/*
* 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/common.h"
#include "avcodec.h"
#include "dsputil.h"
-#include "vp3data.h"
+#include "vp3dsp.h"
#define IdctAdjustBeforeShift 8
#define xC1S7 64277
#define xC6S2 25080
#define xC7S1 12785
-static always_inline void idct(uint8_t *dst, int stride, int16_t *input, int type)
+#define M(a,b) (((a) * (b))>>16)
+
+static av_always_inline void idct(uint8_t *dst, int stride, int16_t *input, int type)
{
int16_t *ip = input;
- uint8_t *cm = cropTbl + MAX_NEG_CROP;
- int A_, B_, C_, D_, _Ad, _Bd, _Cd, _Dd, E_, F_, G_, H_;
- int _Ed, _Gd, _Add, _Bdd, _Fd, _Hd;
- int t1, t2;
+ int A, B, C, D, Ad, Bd, Cd, Dd, E, F, G, H;
+ int Ed, Gd, Add, Bdd, Fd, Hd;
+
+ int i;
- int i, j;
-
/* 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;
-
- t1 = (int32_t)(xC4S4 * (B_ - D_));
- t1 >>= 16;
- _Bd = t1;
-
-
- _Cd = A_ + C_;
- _Dd = B_ + D_;
+ 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)(xC4S4 * (ip[0] + ip[4]));
- t1 >>= 16;
- E_ = t1;
+ Ad = M(xC4S4, (A - C));
+ Bd = M(xC4S4, (B - D));
- t1 = (int32_t)(xC4S4 * (ip[0] - ip[4]));
- t1 >>= 16;
- F_ = t1;
+ Cd = A + C;
+ Dd = B + D;
- t1 = (int32_t)(xC2S6 * ip[2]);
- t2 = (int32_t)(xC6S2 * ip[6]);
- t1 >>= 16;
- t2 >>= 16;
- G_ = t1 + t2;
+ E = M(xC4S4, (ip[0] + ip[4]));
+ F = M(xC4S4, (ip[0] - ip[4]));
- t1 = (int32_t)(xC6S2 * ip[2]);
- t2 = (int32_t)(xC2S6 * ip[6]);
- t1 >>= 16;
- t2 >>= 16;
- H_ = t1 - t2;
+ G = M(xC2S6, ip[2]) + M(xC6S2, ip[6]);
+ H = M(xC6S2, ip[2]) - M(xC2S6, ip[6]);
+ Ed = E - G;
+ Gd = E + G;
- _Ed = E_ - G_;
- _Gd = E_ + G_;
+ Add = F + Ad;
+ Bdd = Bd - H;
- _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] = _Gd + _Cd ;
- ip[7] = _Gd - _Cd ;
-
- ip[1] = _Add + _Hd;
- ip[2] = _Add - _Hd;
+ ip[0] = Gd + Cd ;
+ ip[7] = Gd - Cd ;
- ip[3] = _Ed + _Dd ;
- ip[4] = _Ed - _Dd ;
+ ip[1] = Add + Hd;
+ ip[2] = Add - Hd;
- ip[5] = _Fd + _Bdd;
- ip[6] = _Fd - _Bdd;
+ ip[3] = Ed + Dd ;
+ ip[4] = Ed - Dd ;
+ ip[5] = Fd + Bdd;
+ ip[6] = Fd - Bdd;
}
ip += 8; /* next row */
}
-
+
ip = input;
for ( i = 0; i < 8; i++) {
if ( ip[1 * 8] | ip[2 * 8] | ip[3 * 8] |
ip[4 * 8] | ip[5 * 8] | ip[6 * 8] | ip[7 * 8] ) {
- t1 = (int32_t)(xC1S7 * ip[1*8]);
- t2 = (int32_t)(xC7S1 * ip[7*8]);
- t1 >>= 16;
- t2 >>= 16;
- A_ = t1 + t2;
+ 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)(xC7S1 * ip[1*8]);
- t2 = (int32_t)(xC1S7 * ip[7*8]);
- t1 >>= 16;
- t2 >>= 16;
- B_ = t1 - t2;
+ Ad = M(xC4S4, (A - C));
+ Bd = M(xC4S4, (B - D));
- t1 = (int32_t)(xC3S5 * ip[3*8]);
- t2 = (int32_t)(xC5S3 * ip[5*8]);
- t1 >>= 16;
- t2 >>= 16;
- C_ = t1 + t2;
+ Cd = A + C;
+ Dd = B + D;
- t1 = (int32_t)(xC3S5 * ip[5*8]);
- t2 = (int32_t)(xC5S3 * ip[3*8]);
- t1 >>= 16;
- t2 >>= 16;
- D_ = t1 - t2;
+ E = M(xC4S4, (ip[0*8] + ip[4*8])) + 8;
+ F = M(xC4S4, (ip[0*8] - ip[4*8])) + 8;
+ if(type==1){ //HACK
+ E += 16*128;
+ F += 16*128;
+ }
- t1 = (int32_t)(xC4S4 * (A_ - C_));
- t1 >>= 16;
- _Ad = 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 * (B_ - D_));
- t1 >>= 16;
- _Bd = t1;
+ Ed = E - G;
+ Gd = E + G;
+ Add = F + Ad;
+ Bdd = Bd - H;
- _Cd = A_ + C_;
- _Dd = B_ + D_;
+ Fd = F - Ad;
+ Hd = Bd + H;
- t1 = (int32_t)(xC4S4 * (ip[0*8] + ip[4*8]));
- t1 >>= 16;
- E_ = t1;
+ /* Final sequence of operations over-write original inputs. */
+ if(type==0){
+ ip[0*8] = (Gd + Cd ) >> 4;
+ ip[7*8] = (Gd - Cd ) >> 4;
- t1 = (int32_t)(xC4S4 * (ip[0*8] - ip[4*8]));
- t1 >>= 16;
- F_ = t1;
+ ip[1*8] = (Add + Hd ) >> 4;
+ ip[2*8] = (Add - Hd ) >> 4;
- t1 = (int32_t)(xC2S6 * ip[2*8]);
- t2 = (int32_t)(xC6S2 * ip[6*8]);
- t1 >>= 16;
- t2 >>= 16;
- G_ = t1 + t2;
+ ip[3*8] = (Ed + Dd ) >> 4;
+ ip[4*8] = (Ed - Dd ) >> 4;
- t1 = (int32_t)(xC6S2 * ip[2*8]);
- t2 = (int32_t)(xC2S6 * ip[6*8]);
- t1 >>= 16;
- t2 >>= 16;
- H_ = t1 - t2;
+ ip[5*8] = (Fd + Bdd ) >> 4;
+ ip[6*8] = (Fd - Bdd ) >> 4;
+ }else 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);
- _Ed = E_ - G_;
- _Gd = E_ + G_;
+ dst[3*stride] = av_clip_uint8((Ed + Dd ) >> 4);
+ dst[4*stride] = av_clip_uint8((Ed - Dd ) >> 4);
- _Add = F_ + _Ad;
- _Bdd = _Bd - H_;
+ 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));
- _Fd = F_ - _Ad;
- _Hd = _Bd + H_;
+ dst[1*stride] = av_clip_uint8(dst[1*stride] + ((Add + Hd ) >> 4));
+ dst[2*stride] = av_clip_uint8(dst[2*stride] + ((Add - Hd ) >> 4));
- if(type==1){ //HACK
- _Gd += 16*128;
- _Add+= 16*128;
- _Ed += 16*128;
- _Fd += 16*128;
- }
- _Gd += IdctAdjustBeforeShift;
- _Add += IdctAdjustBeforeShift;
- _Ed += IdctAdjustBeforeShift;
- _Fd += IdctAdjustBeforeShift;
+ dst[3*stride] = av_clip_uint8(dst[3*stride] + ((Ed + Dd ) >> 4));
+ dst[4*stride] = av_clip_uint8(dst[4*stride] + ((Ed - Dd ) >> 4));
- /* Final sequence of operations over-write original inputs. */
- if(type==0){
- ip[0*8] = (_Gd + _Cd ) >> 4;
- ip[7*8] = (_Gd - _Cd ) >> 4;
-
- ip[1*8] = (_Add + _Hd ) >> 4;
- ip[2*8] = (_Add - _Hd ) >> 4;
-
- ip[3*8] = (_Ed + _Dd ) >> 4;
- ip[4*8] = (_Ed - _Dd ) >> 4;
-
- ip[5*8] = (_Fd + _Bdd ) >> 4;
- ip[6*8] = (_Fd - _Bdd ) >> 4;
- }else if(type==1){
- dst[0*stride] = cm[(_Gd + _Cd ) >> 4];
- dst[7*stride] = cm[(_Gd - _Cd ) >> 4];
-
- dst[1*stride] = cm[(_Add + _Hd ) >> 4];
- dst[2*stride] = cm[(_Add - _Hd ) >> 4];
-
- dst[3*stride] = cm[(_Ed + _Dd ) >> 4];
- dst[4*stride] = cm[(_Ed - _Dd ) >> 4];
-
- dst[5*stride] = cm[(_Fd + _Bdd ) >> 4];
- dst[6*stride] = cm[(_Fd - _Bdd ) >> 4];
- }else{
- dst[0*stride] = cm[dst[0*stride] + ((_Gd + _Cd ) >> 4)];
- dst[7*stride] = cm[dst[7*stride] + ((_Gd - _Cd ) >> 4)];
-
- dst[1*stride] = cm[dst[1*stride] + ((_Add + _Hd ) >> 4)];
- dst[2*stride] = cm[dst[2*stride] + ((_Add - _Hd ) >> 4)];
-
- dst[3*stride] = cm[dst[3*stride] + ((_Ed + _Dd ) >> 4)];
- dst[4*stride] = cm[dst[4*stride] + ((_Ed - _Dd ) >> 4)];
-
- dst[5*stride] = cm[dst[5*stride] + ((_Fd + _Bdd ) >> 4)];
- dst[6*stride] = cm[dst[6*stride] + ((_Fd - _Bdd ) >> 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==0){
- ip[0*8] =
- ip[1*8] =
- ip[2*8] =
- ip[3*8] =
- ip[4*8] =
- ip[5*8] =
+ ip[0*8] =
+ ip[1*8] =
+ ip[2*8] =
+ ip[3*8] =
+ ip[4*8] =
+ ip[5*8] =
ip[6*8] =
ip[7*8] = ((xC4S4 * ip[0*8] + (IdctAdjustBeforeShift<<16))>>20);
}else if(type==1){
dst[4*stride]=
dst[5*stride]=
dst[6*stride]=
- dst[7*stride]= 128 + ((xC4S4 * ip[0*8] + (IdctAdjustBeforeShift<<16))>>20);
+ dst[7*stride]= av_clip_uint8(128 + ((xC4S4 * ip[0*8] + (IdctAdjustBeforeShift<<16))>>20));
}else{
if(ip[0*8]){
int v= ((xC4S4 * ip[0*8] + (IdctAdjustBeforeShift<<16))>>20);
- dst[0*stride] = cm[dst[0*stride] + v];
- dst[1*stride] = cm[dst[1*stride] + v];
- dst[2*stride] = cm[dst[2*stride] + v];
- dst[3*stride] = cm[dst[3*stride] + v];
- dst[4*stride] = cm[dst[4*stride] + v];
- dst[5*stride] = cm[dst[5*stride] + v];
- dst[6*stride] = cm[dst[6*stride] + v];
- dst[7*stride] = cm[dst[7*stride] + v];
+ 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);
}
}
}
}
}
-void ff_vp3_idct_c(DCTELEM *block/* align 16*/){
- idct(NULL, 0, block, 0);
-}
-
-void ff_vp3_idct_put_c(uint8_t *dest/*align 8*/, int line_size, DCTELEM *block/*align 16*/){
+static void vp3_idct_put_c(uint8_t *dest/*align 8*/, int line_size, DCTELEM *block/*align 16*/){
idct(dest, line_size, block, 1);
+ memset(block, 0, sizeof(*block) * 64);
}
-void ff_vp3_idct_add_c(uint8_t *dest/*align 8*/, int line_size, DCTELEM *block/*align 16*/){
+static void vp3_idct_add_c(uint8_t *dest/*align 8*/, int line_size, DCTELEM *block/*align 16*/){
idct(dest, line_size, block, 2);
+ memset(block, 0, sizeof(*block) * 64);
+}
+
+static void vp3_idct_dc_add_c(uint8_t *dest/*align 8*/, int line_size,
+ DCTELEM *block/*align 16*/){
+ int i, dc = (block[0] + 15) >> 5;
+
+ 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;
+}
+
+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;
+
+ for (end= first_pixel + 8; first_pixel < end; first_pixel++) {
+ filter_value =
+ (first_pixel[2 * nstride] - first_pixel[ stride])
+ +3*(first_pixel[0 ] - first_pixel[nstride]);
+ filter_value = bounding_values[(filter_value + 4) >> 3];
+ first_pixel[nstride] = av_clip_uint8(first_pixel[nstride] + filter_value);
+ first_pixel[0] = av_clip_uint8(first_pixel[0] - filter_value);
+ }
+}
+
+static void vp3_h_loop_filter_c(uint8_t *first_pixel, int stride,
+ int *bounding_values)
+{
+ unsigned char *end;
+ int filter_value;
+
+ for (end= first_pixel + 8*stride; first_pixel != end; first_pixel += stride) {
+ filter_value =
+ (first_pixel[-2] - first_pixel[ 1])
+ +3*(first_pixel[ 0] - first_pixel[-1]);
+ filter_value = bounding_values[(filter_value + 4) >> 3];
+ first_pixel[-1] = av_clip_uint8(first_pixel[-1] + filter_value);
+ first_pixel[ 0] = av_clip_uint8(first_pixel[ 0] - filter_value);
+ }
+}
+
+av_cold void ff_vp3dsp_init(VP3DSPContext *c, int flags)
+{
+ 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;
+
+ c->idct_perm = FF_NO_IDCT_PERM;
+
+ 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);
}
projects / ffmpeg / blobdiff