Tidy up and optimize a bit

This commit is contained in:
kris 2021-07-19 13:21:32 +01:00
parent e08f25e4cc
commit 4984df7f7a
1 changed files with 32 additions and 69 deletions

View File

@ -53,13 +53,11 @@ cdef int dither_bounds_yb(Dither *dither, int y_res, int y) nogil:
return yb
@cython.boundscheck(False)
@cython.wraparound(False)
# @functools.lru_cache(None)
cdef inline unsigned char lookahead_pixels(unsigned char last_pixel_nbit, unsigned int next_pixels, int lookahead):
cdef inline unsigned char lookahead_pixels(unsigned char last_pixel_nbit, unsigned int next_pixels, int lookahead) nogil:
"""Compute all possible n-bit palette values for upcoming pixels, given x coord and state of n pixels to the left.
Args:
XXX
screen: python screen.Screen object
lookahead: how many pixels to lookahead
last_pixel_nbit: n-bit value representing n pixels to left of current position, which determine available
@ -72,23 +70,6 @@ cdef inline unsigned char lookahead_pixels(unsigned char last_pixel_nbit, unsign
# XXX palette bit depth
return (last_pixel_nbit >> (lookahead+1)) | (next_pixels << (8 - (lookahead + 1)))
# 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
# # 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.
#
@ -109,15 +90,14 @@ cdef inline unsigned char lookahead_pixels(unsigned char last_pixel_nbit, unsign
@cython.wraparound(False)
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
int x_res) nogil:
cdef int i, j, k
cdef float[3] quant_error
cdef unsigned char bit4
cdef int best
cdef float best_error = 2**31-1
cdef float total_error
cdef long flat, dist
cdef long r, g, b
cdef unsigned char next_pixels
cdef int phase
# Don't bother dithering past the lookahead horizon or edge of screen.
cdef int xxr = min(x + lookahead, x_res)
@ -125,17 +105,6 @@ cdef int dither_lookahead(Dither* dither, float[:, :, ::1] palette_rgb,
cdef int lah_shape2 = 3
cdef float *lah_image_rgb = <float *> malloc(lah_shape1 * lah_shape2 * sizeof(float))
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.
@ -151,8 +120,9 @@ 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)
phase = (x + j) % 4
next_pixels = lookahead_pixels(last_pixels, i, j)
next_pixels = lookahead_pixels(last_pixels, next_pixels=i, lookahead=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
@ -160,13 +130,12 @@ cdef int dither_lookahead(Dither* dither, float[:, :, ::1] palette_rgb,
# 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[next_pixels, (x+j) % 4, k]
quant_error[k] = lah_image_rgb[j * lah_shape2 + k] - palette_rgb[next_pixels, phase, k]
apply_one_line(dither, xl, xr, j, lah_image_rgb, lah_shape2, quant_error)
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])
#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]))
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, phase])
if total_error >= best_error:
break
@ -178,13 +147,12 @@ cdef int dither_lookahead(Dither* dither, float[:, :, ::1] palette_rgb,
free(lah_image_rgb)
return best
# XXX fix signature
@cython.boundscheck(False)
@cython.wraparound(False)
cdef inline float colour_distance_squared(float colour1_0, float colour1_1, float colour1_2, float[::1] colour2):
# print("color1=%f,%f,%f color2=%f,%f,%f" % (colour1_0, colour1_1, colour1_2, colour2[0], colour2[1], colour2[2]))
cdef inline float colour_distance_squared(float colour1_0, float colour1_1, float colour1_2, float[::1] colour2) nogil:
return (colour1_0 - colour2[0])**2 + (colour1_1 - colour2[1])**2 + (colour1_2 - colour2[2])**2
# print(" --> %f" % dist)
# return dist
# Perform error diffusion to a single image row.
@ -223,7 +191,7 @@ cdef void apply_one_line(Dither* dither, int xl, int xr, int x, float[] image, i
#
@cython.boundscheck(False)
@cython.wraparound(False)
cdef void apply(Dither* dither, int x_res, int y_res, int x, int y, float[:,:,::1] image, float[] quant_error):
cdef void apply(Dither* dither, int x_res, int y_res, int x, int y, float[:,:,::1] image, float[] quant_error) nogil:
cdef int i, j, k
@ -288,15 +256,12 @@ 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, k
cdef float[3] input_pixel_rgb
cdef int y, x, i, j, k
# cdef float[3] input_pixel_rgb
cdef float[3] quant_error
cdef unsigned char [:, ::1] options_nbit
cdef float[:, :, ::1] options_rgb
cdef unsigned char [:, ::1] lookahead_palette_choices_nbit
cdef unsigned char [::1] palette_choices_nbit
cdef unsigned char output_pixel_nbit
cdef float[::1] output_pixel_rgb
cdef unsigned char best_next_pixels
cdef float[3] output_pixel_rgb
# Hoist some python attribute accesses into C variables for efficient access during the main loop
@ -305,11 +270,9 @@ def dither_image(screen, float[:, :, ::1] image_rgb, dither, int lookahead, unsi
# XXX not rgb any more
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, k, j] = screen.palette.CAM02UCS[i, k][j]
# cdef (unsigned char)[:, ::1] distances = screen.palette.distances
for i, j in screen.palette.CAM02UCS.keys():
for k in range(3):
palette_rgb[i, j, k] = screen.palette.CAM02UCS[i, j][k]
cdef Dither cdither
cdither.y_shape = dither.PATTERN.shape[0]
@ -330,30 +293,30 @@ def dither_image(screen, float[:, :, ::1] image_rgb, dither, int lookahead, unsi
print("%d/%d" % (y, yres))
output_pixel_nbit = 0
for x in range(xres):
for i in range(3):
input_pixel_rgb[i] = image_rgb[y,x,i]
#for i in range(3):
# 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(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
next_pixels = dither_lookahead(
best_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]
# Apply best choice for next 1 pixel
output_pixel_nbit = lookahead_pixels(output_pixel_nbit, best_next_pixels, lookahead=0)
#else:
# # Choose the closest colour among the available n-bit palette options
# palette_choices_nbit = screen.pixel_palette_options(output_pixel_nbit, x)
# output_pixel_nbit = find_nearest_colour(input_pixel_rgb, palette_choices_nbit, distances)
# Apply error diffusion from chosen output pixel value
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] = output_pixel_nbit
output_pixel_rgb[i] = palette_rgb[output_pixel_nbit, x % 4, i]
quant_error[i] = image_rgb[y,x,i] - output_pixel_rgb[i]
apply(&cdither, xres, yres, x, y, image_rgb, quant_error)
# Update image with our chosen image pixel
image_nbit[y, x] = output_pixel_nbit
for i in range(3):
image_rgb[y, x, i] = output_pixel_rgb[i]