falkTX
/
FFmpeg
mirror of https://github.com/falkTX/FFmpeg.git
1
0
Fork 0
Code Issues 0 Releases 338 Wiki Activity
You can not select more than 25 topics Topics must start with a letter or number, can include dashes ('-') and can be up to 35 characters long.
34063 Commits
32 Branches
283MB
Tree: 25b51b2c44
Branches Tags
${ item.name }
Create branch ${ searchTerm }
from '25b51b2c44'
${ noResults }
FFmpeg/libavcodec/vp3dsp.c

271 lines
8.7KB
Raw Blame History 

  1. /*

  2.  * Copyright (C) 2004 the ffmpeg project

  3.  *

  4.  * This file is part of Libav.

  5.  *

  6.  * Libav 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.  * Libav 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 Libav; if not, write to the Free Software

  18.  * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA

  19.  */

  20. 

  21. /**

  22.  * @file

  23.  * Standard C DSP-oriented functions cribbed from the original VP3

  24.  * source code.

  25.  */

  26. 

  27. #include "avcodec.h"

  28. #include "dsputil.h"

  29. 

  30. #define IdctAdjustBeforeShift 8

  31. #define xC1S7 64277

  32. #define xC2S6 60547

  33. #define xC3S5 54491

  34. #define xC4S4 46341

  35. #define xC5S3 36410

  36. #define xC6S2 25080

  37. #define xC7S1 12785

  38. 

  39. #define M(a,b) (((a) * (b))>>16)

  40. 

  41. static av_always_inline void idct(uint8_t *dst, int stride, int16_t *input, int type)

  42. {

  43.     int16_t *ip = input;

  44. 

  45.     int A, B, C, D, Ad, Bd, Cd, Dd, E, F, G, H;

  46.     int Ed, Gd, Add, Bdd, Fd, Hd;

  47. 

  48.     int i;

  49. 

  50.     /* Inverse DCT on the rows now */

  51.     for (i = 0; i < 8; i++) {

  52.         /* Check for non-zero values */

  53.         if ( ip[0] | ip[1] | ip[2] | ip[3] | ip[4] | ip[5] | ip[6] | ip[7] ) {

  54.             A = M(xC1S7, ip[1]) + M(xC7S1, ip[7]);

  55.             B = M(xC7S1, ip[1]) - M(xC1S7, ip[7]);

  56.             C = M(xC3S5, ip[3]) + M(xC5S3, ip[5]);

  57.             D = M(xC3S5, ip[5]) - M(xC5S3, ip[3]);

  58. 

  59.             Ad = M(xC4S4, (A - C));

  60.             Bd = M(xC4S4, (B - D));

  61. 

  62.             Cd = A + C;

  63.             Dd = B + D;

  64. 

  65.             E = M(xC4S4, (ip[0] + ip[4]));

  66.             F = M(xC4S4, (ip[0] - ip[4]));

  67. 

  68.             G = M(xC2S6, ip[2]) + M(xC6S2, ip[6]);

  69.             H = M(xC6S2, ip[2]) - M(xC2S6, ip[6]);

  70. 

  71.             Ed = E - G;

  72.             Gd = E + G;

  73. 

  74.             Add = F + Ad;

  75.             Bdd = Bd - H;

  76. 

  77.             Fd = F - Ad;

  78.             Hd = Bd + H;

  79. 

  80.             /*  Final sequence of operations over-write original inputs. */

  81.             ip[0] = Gd + Cd ;

  82.             ip[7] = Gd - Cd ;

  83. 

  84.             ip[1] = Add + Hd;

  85.             ip[2] = Add - Hd;

  86. 

  87.             ip[3] = Ed + Dd ;

  88.             ip[4] = Ed - Dd ;

  89. 

  90.             ip[5] = Fd + Bdd;

  91.             ip[6] = Fd - Bdd;

  92.         }

  93. 

  94.         ip += 8;            /* next row */

  95.     }

  96. 

  97.     ip = input;

  98. 

  99.     for ( i = 0; i < 8; i++) {

  100.         /* Check for non-zero values (bitwise or faster than ||) */

  101.         if ( ip[1 * 8] | ip[2 * 8] | ip[3 * 8] |

  102.              ip[4 * 8] | ip[5 * 8] | ip[6 * 8] | ip[7 * 8] ) {

  103. 

  104.             A = M(xC1S7, ip[1*8]) + M(xC7S1, ip[7*8]);

  105.             B = M(xC7S1, ip[1*8]) - M(xC1S7, ip[7*8]);

  106.             C = M(xC3S5, ip[3*8]) + M(xC5S3, ip[5*8]);

  107.             D = M(xC3S5, ip[5*8]) - M(xC5S3, ip[3*8]);

  108. 

  109.             Ad = M(xC4S4, (A - C));

  110.             Bd = M(xC4S4, (B - D));

  111. 

  112.             Cd = A + C;

  113.             Dd = B + D;

  114. 

  115.             E = M(xC4S4, (ip[0*8] + ip[4*8])) + 8;

  116.             F = M(xC4S4, (ip[0*8] - ip[4*8])) + 8;

  117. 

  118.             if(type==1){  //HACK

  119.                 E += 16*128;

  120.                 F += 16*128;

  121.             }

  122. 

  123.             G = M(xC2S6, ip[2*8]) + M(xC6S2, ip[6*8]);

  124.             H = M(xC6S2, ip[2*8]) - M(xC2S6, ip[6*8]);

  125. 

  126.             Ed = E - G;

  127.             Gd = E + G;

  128. 

  129.             Add = F + Ad;

  130.             Bdd = Bd - H;

  131. 

  132.             Fd = F - Ad;

  133.             Hd = Bd + H;

  134. 

  135.             /* Final sequence of operations over-write original inputs. */

  136.             if(type==0){

  137.                 ip[0*8] = (Gd + Cd )  >> 4;

  138.                 ip[7*8] = (Gd - Cd )  >> 4;

  139. 

  140.                 ip[1*8] = (Add + Hd ) >> 4;

  141.                 ip[2*8] = (Add - Hd ) >> 4;

  142. 

  143.                 ip[3*8] = (Ed + Dd )  >> 4;

  144.                 ip[4*8] = (Ed - Dd )  >> 4;

  145. 

  146.                 ip[5*8] = (Fd + Bdd ) >> 4;

  147.                 ip[6*8] = (Fd - Bdd ) >> 4;

  148.             }else if(type==1){

  149.                 dst[0*stride] = av_clip_uint8((Gd + Cd )  >> 4);

  150.                 dst[7*stride] = av_clip_uint8((Gd - Cd )  >> 4);

  151. 

  152.                 dst[1*stride] = av_clip_uint8((Add + Hd ) >> 4);

  153.                 dst[2*stride] = av_clip_uint8((Add - Hd ) >> 4);

  154. 

  155.                 dst[3*stride] = av_clip_uint8((Ed + Dd )  >> 4);

  156.                 dst[4*stride] = av_clip_uint8((Ed - Dd )  >> 4);

  157. 

  158.                 dst[5*stride] = av_clip_uint8((Fd + Bdd ) >> 4);

  159.                 dst[6*stride] = av_clip_uint8((Fd - Bdd ) >> 4);

  160.             }else{

  161.                 dst[0*stride] = av_clip_uint8(dst[0*stride] + ((Gd + Cd )  >> 4));

  162.                 dst[7*stride] = av_clip_uint8(dst[7*stride] + ((Gd - Cd )  >> 4));

  163. 

  164.                 dst[1*stride] = av_clip_uint8(dst[1*stride] + ((Add + Hd ) >> 4));

  165.                 dst[2*stride] = av_clip_uint8(dst[2*stride] + ((Add - Hd ) >> 4));

  166. 

  167.                 dst[3*stride] = av_clip_uint8(dst[3*stride] + ((Ed + Dd )  >> 4));

  168.                 dst[4*stride] = av_clip_uint8(dst[4*stride] + ((Ed - Dd )  >> 4));

  169. 

  170.                 dst[5*stride] = av_clip_uint8(dst[5*stride] + ((Fd + Bdd ) >> 4));

  171.                 dst[6*stride] = av_clip_uint8(dst[6*stride] + ((Fd - Bdd ) >> 4));

  172.             }

  173. 

  174.         } else {

  175.             if(type==0){

  176.                 ip[0*8] =

  177.                 ip[1*8] =

  178.                 ip[2*8] =

  179.                 ip[3*8] =

  180.                 ip[4*8] =

  181.                 ip[5*8] =

  182.                 ip[6*8] =

  183.                 ip[7*8] = ((xC4S4 * ip[0*8] + (IdctAdjustBeforeShift<<16))>>20);

  184.             }else if(type==1){

  185.                 dst[0*stride]=

  186.                 dst[1*stride]=

  187.                 dst[2*stride]=

  188.                 dst[3*stride]=

  189.                 dst[4*stride]=

  190.                 dst[5*stride]=

  191.                 dst[6*stride]=

  192.                 dst[7*stride]= av_clip_uint8(128 + ((xC4S4 * ip[0*8] + (IdctAdjustBeforeShift<<16))>>20));

  193.             }else{

  194.                 if(ip[0*8]){

  195.                     int v= ((xC4S4 * ip[0*8] + (IdctAdjustBeforeShift<<16))>>20);

  196.                     dst[0*stride] = av_clip_uint8(dst[0*stride] + v);

  197.                     dst[1*stride] = av_clip_uint8(dst[1*stride] + v);

  198.                     dst[2*stride] = av_clip_uint8(dst[2*stride] + v);

  199.                     dst[3*stride] = av_clip_uint8(dst[3*stride] + v);

  200.                     dst[4*stride] = av_clip_uint8(dst[4*stride] + v);

  201.                     dst[5*stride] = av_clip_uint8(dst[5*stride] + v);

  202.                     dst[6*stride] = av_clip_uint8(dst[6*stride] + v);

  203.                     dst[7*stride] = av_clip_uint8(dst[7*stride] + v);

  204.                 }

  205.             }

  206.         }

  207. 

  208.         ip++;            /* next column */

  209.         dst++;

  210.     }

  211. }

  212. 

  213. void ff_vp3_idct_c(DCTELEM *block/* align 16*/){

  214.     idct(NULL, 0, block, 0);

  215. }

  216. 

  217. void ff_vp3_idct_put_c(uint8_t *dest/*align 8*/, int line_size, DCTELEM *block/*align 16*/){

  218.     idct(dest, line_size, block, 1);

  219. }

  220. 

  221. void ff_vp3_idct_add_c(uint8_t *dest/*align 8*/, int line_size, DCTELEM *block/*align 16*/){

  222.     idct(dest, line_size, block, 2);

  223. }

  224. 

  225. void ff_vp3_idct_dc_add_c(uint8_t *dest/*align 8*/, int line_size, const DCTELEM *block/*align 16*/){

  226.     int i, dc = (block[0] + 15) >> 5;

  227. 

  228.     for(i = 0; i < 8; i++){

  229.         dest[0] = av_clip_uint8(dest[0] + dc);

  230.         dest[1] = av_clip_uint8(dest[1] + dc);

  231.         dest[2] = av_clip_uint8(dest[2] + dc);

  232.         dest[3] = av_clip_uint8(dest[3] + dc);

  233.         dest[4] = av_clip_uint8(dest[4] + dc);

  234.         dest[5] = av_clip_uint8(dest[5] + dc);

  235.         dest[6] = av_clip_uint8(dest[6] + dc);

  236.         dest[7] = av_clip_uint8(dest[7] + dc);

  237.         dest += line_size;

  238.     }

  239. }

  240. 

  241. void ff_vp3_v_loop_filter_c(uint8_t *first_pixel, int stride, int *bounding_values)

  242. {

  243.     unsigned char *end;

  244.     int filter_value;

  245.     const int nstride= -stride;

  246. 

  247.     for (end= first_pixel + 8; first_pixel < end; first_pixel++) {

  248.         filter_value =

  249.             (first_pixel[2 * nstride] - first_pixel[ stride])

  250.          +3*(first_pixel[0          ] - first_pixel[nstride]);

  251.         filter_value = bounding_values[(filter_value + 4) >> 3];

  252.         first_pixel[nstride] = av_clip_uint8(first_pixel[nstride] + filter_value);

  253.         first_pixel[0] = av_clip_uint8(first_pixel[0] - filter_value);

  254.     }

  255. }

  256. 

  257. void ff_vp3_h_loop_filter_c(uint8_t *first_pixel, int stride, int *bounding_values)

  258. {

  259.     unsigned char *end;

  260.     int filter_value;

  261. 

  262.     for (end= first_pixel + 8*stride; first_pixel != end; first_pixel += stride) {

  263.         filter_value =

  264.             (first_pixel[-2] - first_pixel[ 1])

  265.          +3*(first_pixel[ 0] - first_pixel[-1]);

  266.         filter_value = bounding_values[(filter_value + 4) >> 3];

  267.         first_pixel[-1] = av_clip_uint8(first_pixel[-1] + filter_value);

  268.         first_pixel[ 0] = av_clip_uint8(first_pixel[ 0] - filter_value);

  269.     }

  270. }