circuitpython/shared-module/bitmaptools/__init__.c

860 lines
29 KiB
C

/*
* This file is part of the Micro Python project, http://micropython.org/
*
* The MIT License (MIT)
*
* Copyright (c) 2021 Kevin Matocha
*
* Permission is hereby granted, free of charge, to any person obtaining a copy
* of this software and associated documentation files (the "Software"), to deal
* in the Software without restriction, including without limitation the rights
* to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
* copies of the Software, and to permit persons to whom the Software is
* furnished to do so, subject to the following conditions:
*
* The above copyright notice and this permission notice shall be included in
* all copies or substantial portions of the Software.
*
* THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
* IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
* FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
* AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
* LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
* OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN
* THE SOFTWARE.
*/
#include "shared-bindings/bitmaptools/__init__.h"
#include "shared-bindings/displayio/Bitmap.h"
#include "shared-bindings/displayio/Palette.h"
#include "shared-bindings/displayio/ColorConverter.h"
#include "shared-module/displayio/Bitmap.h"
#include "py/mperrno.h"
#include "py/runtime.h"
#include "py/stream.h"
#include <math.h>
#include <stdlib.h>
#include <stdio.h>
#include <string.h>
void common_hal_bitmaptools_rotozoom(displayio_bitmap_t *self, int16_t ox, int16_t oy,
int16_t dest_clip0_x, int16_t dest_clip0_y,
int16_t dest_clip1_x, int16_t dest_clip1_y,
displayio_bitmap_t *source, int16_t px, int16_t py,
int16_t source_clip0_x, int16_t source_clip0_y,
int16_t source_clip1_x, int16_t source_clip1_y,
mp_float_t angle,
mp_float_t scale,
uint32_t skip_index, bool skip_index_none) {
// Copies region from source to the destination bitmap, including rotation,
// scaling and clipping of either the source or destination regions
//
// *self: destination bitmap
// ox: the (ox, oy) destination point where the source (px,py) point is placed
// oy:
// dest_clip0: (x,y) is the corner of the clip window on the destination bitmap
// dest_clip1: (x,y) is the other corner of the clip window of the destination bitmap
// *source: the source bitmap
// px: the (px, py) point of rotation of the source bitmap
// py:
// source_clip0: (x,y) is the corner of the clip window on the source bitmap
// source_clip1: (x,y) is the other of the clip window on the source bitmap
// angle: angle of rotation in radians, positive is clockwise
// scale: scale factor
// skip_index: color index that should be ignored (and not copied over)
// skip_index_none: if skip_index_none is True, then all color indexes should be copied
// (that is, no color indexes should be skipped)
// Copy complete "source" bitmap into "self" bitmap at location x,y in the "self"
// Add a boolean to determine if all values are copied, or only if non-zero
// If skip_value is encountered in the source bitmap, it will not be copied.
// If skip_value is `None`, then all pixels are copied.
// # Credit from https://github.com/wernsey/bitmap
// # MIT License from
// # * Copyright (c) 2017 Werner Stoop <wstoop@gmail.com>
// #
// # *
// # * #### `void bm_rotate_blit(Bitmap *dst, int ox, int oy, Bitmap *src, int px, int py, double angle, double scale);`
// # *
// # * Rotates a source bitmap `src` around a pivot point `px,py` and blits it onto a destination bitmap `dst`.
// # *
// # * The bitmap is positioned such that the point `px,py` on the source is at the offset `ox,oy` on the destination.
// # *
// # * The `angle` is clockwise, in radians. The bitmap is also scaled by the factor `scale`.
// #
// # void bm_rotate_blit(Bitmap *dst, int ox, int oy, Bitmap *src, int px, int py, double angle, double scale);
// # /*
// # Reference:
// # "Fast Bitmap Rotation and Scaling" By Steven Mortimer, Dr Dobbs' Journal, July 01, 2001
// # http://www.drdobbs.com/architecture-and-design/fast-bitmap-rotation-and-scaling/184416337
// # See also http://www.efg2.com/Lab/ImageProcessing/RotateScanline.htm
// # */
int16_t x,y;
int16_t minx = dest_clip1_x;
int16_t miny = dest_clip1_y;
int16_t maxx = dest_clip0_x;
int16_t maxy = dest_clip0_y;
mp_float_t sinAngle = MICROPY_FLOAT_C_FUN(sin)(angle);
mp_float_t cosAngle = MICROPY_FLOAT_C_FUN(cos)(angle);
mp_float_t dx, dy;
/* Compute the position of where each corner on the source bitmap
will be on the destination to get a bounding box for scanning */
dx = -cosAngle * px * scale + sinAngle * py * scale + ox;
dy = -sinAngle * px * scale - cosAngle * py * scale + oy;
if (dx < minx) {
minx = (int16_t)dx;
}
if (dx > maxx) {
maxx = (int16_t)dx;
}
if (dy < miny) {
miny = (int16_t)dy;
}
if (dy > maxy) {
maxy = (int16_t)dy;
}
dx = cosAngle * (source->width - px) * scale + sinAngle * py * scale + ox;
dy = sinAngle * (source->width - px) * scale - cosAngle * py * scale + oy;
if (dx < minx) {
minx = (int16_t)dx;
}
if (dx > maxx) {
maxx = (int16_t)dx;
}
if (dy < miny) {
miny = (int16_t)dy;
}
if (dy > maxy) {
maxy = (int16_t)dy;
}
dx = cosAngle * (source->width - px) * scale - sinAngle * (source->height - py) * scale + ox;
dy = sinAngle * (source->width - px) * scale + cosAngle * (source->height - py) * scale + oy;
if (dx < minx) {
minx = (int16_t)dx;
}
if (dx > maxx) {
maxx = (int16_t)dx;
}
if (dy < miny) {
miny = (int16_t)dy;
}
if (dy > maxy) {
maxy = (int16_t)dy;
}
dx = -cosAngle * px * scale - sinAngle * (source->height - py) * scale + ox;
dy = -sinAngle * px * scale + cosAngle * (source->height - py) * scale + oy;
if (dx < minx) {
minx = (int16_t)dx;
}
if (dx > maxx) {
maxx = (int16_t)dx;
}
if (dy < miny) {
miny = (int16_t)dy;
}
if (dy > maxy) {
maxy = (int16_t)dy;
}
/* Clipping */
if (minx < dest_clip0_x) {
minx = dest_clip0_x;
}
if (maxx > dest_clip1_x - 1) {
maxx = dest_clip1_x - 1;
}
if (miny < dest_clip0_y) {
miny = dest_clip0_y;
}
if (maxy > dest_clip1_y - 1) {
maxy = dest_clip1_y - 1;
}
mp_float_t dvCol = cosAngle / scale;
mp_float_t duCol = sinAngle / scale;
mp_float_t duRow = dvCol;
mp_float_t dvRow = -duCol;
mp_float_t startu = px - (ox * dvCol + oy * duCol);
mp_float_t startv = py - (ox * dvRow + oy * duRow);
mp_float_t rowu = startu + miny * duCol;
mp_float_t rowv = startv + miny * dvCol;
displayio_area_t dirty_area = {minx, miny, maxx + 1, maxy + 1, NULL};
displayio_bitmap_set_dirty_area(self, &dirty_area);
for (y = miny; y <= maxy; y++) {
mp_float_t u = rowu + minx * duRow;
mp_float_t v = rowv + minx * dvRow;
for (x = minx; x <= maxx; x++) {
if (u >= source_clip0_x && u < source_clip1_x && v >= source_clip0_y && v < source_clip1_y) {
uint32_t c = common_hal_displayio_bitmap_get_pixel(source, (int)u, (int)v);
if ((skip_index_none) || (c != skip_index)) {
displayio_bitmap_write_pixel(self, x, y, c);
}
}
u += duRow;
v += dvRow;
}
rowu += duCol;
rowv += dvCol;
}
}
void common_hal_bitmaptools_fill_region(displayio_bitmap_t *destination,
int16_t x1, int16_t y1,
int16_t x2, int16_t y2,
uint32_t value) {
// writes the value (a bitmap color index) into a bitmap in the specified rectangular region
//
// input checks should ensure that x1 < x2 and y1 < y2 and are within the bitmap region
displayio_area_t area = { x1, y1, x2, y2, NULL };
displayio_area_canon(&area);
displayio_area_t bitmap_area = { 0, 0, destination->width, destination->height, NULL };
displayio_area_compute_overlap(&area, &bitmap_area, &area);
// update the dirty rectangle
displayio_bitmap_set_dirty_area(destination, &area);
int16_t x, y;
for (x = area.x1; x < area.x2; x++) {
for (y = area.y1; y < area.y2; y++) {
displayio_bitmap_write_pixel(destination, x, y, value);
}
}
}
void common_hal_bitmaptools_boundary_fill(displayio_bitmap_t *destination,
int16_t x, int16_t y,
uint32_t fill_color_value, uint32_t replaced_color_value) {
if (fill_color_value == replaced_color_value) {
// There is nothing to do
return;
}
uint32_t current_point_color_value;
// the list of points that we'll check
mp_obj_t fill_area = mp_obj_new_list(0, NULL);
// first point is the one user passed in
mp_obj_t point[] = { mp_obj_new_int(x), mp_obj_new_int(y) };
mp_obj_list_append(
fill_area,
mp_obj_new_tuple(2, point)
);
int16_t minx = x;
int16_t miny = y;
int16_t maxx = x;
int16_t maxy = y;
if (replaced_color_value == INT_MAX) {
current_point_color_value = common_hal_displayio_bitmap_get_pixel(
destination,
mp_obj_get_int(point[0]),
mp_obj_get_int(point[1]));
replaced_color_value = (uint32_t)current_point_color_value;
}
mp_obj_t *fill_points;
size_t list_length = 0;
mp_obj_list_get(fill_area, &list_length, &fill_points);
mp_obj_t current_point;
size_t tuple_len = 0;
mp_obj_t *tuple_items;
int cur_x, cur_y;
// while there are still points to check
while (list_length > 0) {
mp_obj_list_get(fill_area, &list_length, &fill_points);
current_point = mp_obj_list_pop(fill_area, 0);
mp_obj_tuple_get(current_point, &tuple_len, &tuple_items);
current_point_color_value = common_hal_displayio_bitmap_get_pixel(
destination,
mp_obj_get_int(tuple_items[0]),
mp_obj_get_int(tuple_items[1]));
// if the current point is not background color ignore it
if (current_point_color_value != replaced_color_value) {
mp_obj_list_get(fill_area, &list_length, &fill_points);
continue;
}
cur_x = mp_obj_int_get_checked(tuple_items[0]);
cur_y = mp_obj_int_get_checked(tuple_items[1]);
if (cur_x < minx) {
minx = (int16_t)cur_x;
}
if (cur_x > maxx) {
maxx = (int16_t)cur_x;
}
if (cur_y < miny) {
miny = (int16_t)cur_y;
}
if (cur_y > maxy) {
maxy = (int16_t)cur_y;
}
// fill the current point with fill color
displayio_bitmap_write_pixel(
destination,
mp_obj_get_int(tuple_items[0]),
mp_obj_get_int(tuple_items[1]),
fill_color_value);
// add all 4 surrounding points to the list to check
// ignore points outside of the bitmap
if (mp_obj_int_get_checked(tuple_items[1]) - 1 >= 0) {
mp_obj_t above_point[] = {
tuple_items[0],
MP_OBJ_NEW_SMALL_INT(mp_obj_int_get_checked(tuple_items[1]) - 1)
};
mp_obj_list_append(
fill_area,
mp_obj_new_tuple(2, above_point));
}
// ignore points outside of the bitmap
if (mp_obj_int_get_checked(tuple_items[0]) - 1 >= 0) {
mp_obj_t left_point[] = {
MP_OBJ_NEW_SMALL_INT(mp_obj_int_get_checked(tuple_items[0]) - 1),
tuple_items[1]
};
mp_obj_list_append(
fill_area,
mp_obj_new_tuple(2, left_point));
}
// ignore points outside of the bitmap
if (mp_obj_int_get_checked(tuple_items[0]) + 1 < destination->width) {
mp_obj_t right_point[] = {
MP_OBJ_NEW_SMALL_INT(mp_obj_int_get_checked(tuple_items[0]) + 1),
tuple_items[1]
};
mp_obj_list_append(
fill_area,
mp_obj_new_tuple(2, right_point));
}
// ignore points outside of the bitmap
if (mp_obj_int_get_checked(tuple_items[1]) + 1 < destination->height) {
mp_obj_t below_point[] = {
tuple_items[0],
MP_OBJ_NEW_SMALL_INT(mp_obj_int_get_checked(tuple_items[1]) + 1)
};
mp_obj_list_append(
fill_area,
mp_obj_new_tuple(2, below_point));
}
mp_obj_list_get(fill_area, &list_length, &fill_points);
}
// set dirty the area so displayio will draw
displayio_area_t area = { minx, miny, maxx + 1, maxy + 1, NULL};
displayio_bitmap_set_dirty_area(destination, &area);
}
void common_hal_bitmaptools_draw_line(displayio_bitmap_t *destination,
int16_t x0, int16_t y0,
int16_t x1, int16_t y1,
uint32_t value) {
//
// adapted from Adafruit_CircuitPython_Display_Shapes.Polygon._line
//
// update the dirty rectangle
int16_t xbb0, xbb1, ybb0, ybb1;
if (x0 < x1) {
xbb0 = x0;
xbb1 = x1 + 1;
} else {
xbb0 = x1;
xbb1 = x0 + 1;
}
if (y0 < y1) {
ybb0 = y0;
ybb1 = y1 + 1;
} else {
ybb0 = y1;
ybb1 = y0 + 1;
}
displayio_area_t area = { xbb0, ybb0, xbb1, ybb1, NULL };
displayio_area_t bitmap_area = { 0, 0, destination->width, destination->height, NULL };
displayio_area_compute_overlap(&area, &bitmap_area, &area);
displayio_bitmap_set_dirty_area(destination, &area);
int16_t temp, x, y;
if (x0 == x1) { // vertical line
if (y0 > y1) { // ensure y1 > y0
temp = y0;
y0 = y1;
y1 = temp;
}
for (y = y0; y < (y1 + 1); y++) { // write a horizontal line
displayio_bitmap_write_pixel(destination, x0, y, value);
}
} else if (y0 == y1) { // horizontal line
if (x0 > x1) { // ensure y1 > y0
temp = x0;
x0 = x1;
x1 = temp;
}
for (x = x0; x < (x1 + 1); x++) { // write a horizontal line
displayio_bitmap_write_pixel(destination, x, y0, value);
}
} else {
bool steep;
steep = (abs(y1 - y0) > abs(x1 - x0));
if (steep) { // flip x0<->y0 and x1<->y1
temp = x0;
x0 = y0;
y0 = temp;
temp = x1;
x1 = y1;
y1 = temp;
}
if (x0 > x1) { // flip x0<->x1 and y0<->y1
temp = x0;
x0 = x1;
x1 = temp;
temp = y0;
y0 = y1;
y1 = temp;
}
int16_t dx, dy, ystep;
dx = x1 - x0;
dy = abs(y1 - y0);
mp_float_t err = dx / 2;
if (y0 < y1) {
ystep = 1;
} else {
ystep = -1;
}
for (x = x0; x < (x1 + 1); x++) {
if (steep) {
displayio_bitmap_write_pixel(destination, y0, x, value);
} else {
displayio_bitmap_write_pixel(destination, x, y0, value);
}
err -= dy;
if (err < 0) {
y0 += ystep;
err += dx;
}
}
}
}
void common_hal_bitmaptools_arrayblit(displayio_bitmap_t *self, void *data, int element_size, int x1, int y1, int x2, int y2, bool skip_specified, uint32_t skip_value) {
uint32_t mask = (1 << common_hal_displayio_bitmap_get_bits_per_value(self)) - 1;
for (int y = y1; y < y2; y++) {
for (int x = x1; x < x2; x++) {
uint32_t value;
switch (element_size) {
default:
case 1:
value = *(uint8_t *)data;
data = (void *)((uint8_t *)data + 1);
break;
case 2:
value = *(uint16_t *)data;
data = (void *)((uint16_t *)data + 1);
break;
case 4:
value = *(uint32_t *)data;
data = (void *)((uint32_t *)data + 1);
break;
}
if (!skip_specified || value != skip_value) {
displayio_bitmap_write_pixel(self, x, y, value & mask);
}
}
}
displayio_area_t area = { x1, y1, x2, y2, NULL };
displayio_bitmap_set_dirty_area(self, &area);
}
void common_hal_bitmaptools_readinto(displayio_bitmap_t *self, mp_obj_t *file, int element_size, int bits_per_pixel, bool reverse_pixels_in_element, bool swap_bytes, bool reverse_rows) {
uint32_t mask = (1 << common_hal_displayio_bitmap_get_bits_per_value(self)) - 1;
const mp_stream_p_t *file_proto = mp_get_stream_raise(file, MP_STREAM_OP_READ);
displayio_area_t a = {0, 0, self->width, self->height, NULL};
displayio_bitmap_set_dirty_area(self, &a);
size_t elements_per_row = (self->width * bits_per_pixel + element_size * 8 - 1) / (element_size * 8);
size_t rowsize = element_size * elements_per_row;
size_t rowsize_in_u32 = (rowsize + sizeof(uint32_t) - 1) / sizeof(uint32_t);
size_t rowsize_in_u16 = (rowsize + sizeof(uint16_t) - 1) / sizeof(uint16_t);
for (int y = 0; y < self->height; y++) {
uint32_t rowdata32[rowsize_in_u32];
uint16_t *rowdata16 = (uint16_t *)rowdata32;
uint8_t *rowdata8 = (uint8_t *)rowdata32;
const int y_draw = reverse_rows ? (self->height) - 1 - y : y;
int error = 0;
mp_uint_t bytes_read = file_proto->read(file, rowdata32, rowsize, &error);
if (error) {
mp_raise_OSError(error);
}
if (bytes_read != rowsize) {
mp_raise_msg(&mp_type_EOFError, NULL);
}
if (swap_bytes) {
switch (element_size) {
case 2:
for (size_t i = 0; i < rowsize_in_u16; i++) {
rowdata16[i] = __builtin_bswap16(rowdata16[i]);
}
break;
case 4:
for (size_t i = 0; i < rowsize_in_u32; i++) {
rowdata32[i] = __builtin_bswap32(rowdata32[i]);
}
default:
break;
}
}
for (int x = 0; x < self->width; x++) {
int value = 0;
switch (bits_per_pixel) {
case 1: {
int byte_offset = x / 8;
int bit_offset = reverse_pixels_in_element ? (7 - x % 8) : x % 8;
value = (rowdata8[byte_offset] >> bit_offset) & 1;
break;
}
case 2: {
int byte_offset = x / 4;
int bit_offset = 2 * (reverse_pixels_in_element ? (3 - x % 4) : x % 4);
value = (rowdata8[byte_offset] >> bit_offset) & 3;
break;
}
case 4: {
int byte_offset = x / 2;
int bit_offset = 4 * (reverse_pixels_in_element ? (1 - x % 2) : x % 2);
value = (rowdata8[byte_offset] >> bit_offset) & 0xf;
break;
}
case 8:
value = rowdata8[x];
break;
case 16:
value = rowdata16[x];
break;
case 24:
value = (rowdata8[x * 3] << 16) | (rowdata8[x * 3 + 1] << 8) | rowdata8[x * 3 + 2];
break;
case 32:
value = rowdata32[x];
break;
}
displayio_bitmap_write_pixel(self, x, y_draw, value & mask);
}
}
}
typedef struct {
uint8_t count; // The number of items in terms[]
uint8_t mx; // the maximum of the absolute value of the dx values
uint8_t dl; // the scaled dither value applied to the pixel at distance [1,0]
struct { // dl is the scaled dither values applied to the pixel at [dx,dy]
int8_t dx, dy, dl;
} terms[];
} bitmaptools_dither_algorithm_info_t;
static bitmaptools_dither_algorithm_info_t atkinson = {
4, 2, 256 / 8, {
{2, 0, 256 / 8},
{-1, 1, 256 / 8},
{0, 1, 256 / 8},
{0, 2, 256 / 8},
}
};
static bitmaptools_dither_algorithm_info_t floyd_stenberg = {
3, 1, 7 * 256 / 16,
{
{-1, 1, 3 * 256 / 16},
{0, 1, 5 * 256 / 16},
{1, 1, 1 * 256 / 16},
}
};
bitmaptools_dither_algorithm_info_t *algorithms[] = {
[DITHER_ALGORITHM_ATKINSON] = &atkinson,
[DITHER_ALGORITHM_FLOYD_STENBERG] = &floyd_stenberg,
};
enum {
SWAP_BYTES = 1 << 0,
SWAP_RB = 1 << 1,
};
STATIC void fill_row(displayio_bitmap_t *bitmap, int swap, int16_t *luminance_data, int y, int mx) {
if (y >= bitmap->height) {
return;
}
// zero out padding area
for (int i = 0; i < mx; i++) {
luminance_data[-mx + i] = 0;
luminance_data[bitmap->width + i] = 0;
}
if (bitmap->bits_per_value == 8) {
uint8_t *pixel_data = (uint8_t *)(bitmap->data + bitmap->stride * y);
for (int x = 0; x < bitmap->width; x++) {
*luminance_data++ = *pixel_data++;
}
} else {
uint16_t *pixel_data = (uint16_t *)(bitmap->data + bitmap->stride * y);
for (int x = 0; x < bitmap->width; x++) {
uint16_t pixel = *pixel_data++;
if (swap & SWAP_BYTES) {
pixel = __builtin_bswap16(pixel);
}
int r = (pixel >> 8) & 0xf8;
int g = (pixel >> 3) & 0xfc;
int b = (pixel << 3) & 0xf8;
if (swap & SWAP_RB) {
uint8_t tmp = r;
r = b;
b = tmp;
}
// ideal coefficients are around .299, .587, .114 (according to
// ppmtopnm), this differs from the 'other' luma-converting
// function in circuitpython (why?)
// we correct for the fact that the input ranges are 0..0xf8 (or
// 0xfc) rather than 0x00..0xff
// Check: (0xf8 * 78 + 0xfc * 154 + 0xf8 * 29) // 256 == 255
*luminance_data++ = (r * 78 + g * 154 + b * 29) / 256;
}
}
}
static void write_pixels(displayio_bitmap_t *bitmap, int y, bool *data) {
if (bitmap->bits_per_value == 1) {
uint32_t *pixel_data = (uint32_t *)(bitmap->data + bitmap->stride * y);
for (int i = 0; i < bitmap->stride; i++) {
uint32_t p = 0;
for (int j = 0; j < 32; i++) {
p = (p << 1);
if (*data++) {
p |= 1;
}
}
*pixel_data++ = p;
}
} else {
uint16_t *pixel_data = (uint16_t *)(bitmap->data + bitmap->stride * y);
for (int i = 0; i < bitmap->width; i++) {
*pixel_data++ = *data++ ? 65535 : 0;
}
}
}
void common_hal_bitmaptools_dither(displayio_bitmap_t *dest_bitmap, displayio_bitmap_t *source_bitmap, displayio_colorspace_t colorspace, bitmaptools_dither_algorithm_t algorithm) {
int height = dest_bitmap->height, width = dest_bitmap->width;
int swap = 0;
if (colorspace == DISPLAYIO_COLORSPACE_RGB565_SWAPPED || colorspace == DISPLAYIO_COLORSPACE_BGR565_SWAPPED) {
swap |= SWAP_BYTES;
}
if (colorspace == DISPLAYIO_COLORSPACE_BGR565 || colorspace == DISPLAYIO_COLORSPACE_BGR565_SWAPPED) {
swap |= SWAP_RB;
}
bitmaptools_dither_algorithm_info_t *info = algorithms[algorithm];
// rowdata holds 3 rows of data. Each one is larger than the input
// bitmap's width, beacuse `mx` extra pixels are allocated at the start and
// end of the row so that no conditionals are needed when storing the error data.
int16_t rowdata[(width + 2 * info->mx) * 3];
int16_t *rows[3] = {
rowdata + info->mx, rowdata + width + info->mx * 3, rowdata + 2 * width + info->mx * 5
};
// out holds one output row of pixels, and is padded to be a multiple of 32 so that the 1bpp storage loop can be simplified
bool out[(width + 31) / 32 * 32];
fill_row(source_bitmap, swap, rows[0], 0, info->mx);
fill_row(source_bitmap, swap, rows[1], 1, info->mx);
fill_row(source_bitmap, swap, rows[2], 2, info->mx);
int16_t err = 0;
for (int y = 0; y < height; y++) {
// Serpentine dither. Going left-to-right...
for (int x = 0; x < width; x++) {
int32_t pixel_in = rows[0][x] + err;
bool pixel_out = pixel_in >= 128;
out[x] = pixel_out;
err = pixel_in - (pixel_out ? 255 : 0);
for (int i = 0; i < info->count; i++) {
int x1 = x + info->terms[i].dx;
int dy = info->terms[i].dy;
rows[dy][x1] = ((info->terms[i].dl * err) / 256) + rows[dy][x1];
}
err = (err * info->dl) / 256;
}
write_pixels(dest_bitmap, y, out);
// Cycle the rows by shuffling pointers, this is faster than copying the data.
int16_t *tmp = rows[0];
rows[0] = rows[1];
rows[1] = rows[2];
rows[2] = tmp;
y++;
if (y == height) {
break;
}
fill_row(source_bitmap, swap, rows[2], y + 2, info->mx);
// Serpentine dither. Going right-to-left...
for (int x = width; x--;) {
int16_t pixel_in = rows[0][x] + err;
bool pixel_out = pixel_in >= 128;
out[x] = pixel_out;
err = pixel_in - (pixel_out ? 255 : 0);
for (int i = 0; i < info->count; i++) {
int x1 = x - info->terms[i].dx;
int dy = info->terms[i].dy;
rows[dy][x1] = ((info->terms[i].dl * err) / 256) + rows[dy][x1];
}
err = (err * info->dl) / 256;
}
write_pixels(dest_bitmap, y, out);
tmp = rows[0];
rows[0] = rows[1];
rows[1] = rows[2];
rows[2] = tmp;
fill_row(source_bitmap, swap, rows[2], y + 3, info->mx);
}
displayio_area_t a = { 0, 0, width, height, NULL };
displayio_bitmap_set_dirty_area(dest_bitmap, &a);
}
void common_hal_bitmaptools_alphablend(displayio_bitmap_t *dest, displayio_bitmap_t *source1, displayio_bitmap_t *source2, displayio_colorspace_t colorspace, mp_float_t factor1, mp_float_t factor2) {
displayio_area_t a = {0, 0, dest->width, dest->height, NULL};
displayio_bitmap_set_dirty_area(dest, &a);
int ifactor1 = (int)(factor1 * 256);
int ifactor2 = (int)(factor2 * 256);
if (colorspace == DISPLAYIO_COLORSPACE_L8) {
for (int y = 0; y < dest->height; y++) {
uint8_t *dptr = (uint8_t *)(dest->data + y * dest->stride);
uint8_t *sptr1 = (uint8_t *)(source1->data + y * source1->stride);
uint8_t *sptr2 = (uint8_t *)(source2->data + y * source2->stride);
for (int x = 0; x < dest->width; x++) {
// This is round(l1*f1 + l2*f2) & clip to range in fixed-point
int pixel = (*sptr1++ *ifactor1 + *sptr2++ *ifactor2 + 128) / 256;
*dptr++ = MIN(255, MAX(0, pixel));
}
}
} else {
bool swap = (colorspace == DISPLAYIO_COLORSPACE_RGB565_SWAPPED) || (colorspace == DISPLAYIO_COLORSPACE_BGR565_SWAPPED);
for (int y = 0; y < dest->height; y++) {
uint16_t *dptr = (uint16_t *)(dest->data + y * dest->stride);
uint16_t *sptr1 = (uint16_t *)(source1->data + y * source1->stride);
uint16_t *sptr2 = (uint16_t *)(source2->data + y * source2->stride);
for (int x = 0; x < dest->width; x++) {
int spix1 = *sptr1++;
int spix2 = *sptr2++;
if (swap) {
spix1 = __builtin_bswap16(spix1);
spix2 = __builtin_bswap16(spix2);
}
const int r_mask = 0xf800; // (or b mask, if BGR)
const int g_mask = 0x07e0;
const int b_mask = 0x001f; // (or r mask, if BGR)
// This is round(r1*f1 + r2*f2) & clip to range in fixed-point
// but avoiding shifting it down to start at bit 0
int r = ((spix1 & r_mask) * ifactor1
+ (spix2 & r_mask) * ifactor2 + r_mask / 2) / 256;
r = MIN(r_mask, MAX(0, r)) & r_mask;
// ditto
int g = ((spix1 & g_mask) * ifactor1
+ (spix2 & g_mask) * ifactor2 + g_mask / 2) / 256;
g = MIN(g_mask, MAX(0, g)) & g_mask;
int b = ((spix1 & b_mask) * ifactor1
+ (spix2 & b_mask) * ifactor2 + b_mask / 2) / 256;
b = MIN(b_mask, MAX(0, b)) & b_mask;
uint16_t pixel = r | g | b;
if (swap) {
pixel = __builtin_bswap16(pixel);
}
*dptr++ = pixel;
}
}
}
}