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.
6123 Commits
32 Branches
944MB
Tree: d26a9ea0df
Branches Tags
${ item.name }
Create branch ${ searchTerm }
from 'd26a9ea0df'
${ noResults }
FFmpeg/libavcodec/vp3dsp.c

230 lines
7.0KB
Raw Blame History 

  1. /*

  2.  * Copyright (C) 2004 the ffmpeg project

  3.  *

  4.  * This library is free software; you can redistribute it and/or

  5.  * modify it under the terms of the GNU Lesser General Public

  6.  * License as published by the Free Software Foundation; either

  7.  * version 2 of the License, or (at your option) any later version.

  8.  *

  9.  * This library is distributed in the hope that it will be useful,

  10.  * but WITHOUT ANY WARRANTY; without even the implied warranty of

  11.  * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU

  12.  * Lesser General Public License for more details.

  13.  *

  14.  * You should have received a copy of the GNU Lesser General Public

  15.  * License along with this library; if not, write to the Free Software

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

  17.  */

  18. 

  19. /**

  20.  * @file vp3dsp.c

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

  22.  * source code.

  23.  */

  24. 

  25. #include "common.h"

  26. #include "avcodec.h"

  27. #include "dsputil.h"

  28. 

  29. #define IdctAdjustBeforeShift 8

  30. #define xC1S7 64277

  31. #define xC2S6 60547

  32. #define xC3S5 54491

  33. #define xC4S4 46341

  34. #define xC5S3 36410

  35. #define xC6S2 25080

  36. #define xC7S1 12785

  37. 

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

  39. 

  40. static always_inline void idct(uint8_t *dst, int stride, int16_t *input, int type)

  41. {

  42.     int16_t *ip = input;

  43.     uint8_t *cm = cropTbl + MAX_NEG_CROP;

  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]));

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

  117. 

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

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

  120. 

  121.             Ed = E - G;

  122.             Gd = E + G;

  123. 

  124.             Add = F + Ad;

  125.             Bdd = Bd - H;

  126. 

  127.             Fd = F - Ad;

  128.             Hd = Bd + H;

  129. 

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

  131.                 Gd += 16*128;

  132.                 Add+= 16*128;

  133.                 Ed += 16*128;

  134.                 Fd += 16*128;

  135.             }

  136.             Gd += IdctAdjustBeforeShift;

  137.             Add += IdctAdjustBeforeShift;

  138.             Ed += IdctAdjustBeforeShift;

  139.             Fd += IdctAdjustBeforeShift;

  140. 

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

  142.             if(type==0){

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

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

  145. 

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

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

  148. 

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

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

  151. 

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

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

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

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

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

  157. 

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

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

  160. 

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

  162.                 dst[4*stride] = cm[(Ed - Dd )  >> 4];

  163. 

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

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

  166.             }else{

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

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

  169. 

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

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

  172. 

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

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

  175. 

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

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

  178.             }

  179. 

  180.         } else {

  181.             if(type==0){

  182.                 ip[0*8] =

  183.                 ip[1*8] =

  184.                 ip[2*8] =

  185.                 ip[3*8] =

  186.                 ip[4*8] =

  187.                 ip[5*8] =

  188.                 ip[6*8] =

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

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

  191.                 dst[0*stride]=

  192.                 dst[1*stride]=

  193.                 dst[2*stride]=

  194.                 dst[3*stride]=

  195.                 dst[4*stride]=

  196.                 dst[5*stride]=

  197.                 dst[6*stride]=

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

  199.             }else{

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

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

  202.                     dst[0*stride] = cm[dst[0*stride] + v];

  203.                     dst[1*stride] = cm[dst[1*stride] + v];

  204.                     dst[2*stride] = cm[dst[2*stride] + v];

  205.                     dst[3*stride] = cm[dst[3*stride] + v];

  206.                     dst[4*stride] = cm[dst[4*stride] + v];

  207.                     dst[5*stride] = cm[dst[5*stride] + v];

  208.                     dst[6*stride] = cm[dst[6*stride] + v];

  209.                     dst[7*stride] = cm[dst[7*stride] + v];

  210.                 }

  211.             }

  212.         }

  213. 

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

  215.         dst++;

  216.     }

  217. }

  218. 

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

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

  221. }

  222. 

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

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

  225. }

  226. 

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

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

  229. }