mozilla
/
aom
зеркало из
1
0
Форкнуть 0
Код Задачи Пакеты Проекты Релизы Вики Активность

Further optimizations of loop restoration 

Change-Id: I4c4300f3f565d8aecf65669b77aaa874bb73a3a0
This commit is contained in:
Debargha Mukherjee 2016-12-16 03:13:02 -08:00
Родитель f10cba2b39
Коммит 519dbcf19b
4 изменённых файлов: 124 добавлений и 110 удалений
Показать статистику Скачать Patch файл Скачать Diff файл Показать все файлы Свернуть все файлы

177
av1/common/restoration.c
Просмотреть файл

@ -27,7 +27,7 @@ static int domaintxfmrf_vtable[DOMAINTXFMRF_ITERS][DOMAINTXFMRF_PARAMS][256];
static const int override_y_only[RESTORE_TYPES] = { 1, 1, 1, 1, 1 }; static const int override_y_only[RESTORE_TYPES] = { 1, 1, 1, 1, 1 };
static const int domaintxfmrf_params[DOMAINTXFMRF_PARAMS] = { static const int domaintxfmrf_params[DOMAINTXFMRF_PARAMS] = {
48, 52, 56, 60, 64, 68, 72, 76, 80, 82, 84, 86, 88, 32, 40, 48, 56, 64, 68, 72, 76, 80, 82, 84, 86, 88,
90, 92, 94, 96, 97, 98, 99, 100, 101, 102, 103, 104, 105, 90, 92, 94, 96, 97, 98, 99, 100, 101, 102, 103, 104, 105,
106, 107, 108, 109, 110, 111, 112, 113, 114, 115, 116, 117, 118, 106, 107, 108, 109, 110, 111, 112, 113, 114, 115, 116, 117, 118,
119, 120, 121, 122, 123, 124, 125, 126, 127, 128, 130, 132, 134, 119, 120, 121, 122, 123, 124, 125, 126, 127, 128, 130, 132, 134,
@ -252,8 +252,8 @@ static void loop_wiener_filter(uint8_t *data, int width, int height, int stride,
} }
} }
static void boxsum(int64_t *src, int width, int height, int src_stride, int r, static void boxsum(int32_t *src, int width, int height, int src_stride, int r,
int sqr, int64_t *dst, int dst_stride, int64_t *tmp, int sqr, int32_t *dst, int dst_stride, int32_t *tmp,
int tmp_stride) { int tmp_stride) {
int i, j; int i, j;
@ -342,11 +342,11 @@ void decode_xq(int *xqd, int *xq) {
} }
#define APPROXIMATE_SGR 1 #define APPROXIMATE_SGR 1
void av1_selfguided_restoration(int64_t *dgd, int width, int height, int stride, void av1_selfguided_restoration(int32_t *dgd, int width, int height, int stride,
int bit_depth, int r, int eps, void *tmpbuf) { int bit_depth, int r, int eps, void *tmpbuf) {
int64_t *A = (int64_t *)tmpbuf; int32_t *A = (int32_t *)tmpbuf;
int64_t *B = A + RESTORATION_TILEPELS_MAX; int32_t *B = A + RESTORATION_TILEPELS_MAX;
int64_t *T = B + RESTORATION_TILEPELS_MAX; int32_t *T = B + RESTORATION_TILEPELS_MAX;
int8_t num[RESTORATION_TILEPELS_MAX]; int8_t num[RESTORATION_TILEPELS_MAX];
int i, j; int i, j;
eps <<= 2 * (bit_depth - 8); eps <<= 2 * (bit_depth - 8);
@ -358,10 +358,9 @@ void av1_selfguided_restoration(int64_t *dgd, int width, int height, int stride,
for (j = 0; j < width; ++j) { for (j = 0; j < width; ++j) {
const int k = i * width + j; const int k = i * width + j;
const int n = num[k]; const int n = num[k];
int64_t den; const int64_t p = A[k] * n - B[k] * B[k];
A[k] = A[k] * n - B[k] * B[k]; const int64_t q = p + n * n * eps;
den = A[k] + n * n * eps; A[k] = (int32_t)((p << SGRPROJ_SGR_BITS) + (q >> 1)) / q;
A[k] = ((A[k] << SGRPROJ_SGR_BITS) + (den >> 1)) / den;
B[k] = ((SGRPROJ_SGR - A[k]) * B[k] + (n >> 1)) / n; B[k] = ((SGRPROJ_SGR - A[k]) * B[k] + (n >> 1)) / n;
} }
} }
@ -372,11 +371,11 @@ void av1_selfguided_restoration(int64_t *dgd, int width, int height, int stride,
const int k = i * width + j; const int k = i * width + j;
const int l = i * stride + j; const int l = i * stride + j;
const int nb = 3; const int nb = 3;
const int64_t a = const int32_t a =
3 * A[k] + 2 * A[k + 1] + 2 * A[k + width] + A[k + width + 1]; 3 * A[k] + 2 * A[k + 1] + 2 * A[k + width] + A[k + width + 1];
const int64_t b = const int32_t b =
3 * B[k] + 2 * B[k + 1] + 2 * B[k + width] + B[k + width + 1]; 3 * B[k] + 2 * B[k + 1] + 2 * B[k + width] + B[k + width + 1];
const int64_t v = const int32_t v =
(((a * dgd[l] + b) << SGRPROJ_RST_BITS) + (1 << nb) / 2) >> nb; (((a * dgd[l] + b) << SGRPROJ_RST_BITS) + (1 << nb) / 2) >> nb;
dgd[l] = ROUND_POWER_OF_TWO(v, SGRPROJ_SGR_BITS); dgd[l] = ROUND_POWER_OF_TWO(v, SGRPROJ_SGR_BITS);
} }
@ -386,11 +385,11 @@ void av1_selfguided_restoration(int64_t *dgd, int width, int height, int stride,
const int k = i * width + j; const int k = i * width + j;
const int l = i * stride + j; const int l = i * stride + j;
const int nb = 3; const int nb = 3;
const int64_t a = const int32_t a =
3 * A[k] + 2 * A[k - 1] + 2 * A[k + width] + A[k + width - 1]; 3 * A[k] + 2 * A[k - 1] + 2 * A[k + width] + A[k + width - 1];
const int64_t b = const int32_t b =
3 * B[k] + 2 * B[k - 1] + 2 * B[k + width] + B[k + width - 1]; 3 * B[k] + 2 * B[k - 1] + 2 * B[k + width] + B[k + width - 1];
const int64_t v = const int32_t v =
(((a * dgd[l] + b) << SGRPROJ_RST_BITS) + (1 << nb) / 2) >> nb; (((a * dgd[l] + b) << SGRPROJ_RST_BITS) + (1 << nb) / 2) >> nb;
dgd[l] = ROUND_POWER_OF_TWO(v, SGRPROJ_SGR_BITS); dgd[l] = ROUND_POWER_OF_TWO(v, SGRPROJ_SGR_BITS);
} }
@ -400,11 +399,11 @@ void av1_selfguided_restoration(int64_t *dgd, int width, int height, int stride,
const int k = i * width + j; const int k = i * width + j;
const int l = i * stride + j; const int l = i * stride + j;
const int nb = 3; const int nb = 3;
const int64_t a = const int32_t a =
3 * A[k] + 2 * A[k + 1] + 2 * A[k - width] + A[k - width + 1]; 3 * A[k] + 2 * A[k + 1] + 2 * A[k - width] + A[k - width + 1];
const int64_t b = const int32_t b =
3 * B[k] + 2 * B[k + 1] + 2 * B[k - width] + B[k - width + 1]; 3 * B[k] + 2 * B[k + 1] + 2 * B[k - width] + B[k - width + 1];
const int64_t v = const int32_t v =
(((a * dgd[l] + b) << SGRPROJ_RST_BITS) + (1 << nb) / 2) >> nb; (((a * dgd[l] + b) << SGRPROJ_RST_BITS) + (1 << nb) / 2) >> nb;
dgd[l] = ROUND_POWER_OF_TWO(v, SGRPROJ_SGR_BITS); dgd[l] = ROUND_POWER_OF_TWO(v, SGRPROJ_SGR_BITS);
} }
@ -414,11 +413,11 @@ void av1_selfguided_restoration(int64_t *dgd, int width, int height, int stride,
const int k = i * width + j; const int k = i * width + j;
const int l = i * stride + j; const int l = i * stride + j;
const int nb = 3; const int nb = 3;
const int64_t a = const int32_t a =
3 * A[k] + 2 * A[k - 1] + 2 * A[k - width] + A[k - width - 1]; 3 * A[k] + 2 * A[k - 1] + 2 * A[k - width] + A[k - width - 1];
const int64_t b = const int32_t b =
3 * B[k] + 2 * B[k - 1] + 2 * B[k - width] + B[k - width - 1]; 3 * B[k] + 2 * B[k - 1] + 2 * B[k - width] + B[k - width - 1];
const int64_t v = const int32_t v =
(((a * dgd[l] + b) << SGRPROJ_RST_BITS) + (1 << nb) / 2) >> nb; (((a * dgd[l] + b) << SGRPROJ_RST_BITS) + (1 << nb) / 2) >> nb;
dgd[l] = ROUND_POWER_OF_TWO(v, SGRPROJ_SGR_BITS); dgd[l] = ROUND_POWER_OF_TWO(v, SGRPROJ_SGR_BITS);
} }
@ -427,11 +426,11 @@ void av1_selfguided_restoration(int64_t *dgd, int width, int height, int stride,
const int k = i * width + j; const int k = i * width + j;
const int l = i * stride + j; const int l = i * stride + j;
const int nb = 3; const int nb = 3;
const int64_t a = A[k] + 2 * (A[k - 1] + A[k + 1]) + A[k + width] + const int32_t a = A[k] + 2 * (A[k - 1] + A[k + 1]) + A[k + width] +
A[k + width - 1] + A[k + width + 1]; A[k + width - 1] + A[k + width + 1];
const int64_t b = B[k] + 2 * (B[k - 1] + B[k + 1]) + B[k + width] + const int32_t b = B[k] + 2 * (B[k - 1] + B[k + 1]) + B[k + width] +
B[k + width - 1] + B[k + width + 1]; B[k + width - 1] + B[k + width + 1];
const int64_t v = const int32_t v =
(((a * dgd[l] + b) << SGRPROJ_RST_BITS) + (1 << nb) / 2) >> nb; (((a * dgd[l] + b) << SGRPROJ_RST_BITS) + (1 << nb) / 2) >> nb;
dgd[l] = ROUND_POWER_OF_TWO(v, SGRPROJ_SGR_BITS); dgd[l] = ROUND_POWER_OF_TWO(v, SGRPROJ_SGR_BITS);
} }
@ -440,11 +439,11 @@ void av1_selfguided_restoration(int64_t *dgd, int width, int height, int stride,
const int k = i * width + j; const int k = i * width + j;
const int l = i * stride + j; const int l = i * stride + j;
const int nb = 3; const int nb = 3;
const int64_t a = A[k] + 2 * (A[k - 1] + A[k + 1]) + A[k - width] + const int32_t a = A[k] + 2 * (A[k - 1] + A[k + 1]) + A[k - width] +
A[k - width - 1] + A[k - width + 1]; A[k - width - 1] + A[k - width + 1];
const int64_t b = B[k] + 2 * (B[k - 1] + B[k + 1]) + B[k - width] + const int32_t b = B[k] + 2 * (B[k - 1] + B[k + 1]) + B[k - width] +
B[k - width - 1] + B[k - width + 1]; B[k - width - 1] + B[k - width + 1];
const int64_t v = const int32_t v =
(((a * dgd[l] + b) << SGRPROJ_RST_BITS) + (1 << nb) / 2) >> nb; (((a * dgd[l] + b) << SGRPROJ_RST_BITS) + (1 << nb) / 2) >> nb;
dgd[l] = ROUND_POWER_OF_TWO(v, SGRPROJ_SGR_BITS); dgd[l] = ROUND_POWER_OF_TWO(v, SGRPROJ_SGR_BITS);
} }
@ -453,11 +452,11 @@ void av1_selfguided_restoration(int64_t *dgd, int width, int height, int stride,
const int k = i * width + j; const int k = i * width + j;
const int l = i * stride + j; const int l = i * stride + j;
const int nb = 3; const int nb = 3;
const int64_t a = A[k] + 2 * (A[k - width] + A[k + width]) + A[k + 1] + const int32_t a = A[k] + 2 * (A[k - width] + A[k + width]) + A[k + 1] +
A[k - width + 1] + A[k + width + 1]; A[k - width + 1] + A[k + width + 1];
const int64_t b = B[k] + 2 * (B[k - width] + B[k + width]) + B[k + 1] + const int32_t b = B[k] + 2 * (B[k - width] + B[k + width]) + B[k + 1] +
B[k - width + 1] + B[k + width + 1]; B[k - width + 1] + B[k + width + 1];
const int64_t v = const int32_t v =
(((a * dgd[l] + b) << SGRPROJ_RST_BITS) + (1 << nb) / 2) >> nb; (((a * dgd[l] + b) << SGRPROJ_RST_BITS) + (1 << nb) / 2) >> nb;
dgd[l] = ROUND_POWER_OF_TWO(v, SGRPROJ_SGR_BITS); dgd[l] = ROUND_POWER_OF_TWO(v, SGRPROJ_SGR_BITS);
} }
@ -466,11 +465,11 @@ void av1_selfguided_restoration(int64_t *dgd, int width, int height, int stride,
const int k = i * width + j; const int k = i * width + j;
const int l = i * stride + j; const int l = i * stride + j;
const int nb = 3; const int nb = 3;
const int64_t a = A[k] + 2 * (A[k - width] + A[k + width]) + A[k - 1] + const int32_t a = A[k] + 2 * (A[k - width] + A[k + width]) + A[k - 1] +
A[k - width - 1] + A[k + width - 1]; A[k - width - 1] + A[k + width - 1];
const int64_t b = B[k] + 2 * (B[k - width] + B[k + width]) + B[k - 1] + const int32_t b = B[k] + 2 * (B[k - width] + B[k + width]) + B[k - 1] +
B[k - width - 1] + B[k + width - 1]; B[k - width - 1] + B[k + width - 1];
const int64_t v = const int32_t v =
(((a * dgd[l] + b) << SGRPROJ_RST_BITS) + (1 << nb) / 2) >> nb; (((a * dgd[l] + b) << SGRPROJ_RST_BITS) + (1 << nb) / 2) >> nb;
dgd[l] = ROUND_POWER_OF_TWO(v, SGRPROJ_SGR_BITS); dgd[l] = ROUND_POWER_OF_TWO(v, SGRPROJ_SGR_BITS);
} }
@ -479,17 +478,17 @@ void av1_selfguided_restoration(int64_t *dgd, int width, int height, int stride,
const int k = i * width + j; const int k = i * width + j;
const int l = i * stride + j; const int l = i * stride + j;
const int nb = 5; const int nb = 5;
const int64_t a = const int32_t a =
(A[k] + A[k - 1] + A[k + 1] + A[k - width] + A[k + width]) * 4 + (A[k] + A[k - 1] + A[k + 1] + A[k - width] + A[k + width]) * 4 +
(A[k - 1 - width] + A[k - 1 + width] + A[k + 1 - width] + (A[k - 1 - width] + A[k - 1 + width] + A[k + 1 - width] +
A[k + 1 + width]) * A[k + 1 + width]) *
3; 3;
const int64_t b = const int32_t b =
(B[k] + B[k - 1] + B[k + 1] + B[k - width] + B[k + width]) * 4 + (B[k] + B[k - 1] + B[k + 1] + B[k - width] + B[k + width]) * 4 +
(B[k - 1 - width] + B[k - 1 + width] + B[k + 1 - width] + (B[k - 1 - width] + B[k - 1 + width] + B[k + 1 - width] +
B[k + 1 + width]) * B[k + 1 + width]) *
3; 3;
const int64_t v = const int32_t v =
(((a * dgd[l] + b) << SGRPROJ_RST_BITS) + (1 << nb) / 2) >> nb; (((a * dgd[l] + b) << SGRPROJ_RST_BITS) + (1 << nb) / 2) >> nb;
dgd[l] = ROUND_POWER_OF_TWO(v, SGRPROJ_SGR_BITS); dgd[l] = ROUND_POWER_OF_TWO(v, SGRPROJ_SGR_BITS);
} }
@ -503,7 +502,7 @@ void av1_selfguided_restoration(int64_t *dgd, int width, int height, int stride,
const int k = i * width + j; const int k = i * width + j;
const int l = i * stride + j; const int l = i * stride + j;
const int n = num[k]; const int n = num[k];
const int64_t v = const int32_t v =
(((A[k] * dgd[l] + B[k]) << SGRPROJ_RST_BITS) + (n >> 1)) / n; (((A[k] * dgd[l] + B[k]) << SGRPROJ_RST_BITS) + (n >> 1)) / n;
dgd[l] = ROUND_POWER_OF_TWO(v, SGRPROJ_SGR_BITS); dgd[l] = ROUND_POWER_OF_TWO(v, SGRPROJ_SGR_BITS);
} }
@ -511,12 +510,13 @@ void av1_selfguided_restoration(int64_t *dgd, int width, int height, int stride,
#endif // APPROXIMATE_SGR #endif // APPROXIMATE_SGR
} }
static void apply_selfguided_restoration(int64_t *dat, int width, int height, static void apply_selfguided_restoration(uint8_t *dat, int width, int height,
int stride, int bit_depth, int eps, int stride, int bit_depth, int eps,
int *xqd, void *tmpbuf) { int *xqd, uint8_t *dst, int dst_stride,
void *tmpbuf) {
int xq[2]; int xq[2];
int64_t *flt1 = (int64_t *)tmpbuf; int32_t *flt1 = (int32_t *)tmpbuf;
int64_t *flt2 = flt1 + RESTORATION_TILEPELS_MAX; int32_t *flt2 = flt1 + RESTORATION_TILEPELS_MAX;
uint8_t *tmpbuf2 = (uint8_t *)(flt2 + RESTORATION_TILEPELS_MAX); uint8_t *tmpbuf2 = (uint8_t *)(flt2 + RESTORATION_TILEPELS_MAX);
int i, j; int i, j;
for (i = 0; i < height; ++i) { for (i = 0; i < height; ++i) {
@ -535,13 +535,14 @@ static void apply_selfguided_restoration(int64_t *dat, int width, int height,
for (j = 0; j < width; ++j) { for (j = 0; j < width; ++j) {
const int k = i * width + j; const int k = i * width + j;
const int l = i * stride + j; const int l = i * stride + j;
const int64_t u = ((int64_t)dat[l] << SGRPROJ_RST_BITS); const int m = i * dst_stride + j;
const int64_t f1 = (int64_t)flt1[k] - u; const int32_t u = ((int32_t)dat[l] << SGRPROJ_RST_BITS);
const int64_t f2 = (int64_t)flt2[k] - u; const int32_t f1 = (int32_t)flt1[k] - u;
const int32_t f2 = (int32_t)flt2[k] - u;
const int64_t v = xq[0] * f1 + xq[1] * f2 + (u << SGRPROJ_PRJ_BITS); const int64_t v = xq[0] * f1 + xq[1] * f2 + (u << SGRPROJ_PRJ_BITS);
const int16_t w = const int16_t w =
(int16_t)ROUND_POWER_OF_TWO(v, SGRPROJ_PRJ_BITS + SGRPROJ_RST_BITS); (int16_t)ROUND_POWER_OF_TWO(v, SGRPROJ_PRJ_BITS + SGRPROJ_RST_BITS);
dat[l] = w; dst[m] = clip_pixel(w);
} }
} }
} }
@ -552,10 +553,9 @@ static void loop_sgrproj_filter_tile(uint8_t *data, int tile_idx, int width,
int dst_stride) { int dst_stride) {
const int tile_width = rst->tile_width >> rst->subsampling_x; const int tile_width = rst->tile_width >> rst->subsampling_x;
const int tile_height = rst->tile_height >> rst->subsampling_y; const int tile_height = rst->tile_height >> rst->subsampling_y;
int i, j;
int h_start, h_end, v_start, v_end; int h_start, h_end, v_start, v_end;
uint8_t *data_p, *dst_p; uint8_t *data_p, *dst_p;
int64_t *dat = (int64_t *)rst->tmpbuf; uint8_t *dat = (uint8_t *)rst->tmpbuf;
uint8_t *tmpbuf = uint8_t *tmpbuf =
(uint8_t *)rst->tmpbuf + RESTORATION_TILEPELS_MAX * sizeof(*dat); (uint8_t *)rst->tmpbuf + RESTORATION_TILEPELS_MAX * sizeof(*dat);
@ -568,22 +568,11 @@ static void loop_sgrproj_filter_tile(uint8_t *data, int tile_idx, int width,
tile_width, tile_height, width, height, 0, 0, tile_width, tile_height, width, height, 0, 0,
&h_start, &h_end, &v_start, &v_end); &h_start, &h_end, &v_start, &v_end);
data_p = data + h_start + v_start * stride; data_p = data + h_start + v_start * stride;
for (i = 0; i < (v_end - v_start); ++i) {
for (j = 0; j < (h_end - h_start); ++j) {
dat[i * (h_end - h_start) + j] = data_p[i * stride + j];
}
}
apply_selfguided_restoration(dat, h_end - h_start, v_end - v_start,
h_end - h_start, 8,
rst->rsi->sgrproj_info[tile_idx].ep,
rst->rsi->sgrproj_info[tile_idx].xqd, tmpbuf);
dst_p = dst + h_start + v_start * dst_stride; dst_p = dst + h_start + v_start * dst_stride;
for (i = 0; i < (v_end - v_start); ++i) { apply_selfguided_restoration(data_p, h_end - h_start, v_end - v_start, stride,
for (j = 0; j < (h_end - h_start); ++j) { 8, rst->rsi->sgrproj_info[tile_idx].ep,
dst_p[i * dst_stride + j] = rst->rsi->sgrproj_info[tile_idx].xqd, dst_p,
clip_pixel((int)dat[i * (h_end - h_start) + j]); dst_stride, tmpbuf);
}
}
} }
static void loop_sgrproj_filter(uint8_t *data, int width, int height, static void loop_sgrproj_filter(uint8_t *data, int width, int height,
@ -857,6 +846,44 @@ static void loop_wiener_filter_highbd(uint8_t *data8, int width, int height,
} }
} }
static void apply_selfguided_restoration_highbd(uint16_t *dat, int width,
int height, int stride,
int bit_depth, int eps,
int *xqd, uint16_t *dst,
int dst_stride, void *tmpbuf) {
int xq[2];
int32_t *flt1 = (int32_t *)tmpbuf;
int32_t *flt2 = flt1 + RESTORATION_TILEPELS_MAX;
uint8_t *tmpbuf2 = (uint8_t *)(flt2 + RESTORATION_TILEPELS_MAX);
int i, j;
for (i = 0; i < height; ++i) {
for (j = 0; j < width; ++j) {
assert(i * width + j < RESTORATION_TILEPELS_MAX);
flt1[i * width + j] = dat[i * stride + j];
flt2[i * width + j] = dat[i * stride + j];
}
}
av1_selfguided_restoration(flt1, width, height, width, bit_depth,
sgr_params[eps].r1, sgr_params[eps].e1, tmpbuf2);
av1_selfguided_restoration(flt2, width, height, width, bit_depth,
sgr_params[eps].r2, sgr_params[eps].e2, tmpbuf2);
decode_xq(xqd, xq);
for (i = 0; i < height; ++i) {
for (j = 0; j < width; ++j) {
const int k = i * width + j;
const int l = i * stride + j;
const int m = i * dst_stride + j;
const int32_t u = ((int32_t)dat[l] << SGRPROJ_RST_BITS);
const int32_t f1 = (int32_t)flt1[k] - u;
const int32_t f2 = (int32_t)flt2[k] - u;
const int64_t v = xq[0] * f1 + xq[1] * f2 + (u << SGRPROJ_PRJ_BITS);
const int16_t w =
(int16_t)ROUND_POWER_OF_TWO(v, SGRPROJ_PRJ_BITS + SGRPROJ_RST_BITS);
dst[m] = (uint16_t)clip_pixel_highbd(w, bit_depth);
}
}
}
static void loop_sgrproj_filter_tile_highbd(uint16_t *data, int tile_idx, static void loop_sgrproj_filter_tile_highbd(uint16_t *data, int tile_idx,
int width, int height, int stride, int width, int height, int stride,
RestorationInternal *rst, RestorationInternal *rst,
@ -864,10 +891,9 @@ static void loop_sgrproj_filter_tile_highbd(uint16_t *data, int tile_idx,
int dst_stride) { int dst_stride) {
const int tile_width = rst->tile_width >> rst->subsampling_x; const int tile_width = rst->tile_width >> rst->subsampling_x;
const int tile_height = rst->tile_height >> rst->subsampling_y; const int tile_height = rst->tile_height >> rst->subsampling_y;
int i, j;
int h_start, h_end, v_start, v_end; int h_start, h_end, v_start, v_end;
uint16_t *data_p, *dst_p; uint16_t *data_p, *dst_p;
int64_t *dat = (int64_t *)rst->tmpbuf; uint16_t *dat = (uint16_t *)rst->tmpbuf;
uint8_t *tmpbuf = uint8_t *tmpbuf =
(uint8_t *)rst->tmpbuf + RESTORATION_TILEPELS_MAX * sizeof(*dat); (uint8_t *)rst->tmpbuf + RESTORATION_TILEPELS_MAX * sizeof(*dat);
@ -880,22 +906,11 @@ static void loop_sgrproj_filter_tile_highbd(uint16_t *data, int tile_idx,
tile_width, tile_height, width, height, 0, 0, tile_width, tile_height, width, height, 0, 0,
&h_start, &h_end, &v_start, &v_end); &h_start, &h_end, &v_start, &v_end);
data_p = data + h_start + v_start * stride; data_p = data + h_start + v_start * stride;
for (i = 0; i < (v_end - v_start); ++i) {
for (j = 0; j < (h_end - h_start); ++j) {
dat[i * (h_end - h_start) + j] = data_p[i * stride + j];
}
}
apply_selfguided_restoration(dat, h_end - h_start, v_end - v_start,
h_end - h_start, bit_depth,
rst->rsi->sgrproj_info[tile_idx].ep,
rst->rsi->sgrproj_info[tile_idx].xqd, tmpbuf);
dst_p = dst + h_start + v_start * dst_stride; dst_p = dst + h_start + v_start * dst_stride;
for (i = 0; i < (v_end - v_start); ++i) { apply_selfguided_restoration_highbd(
for (j = 0; j < (h_end - h_start); ++j) { data_p, h_end - h_start, v_end - v_start, stride, bit_depth,
dst_p[i * dst_stride + j] = rst->rsi->sgrproj_info[tile_idx].ep, rst->rsi->sgrproj_info[tile_idx].xqd,
clip_pixel_highbd((int)dat[i * (h_end - h_start) + j], bit_depth); dst_p, dst_stride, tmpbuf);
}
}
} }
static void loop_sgrproj_filter_highbd(uint8_t *data8, int width, int height, static void loop_sgrproj_filter_highbd(uint8_t *data8, int width, int height,

7
av1/common/restoration.h
Просмотреть файл

@ -42,11 +42,10 @@ extern "C" {
#define DOMAINTXFMRF_TMPBUF_SIZE (RESTORATION_TILEPELS_MAX * sizeof(int32_t)) #define DOMAINTXFMRF_TMPBUF_SIZE (RESTORATION_TILEPELS_MAX * sizeof(int32_t))
#define DOMAINTXFMRF_BITS (DOMAINTXFMRF_PARAMS_BITS) #define DOMAINTXFMRF_BITS (DOMAINTXFMRF_PARAMS_BITS)
// 6 highprecision 64-bit buffers needed for the filter: // 6 highprecision buffers needed for the filter:
// 1 for the degraded frame, 2 for the restored versions and // 1 for the degraded frame, 2 for the restored versions and
// 3 for each restoration operation // 3 for each restoration operation
// TODO(debargha): Explore if we can use 32-bit buffers #define SGRPROJ_TMPBUF_SIZE (RESTORATION_TILEPELS_MAX * 6 * sizeof(int32_t))
#define SGRPROJ_TMPBUF_SIZE (RESTORATION_TILEPELS_MAX * 6 * sizeof(int64_t))
#define SGRPROJ_PARAMS_BITS 3 #define SGRPROJ_PARAMS_BITS 3
#define SGRPROJ_PARAMS (1 << SGRPROJ_PARAMS_BITS) #define SGRPROJ_PARAMS (1 << SGRPROJ_PARAMS_BITS)
@ -211,7 +210,7 @@ int av1_alloc_restoration_struct(RestorationInfo *rst_info, int width,
int height); int height);
void av1_free_restoration_struct(RestorationInfo *rst_info); void av1_free_restoration_struct(RestorationInfo *rst_info);
void av1_selfguided_restoration(int64_t *dgd, int width, int height, int stride, void av1_selfguided_restoration(int32_t *dgd, int width, int height, int stride,
int bit_depth, int r, int eps, void *tmpbuf); int bit_depth, int r, int eps, void *tmpbuf);
void av1_domaintxfmrf_restoration(uint8_t *dgd, int width, int height, void av1_domaintxfmrf_restoration(uint8_t *dgd, int width, int height,
int stride, int param, uint8_t *dst, int stride, int param, uint8_t *dst,

2
av1/encoder/encoder.c
Просмотреть файл

@ -739,7 +739,7 @@ static void alloc_util_frame_buffers(AV1_COMP *cpi) {
aom_internal_error(&cm->error, AOM_CODEC_MEM_ERROR, aom_internal_error(&cm->error, AOM_CODEC_MEM_ERROR,
"Failed to allocate trial restored frame buffer"); "Failed to allocate trial restored frame buffer");
cpi->extra_rstbuf = (uint8_t *)aom_realloc( cpi->extra_rstbuf = (uint8_t *)aom_realloc(
cpi->extra_rstbuf, RESTORATION_TILEPELS_MAX * sizeof(int64_t)); cpi->extra_rstbuf, RESTORATION_TILEPELS_MAX * sizeof(int32_t));
if (!cpi->extra_rstbuf) if (!cpi->extra_rstbuf)
aom_internal_error(&cm->error, AOM_CODEC_MEM_ERROR, aom_internal_error(&cm->error, AOM_CODEC_MEM_ERROR,
"Failed to allocate extra rstbuf for restoration"); "Failed to allocate extra rstbuf for restoration");

48
av1/encoder/pickrst.c
Просмотреть файл

@ -121,10 +121,10 @@ static int64_t try_restoration_frame(const YV12_BUFFER_CONFIG *src,
return filt_err; return filt_err;
} }
static int64_t get_pixel_proj_error(int64_t *src, int width, int height, static int64_t get_pixel_proj_error(int32_t *src, int width, int height,
int src_stride, int64_t *dgd, int src_stride, int32_t *dgd,
int dgd_stride, int64_t *flt1, int dgd_stride, int32_t *flt1,
int flt1_stride, int64_t *flt2, int flt1_stride, int32_t *flt2,
int flt2_stride, int *xqd) { int flt2_stride, int *xqd) {
int i, j; int i, j;
int64_t err = 0; int64_t err = 0;
@ -132,12 +132,12 @@ static int64_t get_pixel_proj_error(int64_t *src, int width, int height,
decode_xq(xqd, xq); decode_xq(xqd, xq);
for (i = 0; i < height; ++i) { for (i = 0; i < height; ++i) {
for (j = 0; j < width; ++j) { for (j = 0; j < width; ++j) {
const int64_t s = (int64_t)src[i * src_stride + j]; const int32_t s = (int32_t)src[i * src_stride + j];
const int64_t u = (int64_t)dgd[i * dgd_stride + j]; const int32_t u = (int32_t)dgd[i * dgd_stride + j];
const int64_t f1 = (int64_t)flt1[i * flt1_stride + j] - u; const int32_t f1 = (int32_t)flt1[i * flt1_stride + j] - u;
const int64_t f2 = (int64_t)flt2[i * flt2_stride + j] - u; const int32_t f2 = (int32_t)flt2[i * flt2_stride + j] - u;
const int64_t v = xq[0] * f1 + xq[1] * f2 + (u << SGRPROJ_PRJ_BITS); const int64_t v = xq[0] * f1 + xq[1] * f2 + (u << SGRPROJ_PRJ_BITS);
const int64_t e = const int32_t e =
ROUND_POWER_OF_TWO(v, SGRPROJ_RST_BITS + SGRPROJ_PRJ_BITS) - ROUND_POWER_OF_TWO(v, SGRPROJ_RST_BITS + SGRPROJ_PRJ_BITS) -
ROUND_POWER_OF_TWO(s, SGRPROJ_RST_BITS); ROUND_POWER_OF_TWO(s, SGRPROJ_RST_BITS);
err += e * e; err += e * e;
@ -146,9 +146,9 @@ static int64_t get_pixel_proj_error(int64_t *src, int width, int height,
return err; return err;
} }
static void get_proj_subspace(int64_t *src, int width, int height, static void get_proj_subspace(int32_t *src, int width, int height,
int src_stride, int64_t *dgd, int dgd_stride, int src_stride, int32_t *dgd, int dgd_stride,
int64_t *flt1, int flt1_stride, int64_t *flt2, int32_t *flt1, int flt1_stride, int32_t *flt2,
int flt2_stride, int *xq) { int flt2_stride, int *xq) {
int i, j; int i, j;
double H[2][2] = { { 0, 0 }, { 0, 0 } }; double H[2][2] = { { 0, 0 }, { 0, 0 } };
@ -198,10 +198,10 @@ static void search_selfguided_restoration(uint8_t *dat8, int width, int height,
int src_stride, int bit_depth, int src_stride, int bit_depth,
int *eps, int *xqd, void *srcbuf, int *eps, int *xqd, void *srcbuf,
void *rstbuf) { void *rstbuf) {
int64_t *srd = (int64_t *)srcbuf; int32_t *srd = (int32_t *)srcbuf;
int64_t *dgd = (int64_t *)rstbuf; int32_t *dgd = (int32_t *)rstbuf;
int64_t *flt1 = dgd + RESTORATION_TILEPELS_MAX; int32_t *flt1 = dgd + RESTORATION_TILEPELS_MAX;
int64_t *flt2 = flt1 + RESTORATION_TILEPELS_MAX; int32_t *flt2 = flt1 + RESTORATION_TILEPELS_MAX;
uint8_t *tmpbuf2 = (uint8_t *)(flt2 + RESTORATION_TILEPELS_MAX); uint8_t *tmpbuf2 = (uint8_t *)(flt2 + RESTORATION_TILEPELS_MAX);
int i, j, ep, bestep = 0; int i, j, ep, bestep = 0;
int64_t err, besterr = -1; int64_t err, besterr = -1;
@ -213,11 +213,11 @@ static void search_selfguided_restoration(uint8_t *dat8, int width, int height,
uint16_t *dat = CONVERT_TO_SHORTPTR(dat8); uint16_t *dat = CONVERT_TO_SHORTPTR(dat8);
for (i = 0; i < height; ++i) { for (i = 0; i < height; ++i) {
for (j = 0; j < width; ++j) { for (j = 0; j < width; ++j) {
flt1[i * width + j] = (int64_t)dat[i * dat_stride + j]; flt1[i * width + j] = (int32_t)dat[i * dat_stride + j];
flt2[i * width + j] = (int64_t)dat[i * dat_stride + j]; flt2[i * width + j] = (int32_t)dat[i * dat_stride + j];
dgd[i * width + j] = (int64_t)dat[i * dat_stride + j] dgd[i * width + j] = (int32_t)dat[i * dat_stride + j]
<< SGRPROJ_RST_BITS; << SGRPROJ_RST_BITS;
srd[i * width + j] = (int64_t)src[i * src_stride + j] srd[i * width + j] = (int32_t)src[i * src_stride + j]
<< SGRPROJ_RST_BITS; << SGRPROJ_RST_BITS;
} }
} }
@ -228,10 +228,10 @@ static void search_selfguided_restoration(uint8_t *dat8, int width, int height,
for (j = 0; j < width; ++j) { for (j = 0; j < width; ++j) {
const int k = i * width + j; const int k = i * width + j;
const int l = i * dat_stride + j; const int l = i * dat_stride + j;
flt1[k] = (int64_t)dat[l]; flt1[k] = (int32_t)dat[l];
flt2[k] = (int64_t)dat[l]; flt2[k] = (int32_t)dat[l];
dgd[k] = (int64_t)dat[l] << SGRPROJ_RST_BITS; dgd[k] = (int32_t)dat[l] << SGRPROJ_RST_BITS;
srd[k] = (int64_t)src[i * src_stride + j] << SGRPROJ_RST_BITS; srd[k] = (int32_t)src[i * src_stride + j] << SGRPROJ_RST_BITS;
} }
} }
} }