/*
* Copyright (C) 2004 the ffmpeg project
*
- * This library is free software; you can redistribute it and/or
+ * This file is part of Libav.
+ *
+ * Libav 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 of the License, or (at your option) any later version.
+ * version 2.1 of the License, or (at your option) any later version.
*
- * This library is distributed in the hope that it will be useful,
+ * Libav 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 this library; if not, write to the Free Software
- * Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA
+ * License along with Libav; if not, write to the Free Software
+ * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
*/
/**
- * @file vp3dsp.c
- * Standard C DSP-oriented functions cribbed from the original VP3
+ * @file
+ * Standard C DSP-oriented functions cribbed from the original VP3
* source code.
*/
-#include "common.h"
+#include "libavutil/attributes.h"
+#include "libavutil/intreadwrite.h"
+#include "libavutil/common.h"
+#include "libavutil/intreadwrite.h"
+
#include "avcodec.h"
-#include "dsputil.h"
-#include "vp3data.h"
+#include "rnd_avg.h"
+#include "vp3dsp.h"
#define IdctAdjustBeforeShift 8
#define xC1S7 64277
#define xC6S2 25080
#define xC7S1 12785
-void vp3_dsp_init_c(void)
-{
- /* nop */
-}
+#define M(a, b) (((a) * (b)) >> 16)
-void vp3_idct_c(int16_t *input_data, int16_t *dequant_matrix,
- int coeff_count, int16_t *output_data)
+static av_always_inline void idct(uint8_t *dst, int stride,
+ int16_t *input, int type)
{
- int32_t dequantized_data[64];
- int32_t *ip = dequantized_data;
- int16_t *op = output_data;
-
- int32_t A_, B_, C_, D_, _Ad, _Bd, _Cd, _Dd, E_, F_, G_, H_;
- int32_t _Ed, _Gd, _Add, _Bdd, _Fd, _Hd;
- int32_t t1, t2;
-
- int i, j;
-
- /* de-zigzag and dequantize */
- for (i = 0; i < coeff_count; i++) {
- dequantized_data[i] = dequant_matrix[i] * input_data[i];
- }
+ int16_t *ip = input;
+
+ int A, B, C, D, Ad, Bd, Cd, Dd, E, F, G, H;
+ int Ed, Gd, Add, Bdd, Fd, Hd;
+
+ int i;
/* Inverse DCT on the rows now */
for (i = 0; i < 8; i++) {
/* Check for non-zero values */
- if ( ip[0] | ip[1] | ip[2] | ip[3] | ip[4] | ip[5] | ip[6] | ip[7] ) {
- t1 = (int32_t)(xC1S7 * ip[1]);
- t2 = (int32_t)(xC7S1 * ip[7]);
- t1 >>= 16;
- t2 >>= 16;
- A_ = t1 + t2;
-
- t1 = (int32_t)(xC7S1 * ip[1]);
- t2 = (int32_t)(xC1S7 * ip[7]);
- t1 >>= 16;
- t2 >>= 16;
- B_ = t1 - t2;
-
- t1 = (int32_t)(xC3S5 * ip[3]);
- t2 = (int32_t)(xC5S3 * ip[5]);
- t1 >>= 16;
- t2 >>= 16;
- C_ = t1 + t2;
-
- t1 = (int32_t)(xC3S5 * ip[5]);
- t2 = (int32_t)(xC5S3 * ip[3]);
- t1 >>= 16;
- t2 >>= 16;
- D_ = t1 - t2;
-
-
- t1 = (int32_t)(xC4S4 * (A_ - C_));
- t1 >>= 16;
- _Ad = t1;
+ if (ip[0 * 8] | ip[1 * 8] | ip[2 * 8] | ip[3 * 8] |
+ ip[4 * 8] | ip[5 * 8] | ip[6 * 8] | ip[7 * 8]) {
+ A = M(xC1S7, ip[1 * 8]) + M(xC7S1, ip[7 * 8]);
+ B = M(xC7S1, ip[1 * 8]) - M(xC1S7, ip[7 * 8]);
+ C = M(xC3S5, ip[3 * 8]) + M(xC5S3, ip[5 * 8]);
+ D = M(xC3S5, ip[5 * 8]) - M(xC5S3, ip[3 * 8]);
- t1 = (int32_t)(xC4S4 * (B_ - D_));
- t1 >>= 16;
- _Bd = t1;
+ Ad = M(xC4S4, (A - C));
+ Bd = M(xC4S4, (B - D));
+ Cd = A + C;
+ Dd = B + D;
- _Cd = A_ + C_;
- _Dd = B_ + D_;
+ E = M(xC4S4, (ip[0 * 8] + ip[4 * 8]));
+ F = M(xC4S4, (ip[0 * 8] - ip[4 * 8]));
- t1 = (int32_t)(xC4S4 * (ip[0] + ip[4]));
- t1 >>= 16;
- E_ = t1;
+ G = M(xC2S6, ip[2 * 8]) + M(xC6S2, ip[6 * 8]);
+ H = M(xC6S2, ip[2 * 8]) - M(xC2S6, ip[6 * 8]);
- t1 = (int32_t)(xC4S4 * (ip[0] - ip[4]));
- t1 >>= 16;
- F_ = t1;
+ Ed = E - G;
+ Gd = E + G;
- t1 = (int32_t)(xC2S6 * ip[2]);
- t2 = (int32_t)(xC6S2 * ip[6]);
- t1 >>= 16;
- t2 >>= 16;
- G_ = t1 + t2;
+ Add = F + Ad;
+ Bdd = Bd - H;
- t1 = (int32_t)(xC6S2 * ip[2]);
- t2 = (int32_t)(xC2S6 * ip[6]);
- t1 >>= 16;
- t2 >>= 16;
- H_ = t1 - t2;
-
-
- _Ed = E_ - G_;
- _Gd = E_ + G_;
-
- _Add = F_ + _Ad;
- _Bdd = _Bd - H_;
-
- _Fd = F_ - _Ad;
- _Hd = _Bd + H_;
+ Fd = F - Ad;
+ Hd = Bd + H;
/* Final sequence of operations over-write original inputs. */
- ip[0] = (int16_t)((_Gd + _Cd ) >> 0);
- ip[7] = (int16_t)((_Gd - _Cd ) >> 0);
-
- ip[1] = (int16_t)((_Add + _Hd ) >> 0);
- ip[2] = (int16_t)((_Add - _Hd ) >> 0);
+ ip[0 * 8] = Gd + Cd;
+ ip[7 * 8] = Gd - Cd;
- ip[3] = (int16_t)((_Ed + _Dd ) >> 0);
- ip[4] = (int16_t)((_Ed - _Dd ) >> 0);
+ ip[1 * 8] = Add + Hd;
+ ip[2 * 8] = Add - Hd;
- ip[5] = (int16_t)((_Fd + _Bdd ) >> 0);
- ip[6] = (int16_t)((_Fd - _Bdd ) >> 0);
+ ip[3 * 8] = Ed + Dd;
+ ip[4 * 8] = Ed - Dd;
+ ip[5 * 8] = Fd + Bdd;
+ ip[6 * 8] = Fd - Bdd;
}
- ip += 8; /* next row */
+ ip += 1; /* next row */
}
- ip = dequantized_data;
+ ip = input;
- for ( i = 0; i < 8; i++) {
+ for (i = 0; i < 8; i++) {
/* Check for non-zero values (bitwise or faster than ||) */
- if ( ip[0 * 8] | ip[1 * 8] | ip[2 * 8] | ip[3 * 8] |
- ip[4 * 8] | ip[5 * 8] | ip[6 * 8] | ip[7 * 8] ) {
+ if (ip[1] | ip[2] | ip[3] |
+ ip[4] | ip[5] | ip[6] | ip[7]) {
+ A = M(xC1S7, ip[1]) + M(xC7S1, ip[7]);
+ B = M(xC7S1, ip[1]) - M(xC1S7, ip[7]);
+ C = M(xC3S5, ip[3]) + M(xC5S3, ip[5]);
+ D = M(xC3S5, ip[5]) - M(xC5S3, ip[3]);
- t1 = (int32_t)(xC1S7 * ip[1*8]);
- t2 = (int32_t)(xC7S1 * ip[7*8]);
- t1 >>= 16;
- t2 >>= 16;
- A_ = t1 + t2;
+ Ad = M(xC4S4, (A - C));
+ Bd = M(xC4S4, (B - D));
- t1 = (int32_t)(xC7S1 * ip[1*8]);
- t2 = (int32_t)(xC1S7 * ip[7*8]);
- t1 >>= 16;
- t2 >>= 16;
- B_ = t1 - t2;
+ Cd = A + C;
+ Dd = B + D;
- t1 = (int32_t)(xC3S5 * ip[3*8]);
- t2 = (int32_t)(xC5S3 * ip[5*8]);
- t1 >>= 16;
- t2 >>= 16;
- C_ = t1 + t2;
+ E = M(xC4S4, (ip[0] + ip[4])) + 8;
+ F = M(xC4S4, (ip[0] - ip[4])) + 8;
- t1 = (int32_t)(xC3S5 * ip[5*8]);
- t2 = (int32_t)(xC5S3 * ip[3*8]);
- t1 >>= 16;
- t2 >>= 16;
- D_ = t1 - t2;
+ if (type == 1) { // HACK
+ E += 16 * 128;
+ F += 16 * 128;
+ }
+ G = M(xC2S6, ip[2]) + M(xC6S2, ip[6]);
+ H = M(xC6S2, ip[2]) - M(xC2S6, ip[6]);
- t1 = (int32_t)(xC4S4 * (A_ - C_));
- t1 >>= 16;
- _Ad = t1;
+ Ed = E - G;
+ Gd = E + G;
- t1 = (int32_t)(xC4S4 * (B_ - D_));
- t1 >>= 16;
- _Bd = t1;
+ Add = F + Ad;
+ Bdd = Bd - H;
+ Fd = F - Ad;
+ Hd = Bd + H;
- _Cd = A_ + C_;
- _Dd = B_ + D_;
+ /* Final sequence of operations over-write original inputs. */
+ if (type == 1) {
+ dst[0 * stride] = av_clip_uint8((Gd + Cd) >> 4);
+ dst[7 * stride] = av_clip_uint8((Gd - Cd) >> 4);
+
+ dst[1 * stride] = av_clip_uint8((Add + Hd) >> 4);
+ dst[2 * stride] = av_clip_uint8((Add - Hd) >> 4);
+
+ dst[3 * stride] = av_clip_uint8((Ed + Dd) >> 4);
+ dst[4 * stride] = av_clip_uint8((Ed - Dd) >> 4);
- t1 = (int32_t)(xC4S4 * (ip[0*8] + ip[4*8]));
- t1 >>= 16;
- E_ = t1;
+ dst[5 * stride] = av_clip_uint8((Fd + Bdd) >> 4);
+ dst[6 * stride] = av_clip_uint8((Fd - Bdd) >> 4);
+ } else {
+ dst[0 * stride] = av_clip_uint8(dst[0 * stride] + ((Gd + Cd) >> 4));
+ dst[7 * stride] = av_clip_uint8(dst[7 * stride] + ((Gd - Cd) >> 4));
- t1 = (int32_t)(xC4S4 * (ip[0*8] - ip[4*8]));
- t1 >>= 16;
- F_ = t1;
+ dst[1 * stride] = av_clip_uint8(dst[1 * stride] + ((Add + Hd) >> 4));
+ dst[2 * stride] = av_clip_uint8(dst[2 * stride] + ((Add - Hd) >> 4));
+
+ dst[3 * stride] = av_clip_uint8(dst[3 * stride] + ((Ed + Dd) >> 4));
+ dst[4 * stride] = av_clip_uint8(dst[4 * stride] + ((Ed - Dd) >> 4));
+
+ dst[5 * stride] = av_clip_uint8(dst[5 * stride] + ((Fd + Bdd) >> 4));
+ dst[6 * stride] = av_clip_uint8(dst[6 * stride] + ((Fd - Bdd) >> 4));
+ }
+ } else {
+ if (type == 1) {
+ dst[0*stride] =
+ dst[1*stride] =
+ dst[2*stride] =
+ dst[3*stride] =
+ dst[4*stride] =
+ dst[5*stride] =
+ dst[6*stride] =
+ dst[7*stride] = av_clip_uint8(128 + ((xC4S4 * ip[0] + (IdctAdjustBeforeShift << 16)) >> 20));
+ } else {
+ if (ip[0]) {
+ int v = (xC4S4 * ip[0] + (IdctAdjustBeforeShift << 16)) >> 20;
+ dst[0 * stride] = av_clip_uint8(dst[0 * stride] + v);
+ dst[1 * stride] = av_clip_uint8(dst[1 * stride] + v);
+ dst[2 * stride] = av_clip_uint8(dst[2 * stride] + v);
+ dst[3 * stride] = av_clip_uint8(dst[3 * stride] + v);
+ dst[4 * stride] = av_clip_uint8(dst[4 * stride] + v);
+ dst[5 * stride] = av_clip_uint8(dst[5 * stride] + v);
+ dst[6 * stride] = av_clip_uint8(dst[6 * stride] + v);
+ dst[7 * stride] = av_clip_uint8(dst[7 * stride] + v);
+ }
+ }
+ }
- t1 = (int32_t)(xC2S6 * ip[2*8]);
- t2 = (int32_t)(xC6S2 * ip[6*8]);
- t1 >>= 16;
- t2 >>= 16;
- G_ = t1 + t2;
+ ip += 8; /* next column */
+ dst++;
+ }
+}
- t1 = (int32_t)(xC6S2 * ip[2*8]);
- t2 = (int32_t)(xC2S6 * ip[6*8]);
- t1 >>= 16;
- t2 >>= 16;
- H_ = t1 - t2;
+static void vp3_idct_put_c(uint8_t *dest /* align 8 */, int line_size,
+ int16_t *block /* align 16 */)
+{
+ idct(dest, line_size, block, 1);
+ memset(block, 0, sizeof(*block) * 64);
+}
+static void vp3_idct_add_c(uint8_t *dest /* align 8 */, int line_size,
+ int16_t *block /* align 16 */)
+{
+ idct(dest, line_size, block, 2);
+ memset(block, 0, sizeof(*block) * 64);
+}
- _Ed = E_ - G_;
- _Gd = E_ + G_;
+static void vp3_idct_dc_add_c(uint8_t *dest /* align 8 */, int line_size,
+ int16_t *block /* align 16 */)
+{
+ int i, dc = (block[0] + 15) >> 5;
- _Add = F_ + _Ad;
- _Bdd = _Bd - H_;
+ for (i = 0; i < 8; i++) {
+ dest[0] = av_clip_uint8(dest[0] + dc);
+ dest[1] = av_clip_uint8(dest[1] + dc);
+ dest[2] = av_clip_uint8(dest[2] + dc);
+ dest[3] = av_clip_uint8(dest[3] + dc);
+ dest[4] = av_clip_uint8(dest[4] + dc);
+ dest[5] = av_clip_uint8(dest[5] + dc);
+ dest[6] = av_clip_uint8(dest[6] + dc);
+ dest[7] = av_clip_uint8(dest[7] + dc);
+ dest += line_size;
+ }
+ block[0] = 0;
+}
- _Fd = F_ - _Ad;
- _Hd = _Bd + H_;
+static void vp3_v_loop_filter_c(uint8_t *first_pixel, int stride,
+ int *bounding_values)
+{
+ unsigned char *end;
+ int filter_value;
+ const int nstride = -stride;
- _Gd += IdctAdjustBeforeShift;
- _Add += IdctAdjustBeforeShift;
- _Ed += IdctAdjustBeforeShift;
- _Fd += IdctAdjustBeforeShift;
+ for (end = first_pixel + 8; first_pixel < end; first_pixel++) {
+ filter_value = (first_pixel[2 * nstride] - first_pixel[stride]) +
+ (first_pixel[0] - first_pixel[nstride]) * 3;
+ filter_value = bounding_values[(filter_value + 4) >> 3];
- /* Final sequence of operations over-write original inputs. */
- op[0*8] = (int16_t)((_Gd + _Cd ) >> 4);
- op[7*8] = (int16_t)((_Gd - _Cd ) >> 4);
+ first_pixel[nstride] = av_clip_uint8(first_pixel[nstride] + filter_value);
+ first_pixel[0] = av_clip_uint8(first_pixel[0] - filter_value);
+ }
+}
- op[1*8] = (int16_t)((_Add + _Hd ) >> 4);
- op[2*8] = (int16_t)((_Add - _Hd ) >> 4);
+static void vp3_h_loop_filter_c(uint8_t *first_pixel, int stride,
+ int *bounding_values)
+{
+ unsigned char *end;
+ int filter_value;
- op[3*8] = (int16_t)((_Ed + _Dd ) >> 4);
- op[4*8] = (int16_t)((_Ed - _Dd ) >> 4);
+ for (end = first_pixel + 8 * stride; first_pixel != end; first_pixel += stride) {
+ filter_value = (first_pixel[-2] - first_pixel[1]) +
+ (first_pixel[ 0] - first_pixel[-1]) * 3;
+ filter_value = bounding_values[(filter_value + 4) >> 3];
- op[5*8] = (int16_t)((_Fd + _Bdd ) >> 4);
- op[6*8] = (int16_t)((_Fd - _Bdd ) >> 4);
+ first_pixel[-1] = av_clip_uint8(first_pixel[-1] + filter_value);
+ first_pixel[ 0] = av_clip_uint8(first_pixel[ 0] - filter_value);
+ }
+}
- } else {
+static void put_no_rnd_pixels_l2(uint8_t *dst, const uint8_t *src1,
+ const uint8_t *src2, ptrdiff_t stride, int h)
+{
+ int i;
- op[0*8] = 0;
- op[7*8] = 0;
- op[1*8] = 0;
- op[2*8] = 0;
- op[3*8] = 0;
- op[4*8] = 0;
- op[5*8] = 0;
- op[6*8] = 0;
- }
+ for (i = 0; i < h; i++) {
+ uint32_t a, b;
- ip++; /* next column */
- op++;
+ a = AV_RN32(&src1[i * stride]);
+ b = AV_RN32(&src2[i * stride]);
+ AV_WN32A(&dst[i * stride], no_rnd_avg32(a, b));
+ a = AV_RN32(&src1[i * stride + 4]);
+ b = AV_RN32(&src2[i * stride + 4]);
+ AV_WN32A(&dst[i * stride + 4], no_rnd_avg32(a, b));
}
}
+
+av_cold void ff_vp3dsp_init(VP3DSPContext *c, int flags)
+{
+ c->put_no_rnd_pixels_l2 = put_no_rnd_pixels_l2;
+
+ c->idct_put = vp3_idct_put_c;
+ c->idct_add = vp3_idct_add_c;
+ c->idct_dc_add = vp3_idct_dc_add_c;
+ c->v_loop_filter = vp3_v_loop_filter_c;
+ c->h_loop_filter = vp3_h_loop_filter_c;
+
+ if (ARCH_ARM)
+ ff_vp3dsp_init_arm(c, flags);
+ if (ARCH_PPC)
+ ff_vp3dsp_init_ppc(c, flags);
+ if (ARCH_X86)
+ ff_vp3dsp_init_x86(c, flags);
+}