0: np.array((0, 0, 0)), # Black
|
||||
# 8: np.array((148, 12, 125)), # Magenta
|
||||
# 4: np.array((99, 77, 0)), # Brown
|
||||
# 12: np.array((249, 86, 29)), # Orange
|
||||
# 2: np.array((51, 111, 0)), # Dark green
|
||||
# 10: np.array((126, 126, 126)), # Grey2
|
||||
# 6: np.array((67, 200, 0)), # Green
|
||||
# 14: np.array((221, 206, 23)), # Yellow
|
||||
# 1: np.array((32, 54, 212)), # Dark blue
|
||||
# 9: np.array((188, 55, 255)), # Violet
|
||||
# 5: np.array((126, 126, 126)), # Grey1
|
||||
# 13: np.array((255, 129, 236)), # Pink
|
||||
# 3: np.array((7, 168, 225)), # Med blue
|
||||
# 11: np.array((158, 172, 255)), # Light blue
|
||||
# 7: np.array((93, 248, 133)), # Aqua
|
||||
# 15: np.array((255, 255, 255)), # White
|
||||
# }
|
||||
#
|
||||
#
|
||||
# class OpenEmulatorPalette(Palette):
|
||||
# """4-bit palette chosen to approximately match OpenEmulator output."""
|
||||
# DISTANCES_PATH = "data/distances_openemulator.data"
|
||||
# PALETTE_DEPTH = 4
|
||||
#
|
||||
# # OpenEmulator
|
||||
# SRGB = {
|
||||
# 0: np.array((0, 0, 0)), # Black
|
||||
# 8: np.array((203, 0, 121)), # Magenta
|
||||
# 4: np.array((99, 103, 0)), # Brown
|
||||
# 12: np.array((244, 78, 0)), # Orange
|
||||
# 2: np.array((0, 150, 0)), # Dark green
|
||||
# 10: np.array((130, 130, 130)), # Grey2
|
||||
# 6: np.array((0, 235, 0)), # Green
|
||||
# 14: np.array((214, 218, 0)), # Yellow
|
||||
# 1: np.array((20, 0, 246)), # Dark blue
|
||||
# 9: np.array((230, 0, 244)), # Violet
|
||||
# 5: np.array((130, 130, 130)), # Grey1
|
||||
# 13: np.array((244, 105, 235)), # Pink
|
||||
# 3: np.array((0, 174, 243)), # Med blue
|
||||
# 11: np.array((160, 156, 244)), # Light blue
|
||||
# 7: np.array((25, 243, 136)), # Aqua
|
||||
# 15: np.array((244, 247, 244)), # White
|
||||
# }
|
||||
#
|
||||
#
|
||||
# class VirtualIIPalette(Palette):
|
||||
# """4-bit palette exactly matching Virtual II emulator output."""
|
||||
# DISTANCES_PATH = "data/distances_virtualii.data"
|
||||
# PALETTE_DEPTH = 4
|
||||
#
|
||||
# SRGB = {
|
||||
# 0: np.array((0, 0, 0)), # Black
|
||||
# 8: np.array((231, 36, 66)), # Magenta
|
||||
# 4: np.array((154, 104, 0)), # Brown
|
||||
# 12: np.array((255, 124, 0)), # Orange
|
||||
# 2: np.array((0, 135, 45)), # Dark green
|
||||
# 10: np.array((104, 104, 104)), # Grey2
|
||||
# 6: np.array((0, 222, 0)), # Green
|
||||
# 14: np.array((255, 252, 0)), # Yellow
|
||||
# 1: np.array((1, 30, 169)), # Dark blue
|
||||
# 9: np.array((230, 73, 228)), # Violet
|
||||
# 5: np.array((185, 185, 185)), # Grey1
|
||||
# 13: np.array((255, 171, 153)), # Pink
|
||||
# 3: np.array((47, 69, 255)), # Med blue
|
||||
# 11: np.array((120, 187, 255)), # Light blue
|
||||
# 7: np.array((83, 250, 208)), # Aqua
|
||||
# 15: np.array((255, 255, 255)), # White
|
||||
# }
|
||||
|
||||
|
||||
class NTSCPalette(Palette):
|
||||
|
|
@ -129,272 +130,13 @@ class NTSCPalette(Palette):
|
|||
PALETTE_DEPTH = 8
|
||||
|
||||
# Computed using ntsc_colours.py
|
||||
SRGB = {
|
||||
0: np.array((0, 0, 0)),
|
||||
1: np.array((0, 0, 115)),
|
||||
2: np.array((0, 37, 0)),
|
||||
3: np.array((0, 23, 88)),
|
||||
4: np.array((14, 14, 0)),
|
||||
5: np.array((0, 0, 0)),
|
||||
6: np.array((0, 52, 0)),
|
||||
7: np.array((0, 37, 0)),
|
||||
8: np.array((64, 0, 26)),
|
||||
9: np.array((49, 0, 141)),
|
||||
10: np.array((0, 0, 0)),
|
||||
11: np.array((0, 0, 115)),
|
||||
12: np.array((79, 0, 0)),
|
||||
13: np.array((64, 0, 26)),
|
||||
14: np.array((14, 14, 0)),
|
||||
15: np.array((0, 0, 0)),
|
||||
16: np.array((49, 48, 178)),
|
||||
17: np.array((34, 34, 255)),
|
||||
18: np.array((0, 86, 152)),
|
||||
19: np.array((0, 71, 255)),
|
||||
20: np.array((63, 63, 63)),
|
||||
21: np.array((49, 48, 178)),
|
||||
22: np.array((0, 101, 37)),
|
||||
23: np.array((0, 86, 152)),
|
||||
24: np.array((113, 10, 204)),
|
||||
25: np.array((98, 0, 255)),
|
||||
26: np.array((49, 48, 178)),
|
||||
27: np.array((34, 34, 255)),
|
||||
28: np.array((128, 25, 89)),
|
||||
29: np.array((113, 10, 204)),
|
||||
30: np.array((63, 63, 63)),
|
||||
31: np.array((49, 48, 178)),
|
||||
32: np.array((0, 101, 37)),
|
||||
33: np.array((0, 86, 152)),
|
||||
34: np.array((0, 139, 11)),
|
||||
35: np.array((0, 124, 126)),
|
||||
36: np.array((13, 116, 0)),
|
||||
37: np.array((0, 101, 37)),
|
||||
38: np.array((0, 154, 0)),
|
||||
39: np.array((0, 139, 11)),
|
||||
40: np.array((63, 63, 63)),
|
||||
41: np.array((49, 48, 178)),
|
||||
42: np.array((0, 101, 37)),
|
||||
43: np.array((0, 86, 152)),
|
||||
44: np.array((78, 78, 0)),
|
||||
45: np.array((63, 63, 63)),
|
||||
46: np.array((13, 116, 0)),
|
||||
47: np.array((0, 101, 37)),
|
||||
48: np.array((48, 150, 216)),
|
||||
49: np.array((33, 135, 255)),
|
||||
50: np.array((0, 188, 190)),
|
||||
51: np.array((0, 173, 255)),
|
||||
52: np.array((62, 165, 101)),
|
||||
53: np.array((48, 150, 216)),
|
||||
54: np.array((0, 203, 75)),
|
||||
55: np.array((0, 188, 190)),
|
||||
56: np.array((112, 112, 242)),
|
||||
57: np.array((98, 97, 255)),
|
||||
58: np.array((48, 150, 216)),
|
||||
59: np.array((33, 135, 255)),
|
||||
60: np.array((127, 127, 127)),
|
||||
61: np.array((112, 112, 242)),
|
||||
62: np.array((62, 165, 101)),
|
||||
63: np.array((48, 150, 216)),
|
||||
64: np.array((78, 78, 0)),
|
||||
65: np.array((63, 63, 63)),
|
||||
66: np.array((13, 116, 0)),
|
||||
67: np.array((0, 101, 37)),
|
||||
68: np.array((93, 93, 0)),
|
||||
69: np.array((78, 78, 0)),
|
||||
70: np.array((28, 131, 0)),
|
||||
71: np.array((13, 116, 0)),
|
||||
72: np.array((142, 40, 0)),
|
||||
73: np.array((128, 25, 89)),
|
||||
74: np.array((78, 78, 0)),
|
||||
75: np.array((63, 63, 63)),
|
||||
76: np.array((157, 55, 0)),
|
||||
77: np.array((142, 40, 0)),
|
||||
78: np.array((93, 93, 0)),
|
||||
79: np.array((78, 78, 0)),
|
||||
80: np.array((127, 127, 127)),
|
||||
81: np.array((112, 112, 242)),
|
||||
82: np.array((62, 165, 101)),
|
||||
83: np.array((48, 150, 216)),
|
||||
84: np.array((142, 142, 12)),
|
||||
85: np.array((127, 127, 127)),
|
||||
86: np.array((77, 180, 0)),
|
||||
87: np.array((62, 165, 101)),
|
||||
88: np.array((192, 89, 153)),
|
||||
89: np.array((177, 74, 255)),
|
||||
90: np.array((127, 127, 127)),
|
||||
91: np.array((112, 112, 242)),
|
||||
92: np.array((206, 104, 38)),
|
||||
93: np.array((192, 89, 153)),
|
||||
94: np.array((142, 142, 12)),
|
||||
95: np.array((127, 127, 127)),
|
||||
96: np.array((77, 180, 0)),
|
||||
97: np.array((62, 165, 101)),
|
||||
98: np.array((13, 218, 0)),
|
||||
99: np.array((0, 203, 75)),
|
||||
100: np.array((92, 195, 0)),
|
||||
101: np.array((77, 180, 0)),
|
||||
102: np.array((27, 233, 0)),
|
||||
103: np.array((13, 218, 0)),
|
||||
104: np.array((142, 142, 12)),
|
||||
105: np.array((127, 127, 127)),
|
||||
106: np.array((77, 180, 0)),
|
||||
107: np.array((62, 165, 101)),
|
||||
108: np.array((156, 157, 0)),
|
||||
109: np.array((142, 142, 12)),
|
||||
110: np.array((92, 195, 0)),
|
||||
111: np.array((77, 180, 0)),
|
||||
112: np.array((126, 229, 165)),
|
||||
113: np.array((112, 214, 255)),
|
||||
114: np.array((62, 255, 138)),
|
||||
115: np.array((47, 252, 253)),
|
||||
116: np.array((141, 244, 50)),
|
||||
117: np.array((126, 229, 165)),
|
||||
118: np.array((76, 255, 23)),
|
||||
119: np.array((62, 255, 138)),
|
||||
120: np.array((191, 191, 191)),
|
||||
121: np.array((176, 176, 255)),
|
||||
122: np.array((126, 229, 165)),
|
||||
123: np.array((112, 214, 255)),
|
||||
124: np.array((205, 206, 76)),
|
||||
125: np.array((191, 191, 191)),
|
||||
126: np.array((141, 244, 50)),
|
||||
127: np.array((126, 229, 165)),
|
||||
128: np.array((128, 25, 89)),
|
||||
129: np.array((113, 10, 204)),
|
||||
130: np.array((63, 63, 63)),
|
||||
131: np.array((49, 48, 178)),
|
||||
132: np.array((142, 40, 0)),
|
||||
133: np.array((128, 25, 89)),
|
||||
134: np.array((78, 78, 0)),
|
||||
135: np.array((63, 63, 63)),
|
||||
136: np.array((192, 0, 116)),
|
||||
137: np.array((178, 0, 231)),
|
||||
138: np.array((128, 25, 89)),
|
||||
139: np.array((113, 10, 204)),
|
||||
140: np.array((207, 2, 1)),
|
||||
141: np.array((192, 0, 116)),
|
||||
142: np.array((142, 40, 0)),
|
||||
143: np.array((128, 25, 89)),
|
||||
144: np.array((177, 74, 255)),
|
||||
145: np.array((162, 59, 255)),
|
||||
146: np.array((112, 112, 242)),
|
||||
147: np.array((98, 97, 255)),
|
||||
148: np.array((192, 89, 153)),
|
||||
149: np.array((177, 74, 255)),
|
||||
150: np.array((127, 127, 127)),
|
||||
151: np.array((112, 112, 242)),
|
||||
152: np.array((241, 36, 255)),
|
||||
153: np.array((227, 21, 255)),
|
||||
154: np.array((177, 74, 255)),
|
||||
155: np.array((162, 59, 255)),
|
||||
156: np.array((255, 51, 179)),
|
||||
157: np.array((241, 36, 255)),
|
||||
158: np.array((192, 89, 153)),
|
||||
159: np.array((177, 74, 255)),
|
||||
160: np.array((127, 127, 127)),
|
||||
161: np.array((112, 112, 242)),
|
||||
162: np.array((62, 165, 101)),
|
||||
163: np.array((48, 150, 216)),
|
||||
164: np.array((142, 142, 12)),
|
||||
165: np.array((127, 127, 127)),
|
||||
166: np.array((77, 180, 0)),
|
||||
167: np.array((62, 165, 101)),
|
||||
168: np.array((192, 89, 153)),
|
||||
169: np.array((177, 74, 255)),
|
||||
170: np.array((127, 127, 127)),
|
||||
171: np.array((112, 112, 242)),
|
||||
172: np.array((206, 104, 38)),
|
||||
173: np.array((192, 89, 153)),
|
||||
174: np.array((142, 142, 12)),
|
||||
175: np.array((127, 127, 127)),
|
||||
176: np.array((176, 176, 255)),
|
||||
177: np.array((161, 161, 255)),
|
||||
178: np.array((112, 214, 255)),
|
||||
179: np.array((97, 199, 255)),
|
||||
180: np.array((191, 191, 191)),
|
||||
181: np.array((176, 176, 255)),
|
||||
182: np.array((126, 229, 165)),
|
||||
183: np.array((112, 214, 255)),
|
||||
184: np.array((241, 138, 255)),
|
||||
185: np.array((226, 123, 255)),
|
||||
186: np.array((176, 176, 255)),
|
||||
187: np.array((161, 161, 255)),
|
||||
188: np.array((255, 153, 217)),
|
||||
189: np.array((241, 138, 255)),
|
||||
190: np.array((191, 191, 191)),
|
||||
191: np.array((176, 176, 255)),
|
||||
192: np.array((206, 104, 38)),
|
||||
193: np.array((192, 89, 153)),
|
||||
194: np.array((142, 142, 12)),
|
||||
195: np.array((127, 127, 127)),
|
||||
196: np.array((221, 119, 0)),
|
||||
197: np.array((206, 104, 38)),
|
||||
198: np.array((156, 157, 0)),
|
||||
199: np.array((142, 142, 12)),
|
||||
200: np.array((255, 66, 64)),
|
||||
201: np.array((255, 51, 179)),
|
||||
202: np.array((206, 104, 38)),
|
||||
203: np.array((192, 89, 153)),
|
||||
204: np.array((255, 81, 0)),
|
||||
205: np.array((255, 66, 64)),
|
||||
206: np.array((221, 119, 0)),
|
||||
207: np.array((206, 104, 38)),
|
||||
208: np.array((255, 153, 217)),
|
||||
209: np.array((241, 138, 255)),
|
||||
210: np.array((191, 191, 191)),
|
||||
211: np.array((176, 176, 255)),
|
||||
212: np.array((255, 168, 102)),
|
||||
213: np.array((255, 153, 217)),
|
||||
214: np.array((205, 206, 76)),
|
||||
215: np.array((191, 191, 191)),
|
||||
216: np.array((255, 115, 243)),
|
||||
217: np.array((255, 100, 255)),
|
||||
218: np.array((255, 153, 217)),
|
||||
219: np.array((241, 138, 255)),
|
||||
220: np.array((255, 130, 128)),
|
||||
221: np.array((255, 115, 243)),
|
||||
222: np.array((255, 168, 102)),
|
||||
223: np.array((255, 153, 217)),
|
||||
224: np.array((205, 206, 76)),
|
||||
225: np.array((191, 191, 191)),
|
||||
226: np.array((141, 244, 50)),
|
||||
227: np.array((126, 229, 165)),
|
||||
228: np.array((220, 220, 0)),
|
||||
229: np.array((205, 206, 76)),
|
||||
230: np.array((156, 255, 0)),
|
||||
231: np.array((141, 244, 50)),
|
||||
232: np.array((255, 168, 102)),
|
||||
233: np.array((255, 153, 217)),
|
||||
234: np.array((205, 206, 76)),
|
||||
235: np.array((191, 191, 191)),
|
||||
236: np.array((255, 183, 0)),
|
||||
237: np.array((255, 168, 102)),
|
||||
238: np.array((220, 220, 0)),
|
||||
239: np.array((205, 206, 76)),
|
||||
240: np.array((254, 255, 255)),
|
||||
241: np.array((240, 240, 255)),
|
||||
242: np.array((190, 255, 228)),
|
||||
243: np.array((175, 255, 255)),
|
||||
244: np.array((255, 255, 139)),
|
||||
245: np.array((254, 255, 255)),
|
||||
246: np.array((205, 255, 113)),
|
||||
247: np.array((190, 255, 228)),
|
||||
248: np.array((255, 217, 255)),
|
||||
249: np.array((255, 202, 255)),
|
||||
250: np.array((254, 255, 255)),
|
||||
251: np.array((240, 240, 255)),
|
||||
252: np.array((255, 231, 166)),
|
||||
253: np.array((255, 217, 255)),
|
||||
254: np.array((255, 255, 139)),
|
||||
255: np.array((254, 255, 255)),
|
||||
}
|
||||
# 85 unique colours
|
||||
|
||||
SRGB = palette_ntsc.SRGB
|
||||
|
||||
PALETTES = {
|
||||
'openemulator': OpenEmulatorPalette,
|
||||
'virtualii': VirtualIIPalette,
|
||||
'tohgr': ToHgrPalette,
|
||||
# 'openemulator': OpenEmulatorPalette,
|
||||
# 'virtualii': VirtualIIPalette,
|
||||
# 'tohgr': ToHgrPalette,
|
||||
'ntsc': NTSCPalette
|
||||
}
|
||||
|
||||
DEFAULT_PALETTE = 'openemulator'
|
||||
DEFAULT_PALETTE = 'ntsc'
|
||||
|
|
|
|||
57
screen.py
57
screen.py
|
|
@ -151,7 +151,7 @@ class DHGR560Screen(Screen):
|
|||
# XXX wrong
|
||||
|
||||
assert next_dots[(3 - x) % 4 + 4] is None
|
||||
#print(x, last_dots, next_dots)
|
||||
# print(x, last_dots, next_dots)
|
||||
|
||||
next_dots[(3 - x) % 4 + 4] = False
|
||||
next_pixel_nbit_0 = self.palette.DOTS_TO_INDEX[next_dots]
|
||||
|
|
@ -173,9 +173,9 @@ class DHGR560NTSCScreen(Screen):
|
|||
for y in range(self.Y_RES):
|
||||
for x in range(self.X_RES):
|
||||
pixel = image_nbit[y, x]
|
||||
dots = self.palette.DOTS[pixel]
|
||||
phase = x % 4
|
||||
bitmap[y, x] = dots[4 + phase]
|
||||
#dots = self.palette.DOTS[pixel]
|
||||
#phase = x % 4
|
||||
bitmap[y, x] = pixel >> 7 # dots[4 + phase]
|
||||
return bitmap
|
||||
|
||||
def bitmap_to_image_rgb(self, bitmap: np.ndarray) -> np.ndarray:
|
||||
|
|
@ -190,13 +190,14 @@ class DHGR560NTSCScreen(Screen):
|
|||
for y in range(self.Y_RES):
|
||||
pixel = [False, False, False, False, False, False, False, False]
|
||||
for x in range(self.NATIVE_X_RES):
|
||||
pixel[x % 4] = pixel[x % 4 + 4]
|
||||
pixel[x % 4 + 4] = bitmap[y, x]
|
||||
# pixel[x % 4] = pixel[x % 4 + 4]
|
||||
# pixel[x % 4 + 4] = bitmap[y, x]
|
||||
pixel = pixel[1:] + [bitmap[y, x]]
|
||||
dots = self.palette.DOTS_TO_INDEX[tuple(pixel)]
|
||||
image_rgb[y, x, :] = self.palette.RGB[dots]
|
||||
image_rgb[y, x, :] = self.palette.RGB[dots, x % 4]
|
||||
return image_rgb
|
||||
|
||||
def pixel_palette_options(self, last_pixel_nbit, x: int):
|
||||
def pixel_palette_options(self, last_pixel_nbit):
|
||||
# # The two available 8-bit pixel colour choices are given by:
|
||||
# # - Rotating the pixel value from the current x % 4 + 4 position to
|
||||
# # x % 4
|
||||
|
|
@ -210,25 +211,33 @@ class DHGR560NTSCScreen(Screen):
|
|||
# pixel_nbit_0 = self.palette.DOTS_TO_INDEX[tuple(next_dots0)]
|
||||
# pixel_nbit_1 = self.palette.DOTS_TO_INDEX[tuple(next_dots1)]
|
||||
# return np.array([pixel_nbit_0, pixel_nbit_1], dtype=np.uint8)
|
||||
last_dots = list(self.palette.DOTS[last_pixel_nbit][1:]) + [None]
|
||||
#next_dots = last_dots[1:] # list(self.palette.DOTS[
|
||||
# last_pixel_nbit][1:])
|
||||
#return np.array([
|
||||
# self.palette.DOTS_TO_INDEX[tuple(last_dots + [False])],
|
||||
# self.palette.DOTS_TO_INDEX[tuple(next_dots + [True])]],
|
||||
# dtype=np.uint8)
|
||||
|
||||
# rearrange into palette order
|
||||
next_dots = [None] * 8
|
||||
for i in range(4):
|
||||
next_dots[i] = last_dots[(i - x) % 4]
|
||||
next_dots[i + 4] = last_dots[(i - x) % 4 + 4]
|
||||
|
||||
assert next_dots[(3 + x) % 4 + 4] is None
|
||||
# print(x, last_dots, next_dots)
|
||||
|
||||
next_dots[(3 + x) % 4 + 4] = False
|
||||
next_pixel_nbit_0 = self.palette.DOTS_TO_INDEX[tuple(next_dots)]
|
||||
|
||||
next_dots[(3 + x) % 4 + 4] = True
|
||||
next_pixel_nbit_1 = self.palette.DOTS_TO_INDEX[tuple(next_dots)]
|
||||
return np.array([next_pixel_nbit_0, next_pixel_nbit_1],
|
||||
return np.array(last_pixel_nbit >> 1, (last_pixel_nbit >> 1) + 1,
|
||||
dtype=np.uint8)
|
||||
|
||||
# # rearrange into palette order
|
||||
# next_dots = [None] * 8
|
||||
# for i in range(4):
|
||||
# next_dots[i] = last_dots[(i - x) % 4]
|
||||
# next_dots[i + 4] = last_dots[(i - x) % 4 + 4]
|
||||
#
|
||||
# assert next_dots[(3 + x) % 4 + 4] is None
|
||||
# # print(x, last_dots, next_dots)
|
||||
#
|
||||
# next_dots[(3 + x) % 4 + 4] = False
|
||||
# next_pixel_nbit_0 = self.palette.DOTS_TO_INDEX[tuple(next_dots)]
|
||||
#
|
||||
# next_dots[(3 + x) % 4 + 4] = True
|
||||
# next_pixel_nbit_1 = self.palette.DOTS_TO_INDEX[tuple(next_dots)]
|
||||
# return np.array([next_pixel_nbit_0, next_pixel_nbit_1],
|
||||
# dtype=np.uint8)
|
||||
|
||||
def bitmap_to_ntsc(self, bitmap: np.ndarray) -> np.ndarray:
|
||||
y_width = 12
|
||||
u_width = 24
|
||||
|
|
|
|||
Loading…
Reference in New Issue