| 1 | /* |
| 2 | * Lossless video DSP utils |
| 3 | * |
| 4 | * This file is part of FFmpeg. |
| 5 | * |
| 6 | * FFmpeg is free software; you can redistribute it and/or |
| 7 | * modify it under the terms of the GNU Lesser General Public |
| 8 | * License as published by the Free Software Foundation; either |
| 9 | * version 2.1 of the License, or (at your option) any later version. |
| 10 | * |
| 11 | * FFmpeg is distributed in the hope that it will be useful, |
| 12 | * but WITHOUT ANY WARRANTY; without even the implied warranty of |
| 13 | * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU |
| 14 | * Lesser General Public License for more details. |
| 15 | * |
| 16 | * You should have received a copy of the GNU Lesser General Public |
| 17 | * License along with FFmpeg; if not, write to the Free Software |
| 18 | * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA |
| 19 | */ |
| 20 | #include "avcodec.h" |
| 21 | #include "lossless_videodsp.h" |
| 22 | #include "libavcodec/mathops.h" |
| 23 | |
| 24 | static void add_int16_c(uint16_t *dst, const uint16_t *src, unsigned mask, int w){ |
| 25 | long i; |
| 26 | unsigned long pw_lsb = (mask >> 1) * 0x0001000100010001ULL; |
| 27 | unsigned long pw_msb = pw_lsb + 0x0001000100010001ULL; |
| 28 | for (i = 0; i <= w - (int)sizeof(long)/2; i += sizeof(long)/2) { |
| 29 | long a = *(long*)(src+i); |
| 30 | long b = *(long*)(dst+i); |
| 31 | *(long*)(dst+i) = ((a&pw_lsb) + (b&pw_lsb)) ^ ((a^b)&pw_msb); |
| 32 | } |
| 33 | for(; i<w; i++) |
| 34 | dst[i] = (dst[i] + src[i]) & mask; |
| 35 | } |
| 36 | |
| 37 | static void diff_int16_c(uint16_t *dst, const uint16_t *src1, const uint16_t *src2, unsigned mask, int w){ |
| 38 | long i; |
| 39 | #if !HAVE_FAST_UNALIGNED |
| 40 | if((long)src2 & (sizeof(long)-1)){ |
| 41 | for(i=0; i+3<w; i+=4){ |
| 42 | dst[i+0] = (src1[i+0]-src2[i+0]) & mask; |
| 43 | dst[i+1] = (src1[i+1]-src2[i+1]) & mask; |
| 44 | dst[i+2] = (src1[i+2]-src2[i+2]) & mask; |
| 45 | dst[i+3] = (src1[i+3]-src2[i+3]) & mask; |
| 46 | } |
| 47 | }else |
| 48 | #endif |
| 49 | { |
| 50 | unsigned long pw_lsb = (mask >> 1) * 0x0001000100010001ULL; |
| 51 | unsigned long pw_msb = pw_lsb + 0x0001000100010001ULL; |
| 52 | |
| 53 | for (i = 0; i <= w - (int)sizeof(long)/2; i += sizeof(long)/2) { |
| 54 | long a = *(long*)(src1+i); |
| 55 | long b = *(long*)(src2+i); |
| 56 | *(long*)(dst+i) = ((a|pw_msb) - (b&pw_lsb)) ^ ((a^b^pw_msb)&pw_msb); |
| 57 | } |
| 58 | } |
| 59 | for (; i<w; i++) |
| 60 | dst[i] = (src1[i] - src2[i]) & mask; |
| 61 | } |
| 62 | |
| 63 | static void add_hfyu_median_pred_int16_c(uint16_t *dst, const uint16_t *src, const uint16_t *diff, unsigned mask, int w, int *left, int *left_top){ |
| 64 | int i; |
| 65 | uint16_t l, lt; |
| 66 | |
| 67 | l = *left; |
| 68 | lt = *left_top; |
| 69 | |
| 70 | for(i=0; i<w; i++){ |
| 71 | l = (mid_pred(l, src[i], (l + src[i] - lt) & mask) + diff[i]) & mask; |
| 72 | lt = src[i]; |
| 73 | dst[i] = l; |
| 74 | } |
| 75 | |
| 76 | *left = l; |
| 77 | *left_top = lt; |
| 78 | } |
| 79 | |
| 80 | static void sub_hfyu_median_pred_int16_c(uint16_t *dst, const uint16_t *src1, const uint16_t *src2, unsigned mask, int w, int *left, int *left_top){ |
| 81 | int i; |
| 82 | uint16_t l, lt; |
| 83 | |
| 84 | l = *left; |
| 85 | lt = *left_top; |
| 86 | |
| 87 | for(i=0; i<w; i++){ |
| 88 | const int pred = mid_pred(l, src1[i], (l + src1[i] - lt) & mask); |
| 89 | lt = src1[i]; |
| 90 | l = src2[i]; |
| 91 | dst[i] = (l - pred) & mask; |
| 92 | } |
| 93 | |
| 94 | *left = l; |
| 95 | *left_top = lt; |
| 96 | } |
| 97 | |
| 98 | static int add_hfyu_left_pred_int16_c(uint16_t *dst, const uint16_t *src, unsigned mask, int w, unsigned acc){ |
| 99 | int i; |
| 100 | |
| 101 | for(i=0; i<w-1; i++){ |
| 102 | acc+= src[i]; |
| 103 | dst[i]= acc & mask; |
| 104 | i++; |
| 105 | acc+= src[i]; |
| 106 | dst[i]= acc & mask; |
| 107 | } |
| 108 | |
| 109 | for(; i<w; i++){ |
| 110 | acc+= src[i]; |
| 111 | dst[i]= acc & mask; |
| 112 | } |
| 113 | |
| 114 | return acc; |
| 115 | } |
| 116 | |
| 117 | |
| 118 | void ff_llviddsp_init(LLVidDSPContext *c, AVCodecContext *avctx) |
| 119 | { |
| 120 | c->add_int16 = add_int16_c; |
| 121 | c->diff_int16= diff_int16_c; |
| 122 | c->add_hfyu_left_pred_int16 = add_hfyu_left_pred_int16_c; |
| 123 | c->add_hfyu_median_pred_int16 = add_hfyu_median_pred_int16_c; |
| 124 | c->sub_hfyu_median_pred_int16 = sub_hfyu_median_pred_int16_c; |
| 125 | |
| 126 | if (ARCH_X86) |
| 127 | ff_llviddsp_init_x86(c, avctx); |
| 128 | } |