diff options
author | Clément Bœsch <u@pkh.me> | 2014-08-02 18:54:43 +0200 |
---|---|---|
committer | Clément Bœsch <u@pkh.me> | 2014-08-05 21:05:52 +0200 |
commit | 28a2107a8d61af7c7a26f9d4af0716ba12c112a7 (patch) | |
tree | 2dd49b789e0296fe6238adb9cc6be46c304403ce /libavutil/pixelutils.c | |
parent | 6931d1274568ecb1c957157ed92066bde52073d8 (diff) | |
download | ffmpeg-28a2107a8d61af7c7a26f9d4af0716ba12c112a7.tar.gz |
avutil: add pixelutils API
Diffstat (limited to 'libavutil/pixelutils.c')
-rw-r--r-- | libavutil/pixelutils.c | 153 |
1 files changed, 153 insertions, 0 deletions
diff --git a/libavutil/pixelutils.c b/libavutil/pixelutils.c new file mode 100644 index 0000000000..cfdd35afc2 --- /dev/null +++ b/libavutil/pixelutils.c @@ -0,0 +1,153 @@ +/* + * This file is part of FFmpeg. + * + * FFmpeg is free software; you can redistribute it and/or + * modify it under the terms of the GNU Lesser General Public + * License as published by the Free Software Foundation; either + * version 2.1 of the License, or (at your option) any later version. + * + * FFmpeg is distributed in the hope that it will be useful, + * but WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU + * Lesser General Public License for more details. + * + * You should have received a copy of the GNU Lesser General Public + * License along with FFmpeg; if not, write to the Free Software + * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA + */ + +#include "config.h" +#include "common.h" +#include "pixelutils.h" + +#if CONFIG_PIXELUTILS + +#include "x86/pixelutils.h" + +static av_always_inline int sad_wxh(const uint8_t *src1, ptrdiff_t stride1, + const uint8_t *src2, ptrdiff_t stride2, + int w, int h) +{ + int x, y, sum = 0; + + for (y = 0; y < h; y++) { + for (x = 0; x < w; x++) + sum += abs(src1[x] - src2[x]); + src1 += stride1; + src2 += stride2; + } + return sum; +} + +#define DECLARE_BLOCK_FUNCTIONS(size) \ +static int block_sad_##size##x##size##_c(const uint8_t *src1, ptrdiff_t stride1, \ + const uint8_t *src2, ptrdiff_t stride2) \ +{ \ + return sad_wxh(src1, stride1, src2, stride2, size, size); \ +} + +DECLARE_BLOCK_FUNCTIONS(2) +DECLARE_BLOCK_FUNCTIONS(4) +DECLARE_BLOCK_FUNCTIONS(8) +DECLARE_BLOCK_FUNCTIONS(16) + +static const av_pixelutils_sad_fn sad_c[] = { + block_sad_2x2_c, + block_sad_4x4_c, + block_sad_8x8_c, + block_sad_16x16_c, +}; + +#endif /* CONFIG_PIXELUTILS */ + +av_pixelutils_sad_fn av_pixelutils_get_sad_fn(int w_bits, int h_bits, int aligned, void *log_ctx) +{ +#if !CONFIG_PIXELUTILS + av_log(log_ctx, AV_LOG_ERROR, "pixelutils support is required " + "but libavutil is not compiled with it\n"); + return NULL; +#else + av_pixelutils_sad_fn sad[FF_ARRAY_ELEMS(sad_c)]; + + memcpy(sad, sad_c, sizeof(sad)); + + if (w_bits < 1 || w_bits > FF_ARRAY_ELEMS(sad) || + h_bits < 1 || h_bits > FF_ARRAY_ELEMS(sad)) + return NULL; + if (w_bits != h_bits) // only squared sad for now + return NULL; + +#if ARCH_X86 + ff_pixelutils_sad_init_x86(sad, aligned); +#endif + + return sad[w_bits - 1]; +#endif +} + +#ifdef TEST +#define W1 320 +#define H1 240 +#define W2 640 +#define H2 480 + +static int run_test(const char *test, + const uint32_t *b1, const uint32_t *b2) +{ + int i, a, ret = 0; + + for (a = 0; a < 3; a++) { + const uint8_t *block1 = (const uint8_t *)b1; + const uint8_t *block2 = (const uint8_t *)b2; + + switch (a) { + case 0: block1++; block2++; break; + case 1: block2++; break; + case 2: break; + } + for (i = 1; i <= FF_ARRAY_ELEMS(sad_c); i++) { + av_pixelutils_sad_fn f_ref = sad_c[i - 1]; + av_pixelutils_sad_fn f_out = av_pixelutils_get_sad_fn(i, i, a, NULL); + const int out = f_out(block1, W1, block2, W2); + const int ref = f_ref(block1, W1, block2, W2); + printf("[%s] [%c%c] SAD [%s] %dx%d=%d ref=%d\n", + out == ref ? "OK" : "FAIL", + a ? 'A' : 'U', a == 2 ? 'A' : 'U', + test, 1<<i, 1<<i, out, ref); + if (out != ref) + ret = 1; + } + } + return ret; +} + +int main(void) +{ + int i, ret; + DECLARE_ALIGNED(32, uint32_t, buf1)[W1*H1]; + DECLARE_ALIGNED(32, uint32_t, buf2)[W2*H2]; + uint32_t state = 0; + + for (i = 0; i < W1*H1; i++) { + state = state * 1664525 + 1013904223; + buf1[i] = state; + } + for (i = 0; i < W2*H2; i++) { + state = state * 1664525 + 1013904223; + buf2[i] = state; + } + ret = run_test("random", buf1, buf2); + if (ret < 0) + return ret; + + memset(buf1, 0xff, sizeof(buf1)); + memset(buf2, 0x00, sizeof(buf2)); + ret = run_test("max", buf1, buf2); + if (ret < 0) + return ret; + + memset(buf1, 0x90, sizeof(buf1)); + memset(buf2, 0x90, sizeof(buf2)); + return run_test("min", buf1, buf2); +} +#endif /* TEST */ |