X-Git-Url: https://git.sesse.net/?a=blobdiff_plain;f=libavcodec%2Fsimple_idct.c;h=6670740bf3d9d5055b20585a84f064ff5c2a0bc3;hb=8d872e8ac9d17dd004af60787cd69eac69ef8891;hp=8c9ce7b9368d8f93662692441da3d56406aba9d4;hpb=1457ab523343e94e094ad1c60de37077f8dc5589;p=ffmpeg diff --git a/libavcodec/simple_idct.c b/libavcodec/simple_idct.c index 8c9ce7b9368..6670740bf3d 100644 --- a/libavcodec/simple_idct.c +++ b/libavcodec/simple_idct.c @@ -3,26 +3,35 @@ * * Copyright (c) 2001 Michael Niedermayer * - * This library is free software; you can redistribute it and/or + * 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 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, + * 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 this library; if not, write to the Free Software - * Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA + * License along with FFmpeg; if not, write to the Free Software + * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA + */ + +/** + * @file libavcodec/simple_idct.c + * simpleidct in C. */ + /* based upon some outcommented c code from mpeg2dec (idct_mmx.c - written by Aaron Holtzman ) + written by Aaron Holtzman ) */ #include "avcodec.h" #include "dsputil.h" +#include "mathops.h" #include "simple_idct.h" #if 0 @@ -47,67 +56,61 @@ #define COL_SHIFT 20 // 6 #endif -#if defined(ARCH_POWERPC_405) - -/* signed 16x16 -> 32 multiply add accumulate */ -#define MAC16(rt, ra, rb) \ - asm ("maclhw %0, %2, %3" : "=r" (rt) : "0" (rt), "r" (ra), "r" (rb)); - -/* signed 16x16 -> 32 multiply */ -#define MUL16(rt, ra, rb) \ - asm ("mullhw %0, %1, %2" : "=r" (rt) : "r" (ra), "r" (rb)); - -#else - -/* signed 16x16 -> 32 multiply add accumulate */ -#define MAC16(rt, ra, rb) rt += (ra) * (rb) - -/* signed 16x16 -> 32 multiply */ -#define MUL16(rt, ra, rb) rt = (ra) * (rb) - -#endif - -static inline void idctRowCondDC (int16_t * row) +static inline void idctRowCondDC (DCTELEM * row) { - int a0, a1, a2, a3, b0, b1, b2, b3; -#ifdef FAST_64BIT + int a0, a1, a2, a3, b0, b1, b2, b3; +#if HAVE_FAST_64BIT uint64_t temp; #else uint32_t temp; #endif -#ifdef FAST_64BIT -#ifdef WORDS_BIGENDIAN +#if HAVE_FAST_64BIT +#if HAVE_BIGENDIAN #define ROW0_MASK 0xffff000000000000LL #else #define ROW0_MASK 0xffffLL #endif - if ( ((((uint64_t *)row)[0] & ~ROW0_MASK) | - ((uint64_t *)row)[1]) == 0) { - temp = (row[0] << 3) & 0xffff; - temp += temp << 16; - temp += temp << 32; - ((uint64_t *)row)[0] = temp; - ((uint64_t *)row)[1] = temp; - return; - } + if(sizeof(DCTELEM)==2){ + if ( ((((uint64_t *)row)[0] & ~ROW0_MASK) | + ((uint64_t *)row)[1]) == 0) { + temp = (row[0] << 3) & 0xffff; + temp += temp << 16; + temp += temp << 32; + ((uint64_t *)row)[0] = temp; + ((uint64_t *)row)[1] = temp; + return; + } + }else{ + if (!(row[1]|row[2]|row[3]|row[4]|row[5]|row[6]|row[7])) { + row[0]=row[1]=row[2]=row[3]=row[4]=row[5]=row[6]=row[7]= row[0] << 3; + return; + } + } #else - if (!(((uint32_t*)row)[1] | - ((uint32_t*)row)[2] | - ((uint32_t*)row)[3] | - row[1])) { - temp = (row[0] << 3) & 0xffff; - temp += temp << 16; - ((uint32_t*)row)[0]=((uint32_t*)row)[1] = - ((uint32_t*)row)[2]=((uint32_t*)row)[3] = temp; - return; - } + if(sizeof(DCTELEM)==2){ + if (!(((uint32_t*)row)[1] | + ((uint32_t*)row)[2] | + ((uint32_t*)row)[3] | + row[1])) { + temp = (row[0] << 3) & 0xffff; + temp += temp << 16; + ((uint32_t*)row)[0]=((uint32_t*)row)[1] = + ((uint32_t*)row)[2]=((uint32_t*)row)[3] = temp; + return; + } + }else{ + if (!(row[1]|row[2]|row[3]|row[4]|row[5]|row[6]|row[7])) { + row[0]=row[1]=row[2]=row[3]=row[4]=row[5]=row[6]=row[7]= row[0] << 3; + return; + } + } #endif a0 = (W4 * row[0]) + (1 << (ROW_SHIFT - 1)); - a1 = a0; - a2 = a0; - a3 = a0; + a1 = a0; + a2 = a0; + a3 = a0; /* no need to optimize : gcc does it */ a0 += W2 * row[2]; @@ -115,21 +118,21 @@ static inline void idctRowCondDC (int16_t * row) a2 -= W6 * row[2]; a3 -= W2 * row[2]; - MUL16(b0, W1, row[1]); + b0 = MUL16(W1, row[1]); MAC16(b0, W3, row[3]); - MUL16(b1, W3, row[1]); + b1 = MUL16(W3, row[1]); MAC16(b1, -W7, row[3]); - MUL16(b2, W5, row[1]); + b2 = MUL16(W5, row[1]); MAC16(b2, -W1, row[3]); - MUL16(b3, W7, row[1]); + b3 = MUL16(W7, row[1]); MAC16(b3, -W5, row[3]); -#ifdef FAST_64BIT +#if HAVE_FAST_64BIT temp = ((uint64_t*)row)[1]; #else temp = ((uint32_t*)row)[2] | ((uint32_t*)row)[3]; #endif - if (temp != 0) { + if (temp != 0) { a0 += W4*row[4] + W6*row[6]; a1 += - W4*row[4] - W2*row[6]; a2 += - W4*row[4] + W2*row[6]; @@ -137,81 +140,81 @@ static inline void idctRowCondDC (int16_t * row) MAC16(b0, W5, row[5]); MAC16(b0, W7, row[7]); - + MAC16(b1, -W1, row[5]); MAC16(b1, -W5, row[7]); - + MAC16(b2, W7, row[5]); MAC16(b2, W3, row[7]); - + MAC16(b3, W3, row[5]); MAC16(b3, -W1, row[7]); - } - - row[0] = (a0 + b0) >> ROW_SHIFT; - row[7] = (a0 - b0) >> ROW_SHIFT; - row[1] = (a1 + b1) >> ROW_SHIFT; - row[6] = (a1 - b1) >> ROW_SHIFT; - row[2] = (a2 + b2) >> ROW_SHIFT; - row[5] = (a2 - b2) >> ROW_SHIFT; - row[3] = (a3 + b3) >> ROW_SHIFT; - row[4] = (a3 - b3) >> ROW_SHIFT; + } + + row[0] = (a0 + b0) >> ROW_SHIFT; + row[7] = (a0 - b0) >> ROW_SHIFT; + row[1] = (a1 + b1) >> ROW_SHIFT; + row[6] = (a1 - b1) >> ROW_SHIFT; + row[2] = (a2 + b2) >> ROW_SHIFT; + row[5] = (a2 - b2) >> ROW_SHIFT; + row[3] = (a3 + b3) >> ROW_SHIFT; + row[4] = (a3 - b3) >> ROW_SHIFT; } -static inline void idctSparseColPut (UINT8 *dest, int line_size, - int16_t * col) +static inline void idctSparseColPut (uint8_t *dest, int line_size, + DCTELEM * col) { - int a0, a1, a2, a3, b0, b1, b2, b3; - UINT8 *cm = cropTbl + MAX_NEG_CROP; + int a0, a1, a2, a3, b0, b1, b2, b3; + uint8_t *cm = ff_cropTbl + MAX_NEG_CROP; /* XXX: I did that only to give same values as previous code */ - a0 = W4 * (col[8*0] + ((1<<(COL_SHIFT-1))/W4)); - a1 = a0; - a2 = a0; - a3 = a0; + a0 = W4 * (col[8*0] + ((1<<(COL_SHIFT-1))/W4)); + a1 = a0; + a2 = a0; + a3 = a0; a0 += + W2*col[8*2]; a1 += + W6*col[8*2]; a2 += - W6*col[8*2]; a3 += - W2*col[8*2]; - MUL16(b0, W1, col[8*1]); - MUL16(b1, W3, col[8*1]); - MUL16(b2, W5, col[8*1]); - MUL16(b3, W7, col[8*1]); + b0 = MUL16(W1, col[8*1]); + b1 = MUL16(W3, col[8*1]); + b2 = MUL16(W5, col[8*1]); + b3 = MUL16(W7, col[8*1]); MAC16(b0, + W3, col[8*3]); MAC16(b1, - W7, col[8*3]); MAC16(b2, - W1, col[8*3]); MAC16(b3, - W5, col[8*3]); - if(col[8*4]){ + if(col[8*4]){ a0 += + W4*col[8*4]; a1 += - W4*col[8*4]; a2 += - W4*col[8*4]; a3 += + W4*col[8*4]; - } + } - if (col[8*5]) { + if (col[8*5]) { MAC16(b0, + W5, col[8*5]); MAC16(b1, - W1, col[8*5]); MAC16(b2, + W7, col[8*5]); MAC16(b3, + W3, col[8*5]); - } + } - if(col[8*6]){ + if(col[8*6]){ a0 += + W6*col[8*6]; a1 += - W2*col[8*6]; a2 += + W2*col[8*6]; a3 += - W6*col[8*6]; - } + } - if (col[8*7]) { + if (col[8*7]) { MAC16(b0, + W7, col[8*7]); MAC16(b1, - W5, col[8*7]); MAC16(b2, + W3, col[8*7]); MAC16(b3, - W1, col[8*7]); - } + } dest[0] = cm[(a0 + b0) >> COL_SHIFT]; dest += line_size; @@ -230,60 +233,60 @@ static inline void idctSparseColPut (UINT8 *dest, int line_size, dest[0] = cm[(a0 - b0) >> COL_SHIFT]; } -static inline void idctSparseColAdd (UINT8 *dest, int line_size, - int16_t * col) +static inline void idctSparseColAdd (uint8_t *dest, int line_size, + DCTELEM * col) { - int a0, a1, a2, a3, b0, b1, b2, b3; - UINT8 *cm = cropTbl + MAX_NEG_CROP; + int a0, a1, a2, a3, b0, b1, b2, b3; + uint8_t *cm = ff_cropTbl + MAX_NEG_CROP; /* XXX: I did that only to give same values as previous code */ - a0 = W4 * (col[8*0] + ((1<<(COL_SHIFT-1))/W4)); - a1 = a0; - a2 = a0; - a3 = a0; + a0 = W4 * (col[8*0] + ((1<<(COL_SHIFT-1))/W4)); + a1 = a0; + a2 = a0; + a3 = a0; a0 += + W2*col[8*2]; a1 += + W6*col[8*2]; a2 += - W6*col[8*2]; a3 += - W2*col[8*2]; - MUL16(b0, W1, col[8*1]); - MUL16(b1, W3, col[8*1]); - MUL16(b2, W5, col[8*1]); - MUL16(b3, W7, col[8*1]); + b0 = MUL16(W1, col[8*1]); + b1 = MUL16(W3, col[8*1]); + b2 = MUL16(W5, col[8*1]); + b3 = MUL16(W7, col[8*1]); MAC16(b0, + W3, col[8*3]); MAC16(b1, - W7, col[8*3]); MAC16(b2, - W1, col[8*3]); MAC16(b3, - W5, col[8*3]); - if(col[8*4]){ + if(col[8*4]){ a0 += + W4*col[8*4]; a1 += - W4*col[8*4]; a2 += - W4*col[8*4]; a3 += + W4*col[8*4]; - } + } - if (col[8*5]) { + if (col[8*5]) { MAC16(b0, + W5, col[8*5]); MAC16(b1, - W1, col[8*5]); MAC16(b2, + W7, col[8*5]); MAC16(b3, + W3, col[8*5]); - } + } - if(col[8*6]){ + if(col[8*6]){ a0 += + W6*col[8*6]; a1 += - W2*col[8*6]; a2 += + W2*col[8*6]; a3 += - W6*col[8*6]; - } + } - if (col[8*7]) { + if (col[8*7]) { MAC16(b0, + W7, col[8*7]); MAC16(b1, - W5, col[8*7]); MAC16(b2, + W3, col[8*7]); MAC16(b3, - W1, col[8*7]); - } + } dest[0] = cm[dest[0] + ((a0 + b0) >> COL_SHIFT)]; dest += line_size; @@ -302,58 +305,58 @@ static inline void idctSparseColAdd (UINT8 *dest, int line_size, dest[0] = cm[dest[0] + ((a0 - b0) >> COL_SHIFT)]; } -static inline void idctSparseCol (int16_t * col) +static inline void idctSparseCol (DCTELEM * col) { - int a0, a1, a2, a3, b0, b1, b2, b3; + int a0, a1, a2, a3, b0, b1, b2, b3; /* XXX: I did that only to give same values as previous code */ - a0 = W4 * (col[8*0] + ((1<<(COL_SHIFT-1))/W4)); - a1 = a0; - a2 = a0; - a3 = a0; + a0 = W4 * (col[8*0] + ((1<<(COL_SHIFT-1))/W4)); + a1 = a0; + a2 = a0; + a3 = a0; a0 += + W2*col[8*2]; a1 += + W6*col[8*2]; a2 += - W6*col[8*2]; a3 += - W2*col[8*2]; - MUL16(b0, W1, col[8*1]); - MUL16(b1, W3, col[8*1]); - MUL16(b2, W5, col[8*1]); - MUL16(b3, W7, col[8*1]); + b0 = MUL16(W1, col[8*1]); + b1 = MUL16(W3, col[8*1]); + b2 = MUL16(W5, col[8*1]); + b3 = MUL16(W7, col[8*1]); MAC16(b0, + W3, col[8*3]); MAC16(b1, - W7, col[8*3]); MAC16(b2, - W1, col[8*3]); MAC16(b3, - W5, col[8*3]); - if(col[8*4]){ + if(col[8*4]){ a0 += + W4*col[8*4]; a1 += - W4*col[8*4]; a2 += - W4*col[8*4]; a3 += + W4*col[8*4]; - } + } - if (col[8*5]) { + if (col[8*5]) { MAC16(b0, + W5, col[8*5]); MAC16(b1, - W1, col[8*5]); MAC16(b2, + W7, col[8*5]); MAC16(b3, + W3, col[8*5]); - } + } - if(col[8*6]){ + if(col[8*6]){ a0 += + W6*col[8*6]; a1 += - W2*col[8*6]; a2 += + W2*col[8*6]; a3 += - W6*col[8*6]; - } + } - if (col[8*7]) { + if (col[8*7]) { MAC16(b0, + W7, col[8*7]); MAC16(b1, - W5, col[8*7]); MAC16(b2, + W3, col[8*7]); MAC16(b3, - W1, col[8*7]); - } + } col[0 ] = ((a0 + b0) >> COL_SHIFT); col[8 ] = ((a1 + b1) >> COL_SHIFT); @@ -365,32 +368,32 @@ static inline void idctSparseCol (int16_t * col) col[56] = ((a0 - b0) >> COL_SHIFT); } -void simple_idct_put(UINT8 *dest, int line_size, INT16 *block) +void ff_simple_idct_put(uint8_t *dest, int line_size, DCTELEM *block) { int i; for(i=0; i<8; i++) idctRowCondDC(block + i*8); - + for(i=0; i<8; i++) idctSparseColPut(dest + i, line_size, block + i); } -void simple_idct_add(UINT8 *dest, int line_size, INT16 *block) +void ff_simple_idct_add(uint8_t *dest, int line_size, DCTELEM *block) { int i; for(i=0; i<8; i++) idctRowCondDC(block + i*8); - + for(i=0; i<8; i++) idctSparseColAdd(dest + i, line_size, block + i); } -void simple_idct(INT16 *block) +void ff_simple_idct(DCTELEM *block) { int i; for(i=0; i<8; i++) idctRowCondDC(block + i*8); - + for(i=0; i<8; i++) idctSparseCol(block + i); } @@ -406,10 +409,10 @@ void simple_idct(INT16 *block) and the butterfly must be multiplied by 0.5 * sqrt(2.0) */ #define C_SHIFT (4+1+12) -static inline void idct4col(UINT8 *dest, int line_size, const INT16 *col) +static inline void idct4col_put(uint8_t *dest, int line_size, const DCTELEM *col) { int c0, c1, c2, c3, a0, a1, a2, a3; - const UINT8 *cm = cropTbl + MAX_NEG_CROP; + const uint8_t *cm = ff_cropTbl + MAX_NEG_CROP; a0 = col[8*0]; a1 = col[8*2]; @@ -443,11 +446,11 @@ static inline void idct4col(UINT8 *dest, int line_size, const INT16 *col) /* XXX: I think a 1.0/sqrt(2) normalization should be needed to compensate the extra butterfly stage - I don't have the full DV specification */ -void simple_idct248_put(UINT8 *dest, int line_size, INT16 *block) +void ff_simple_idct248_put(uint8_t *dest, int line_size, DCTELEM *block) { int i; - INT16 *ptr; - + DCTELEM *ptr; + /* butterfly */ ptr = block; for(i=0;i<4;i++) { @@ -469,8 +472,8 @@ void simple_idct248_put(UINT8 *dest, int line_size, INT16 *block) /* IDCT4 and store */ for(i=0;i<8;i++) { - idct4col(dest + i, 2 * line_size, block + i); - idct4col(dest + line_size + i, 2 * line_size, block + 8 + i); + idct4col_put(dest + i, 2 * line_size, block + i); + idct4col_put(dest + line_size + i, 2 * line_size, block + 8 + i); } } @@ -486,10 +489,10 @@ void simple_idct248_put(UINT8 *dest, int line_size, INT16 *block) #define C2 C_FIX(0.2705980501) #define C3 C_FIX(0.5) #define C_SHIFT (4+1+12) -static inline void idct4col_add(UINT8 *dest, int line_size, const INT16 *col) +static inline void idct4col_add(uint8_t *dest, int line_size, const DCTELEM *col) { int c0, c1, c2, c3, a0, a1, a2, a3; - const UINT8 *cm = cropTbl + MAX_NEG_CROP; + const uint8_t *cm = ff_cropTbl + MAX_NEG_CROP; a0 = col[8*0]; a1 = col[8*1]; @@ -514,10 +517,10 @@ static inline void idct4col_add(UINT8 *dest, int line_size, const INT16 *col) #define R2 R_FIX(0.2705980501) #define R3 R_FIX(0.5) #define R_SHIFT 11 -static inline void idct4row(INT16 *row) +static inline void idct4row(DCTELEM *row) { int c0, c1, c2, c3, a0, a1, a2, a3; - const UINT8 *cm = cropTbl + MAX_NEG_CROP; + //const uint8_t *cm = ff_cropTbl + MAX_NEG_CROP; a0 = row[0]; a1 = row[1]; @@ -533,7 +536,7 @@ static inline void idct4row(INT16 *row) row[3]= (c0 - c1) >> R_SHIFT; } -void simple_idct84_add(UINT8 *dest, int line_size, INT16 *block) +void ff_simple_idct84_add(uint8_t *dest, int line_size, DCTELEM *block) { int i; @@ -548,7 +551,7 @@ void simple_idct84_add(UINT8 *dest, int line_size, INT16 *block) } } -void simple_idct48_add(UINT8 *dest, int line_size, INT16 *block) +void ff_simple_idct48_add(uint8_t *dest, int line_size, DCTELEM *block) { int i; @@ -563,3 +566,17 @@ void simple_idct48_add(UINT8 *dest, int line_size, INT16 *block) } } +void ff_simple_idct44_add(uint8_t *dest, int line_size, DCTELEM *block) +{ + int i; + + /* IDCT4 on each line */ + for(i=0; i<4; i++) { + idct4row(block + i*8); + } + + /* IDCT4 and store */ + for(i=0; i<4; i++){ + idct4col_add(dest + i, line_size, block + i); + } +}