[cvs] / xvidcore / src / image / interpolate8x8.c Repository:
ViewVC logotype

Diff of /xvidcore/src/image/interpolate8x8.c

Parent Directory Parent Directory | Revision Log Revision Log | View Patch Patch

revision 1.4.2.4, Sun Nov 3 02:04:07 2002 UTC revision 1.6, Sun Nov 17 00:20:30 2002 UTC
# Line 1  Line 1 
1  /**************************************************************************  /*****************************************************************************
2   *   *
3   *      XVID MPEG-4 VIDEO CODEC   *      XVID MPEG-4 VIDEO CODEC
4   *      8x8 block-based halfpel interpolation   *  - 8x8 block-based halfpel interpolation -
5   *   *
6   *      This program is free software; you can redistribute it and/or modify   *  Copyright(C) 2002 Peter Ross <pross@xvid.org>
7   *      it under the terms of the GNU General Public License as published by   *  Copyright(C) 2002 MinChen <chenm002@163.com>
8     *
9     *  This file is part of XviD, a free MPEG-4 video encoder/decoder
10     *
11     *  XviD is free software; you can redistribute it and/or modify it
12     *  under the terms of the GNU General Public License as published by
13   *      the Free Software Foundation; either version 2 of the License, or   *      the Free Software Foundation; either version 2 of the License, or
14   *      (at your option) any later version.   *      (at your option) any later version.
15   *   *
# Line 15  Line 20 
20   *   *
21   *      You should have received a copy of the GNU General Public License   *      You should have received a copy of the GNU General Public License
22   *      along with this program; if not, write to the Free Software   *      along with this program; if not, write to the Free Software
23   *      Foundation, Inc., 675 Mass Ave, Cambridge, MA 02139, USA.   *  Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA  02111-1307 USA
24   *   *
25   *************************************************************************/   *  Under section 8 of the GNU General Public License, the copyright
26     *  holders of XVID explicitly forbid distribution in the following
27  /**************************************************************************   *  countries:
28   *   *
29   *      History:   *    - Japan
30     *    - United States of America
31   *   *
32   *  05.10.2002  new bilinear and qpel interpolation code - Isibaar   *  Linking XviD statically or dynamically with other modules is making a
33   *      27.12.2001      modified "compensate_halfpel"   *  combined work based on XviD.  Thus, the terms and conditions of the
34   *      05.11.2001      initial version; (c)2001 peter ross <pross@cs.rmit.edu.au>   *  GNU General Public License cover the whole combination.
35   *   *
36   *************************************************************************/   *  As a special exception, the copyright holders of XviD give you
37     *  permission to link XviD with independent modules that communicate with
38     *  XviD solely through the VFW1.1 and DShow interfaces, regardless of the
39     *  license terms of these independent modules, and to copy and distribute
40     *  the resulting combined work under terms of your choice, provided that
41     *  every copy of the combined work is accompanied by a complete copy of
42     *  the source code of XviD (the version of XviD used to produce the
43     *  combined work), being distributed under the terms of the GNU General
44     *  Public License plus this exception.  An independent module is a module
45     *  which is not derived from or based on XviD.
46     *
47     *  Note that people who make modified versions of XviD are not obligated
48     *  to grant this special exception for their modified versions; it is
49     *  their choice whether to do so.  The GNU General Public License gives
50     *  permission to release a modified version without this exception; this
51     *  exception also makes it possible to release a modified version which
52     *  carries forward this exception.
53     *
54     * $Id$
55     *
56     ****************************************************************************/
57    
58  #include "../portab.h"  #include "../portab.h"
59  #include "interpolate8x8.h"  #include "interpolate8x8.h"
# Line 38  Line 63 
63  INTERPOLATE8X8_PTR interpolate8x8_halfpel_v;  INTERPOLATE8X8_PTR interpolate8x8_halfpel_v;
64  INTERPOLATE8X8_PTR interpolate8x8_halfpel_hv;  INTERPOLATE8X8_PTR interpolate8x8_halfpel_hv;
65    
 INTERPOLATE8X8_AVG2_PTR interpolate8x8_avg2;  
 INTERPOLATE8X8_AVG4_PTR interpolate8x8_avg4;  
   
 INTERPOLATE_LOWPASS_PTR interpolate8x8_lowpass_h;  
 INTERPOLATE_LOWPASS_PTR interpolate8x8_lowpass_v;  
   
 INTERPOLATE_LOWPASS_PTR interpolate16x16_lowpass_h;  
 INTERPOLATE_LOWPASS_PTR interpolate16x16_lowpass_v;  
   
 INTERPOLATE_LOWPASS_HV_PTR interpolate8x8_lowpass_hv;  
 INTERPOLATE_LOWPASS_HV_PTR interpolate16x16_lowpass_hv;  
   
 INTERPOLATE8X8_6TAP_LOWPASS_PTR interpolate8x8_6tap_lowpass_h;  
 INTERPOLATE8X8_6TAP_LOWPASS_PTR interpolate8x8_6tap_lowpass_v;  
   
 void interpolate8x8_avg2_c(uint8_t *dst, const uint8_t *src1, const uint8_t *src2, const uint32_t stride, const uint32_t rounding)  
 {  
     int32_t i;  
         const int32_t round = 1 - rounding;  
   
     for(i = 0; i < 9; i++)  
     {  
         dst[0] = (src1[0] + src2[0] + round) >> 1;  
         dst[1] = (src1[1] + src2[1] + round) >> 1;  
         dst[2] = (src1[2] + src2[2] + round) >> 1;  
         dst[3] = (src1[3] + src2[3] + round) >> 1;  
         dst[4] = (src1[4] + src2[4] + round) >> 1;  
         dst[5] = (src1[5] + src2[5] + round) >> 1;  
         dst[6] = (src1[6] + src2[6] + round) >> 1;  
         dst[7] = (src1[7] + src2[7] + round) >> 1;  
   
         dst += stride;  
         src1 += stride;  
         src2 += stride;  
     }  
 }  
   
 void interpolate8x8_avg4_c(uint8_t *dst, const uint8_t *src1, const uint8_t *src2, const uint8_t *src3, const uint8_t *src4, const uint32_t stride, const uint32_t rounding)  
 {  
     int32_t i;  
         const int32_t round = 2 - rounding;  
   
     for(i = 0; i < 8; i++)  
     {  
         dst[0] = (src1[0] + src2[0] + src3[0] + src4[0] + round) >> 2;  
         dst[1] = (src1[1] + src2[1] + src3[1] + src4[1] + round) >> 2;  
         dst[2] = (src1[2] + src2[2] + src3[2] + src4[2] + round) >> 2;  
         dst[3] = (src1[3] + src2[3] + src3[3] + src4[3] + round) >> 2;  
         dst[4] = (src1[4] + src2[4] + src3[4] + src4[4] + round) >> 2;  
         dst[5] = (src1[5] + src2[5] + src3[5] + src4[5] + round) >> 2;  
         dst[6] = (src1[6] + src2[6] + src3[6] + src4[6] + round) >> 2;  
         dst[7] = (src1[7] + src2[7] + src3[7] + src4[7] + round) >> 2;  
   
                 dst += stride;  
         src1 += stride;  
         src2 += stride;  
         src3 += stride;  
         src4 += stride;  
     }  
 }  
66    
67  // dst = interpolate(src)  // dst = interpolate(src)
68    
# Line 163  Line 128 
128          }          }
129  }  }
130    
131  /*************************************************************  // add by MinChen <chenm001@163.com>
132   * QPEL STUFF STARTS HERE                                    *  // interpolate8x8 two pred block
133   *************************************************************/  void
134    interpolate8x8_c(uint8_t * const dst,
135  #define CLIP(X,A,B) (X < A) ? (A) : ((X > B) ? (B) : (X))                                   const uint8_t * const src,
136                                     const uint32_t x,
137  void interpolate8x8_6tap_lowpass_h_c(uint8_t *dst, uint8_t *src, int32_t stride, int32_t rounding)                                   const uint32_t y,
138  {                                   const uint32_t stride)
     int32_t i;  
         uint8_t round_add = 16 - rounding;  
   
     for(i = 0; i < 8; i++)  
139      {      {
140            uint32_t i, j;
141    
142          dst[0] = CLIP((((src[-2] + src[3]) + 5 * (((src[0] + src[1])<<2) - (src[-1] + src[2])) + round_add) >> 5), 0, 255);          for (j = 0; j < 8; j++) {
143          dst[1] = CLIP((((src[-1] + src[4]) + 5 * (((src[1] + src[2])<<2) - (src[0] + src[3])) + round_add) >> 5), 0, 255);                  for (i = 0; i < 8; i++) {
144          dst[2] = CLIP((((src[0] + src[5]) + 5 * (((src[2] + src[3])<<2) - (src[1] + src[4])) + round_add) >> 5), 0, 255);                          int32_t tot =
145          dst[3] = CLIP((((src[1] + src[6]) + 5 * (((src[3] + src[4])<<2) - (src[2] + src[5])) + round_add) >> 5), 0, 255);                                  ((src[(y + j) * stride + x + i] +
146          dst[4] = CLIP((((src[2] + src[7]) + 5 * (((src[4] + src[5])<<2) - (src[3] + src[6])) + round_add) >> 5), 0, 255);                                    dst[(y + j) * stride + x + i] + 1) >> 1);
147          dst[5] = CLIP((((src[3] + src[8]) + 5 * (((src[5] + src[6])<<2) - (src[4] + src[7])) + round_add) >> 5), 0, 255);                          dst[(y + j) * stride + x + i] = (uint8_t) tot;
148          dst[6] = CLIP((((src[4] + src[9]) + 5 * (((src[6] + src[7])<<2) - (src[5] + src[8])) + round_add) >> 5), 0, 255);                  }
         dst[7] = CLIP((((src[5] + src[10]) + 5 * (((src[7] + src[8])<<2) - (src[6] + src[9])) + round_add) >> 5), 0, 255);  
   
         dst += stride;  
         src += stride;  
149      }      }
150  }  }
151    
152  void interpolate16x16_lowpass_h_c(uint8_t *dst, uint8_t *src, int32_t stride, int32_t rounding)  /*************************************************************
153  {   * QPEL STUFF STARTS HERE                                    *
154      int32_t i;   *************************************************************/
         uint8_t round_add = 16 - rounding;  
   
     for(i = 0; i < 17; i++)  
     {  
   
         dst[0] = CLIP(((7 * ((src[0]<<1) - src[2]) +  23 * src[1] + 3 * src[3] - src[4] + round_add) >> 5), 0, 255);  
         dst[1] = CLIP(((19 * src[1] + 20 * src[2] - src[5] + 3 * (src[4] - src[0] - (src[3]<<1)) + round_add) >> 5), 0, 255);  
         dst[2] = CLIP(((20 * (src[2] + src[3]) + (src[0]<<1) + 3 * (src[5] - ((src[1] + src[4])<<1)) - src[6] + round_add) >> 5), 0, 255);  
   
         dst[3] = CLIP(((20 * (src[3] + src[4]) + 3 * ((src[6] + src[1]) - ((src[2] + src[5])<<1)) - (src[0] + src[7]) + round_add) >> 5), 0, 255);  
         dst[4] = CLIP(((20 * (src[4] + src[5]) - 3 * (((src[3] + src[6])<<1) - (src[2] + src[7])) - (src[1] + src[8]) + round_add) >> 5), 0, 255);  
         dst[5] = CLIP(((20 * (src[5] + src[6]) - 3 * (((src[4] + src[7])<<1) - (src[3] + src[8])) - (src[2] + src[9]) + round_add) >> 5), 0, 255);  
         dst[6] = CLIP(((20 * (src[6] + src[7]) - 3 * (((src[5] + src[8])<<1) - (src[4] + src[9])) - (src[3] + src[10]) + round_add) >> 5), 0, 255);  
         dst[7] = CLIP(((20 * (src[7] + src[8]) - 3 * (((src[6] + src[9])<<1) - (src[5] + src[10])) - (src[4] + src[11]) + round_add) >> 5), 0, 255);  
         dst[8] = CLIP(((20 * (src[8] + src[9]) - 3 * (((src[7] + src[10])<<1) - (src[6] + src[11])) - (src[5] + src[12]) + round_add) >> 5), 0, 255);  
         dst[9] = CLIP(((20 * (src[9] + src[10]) - 3 * (((src[8] + src[11])<<1) - (src[7] + src[12])) - (src[6] + src[13]) + round_add) >> 5), 0, 255);  
         dst[10] = CLIP(((20 * (src[10] + src[11]) - 3 * (((src[9] + src[12])<<1) - (src[8] + src[13])) - (src[7] + src[14]) + round_add) >> 5), 0, 255);  
         dst[11] = CLIP(((20 * (src[11] + src[12]) - 3 * (((src[10] + src[13])<<1) - (src[9] + src[14])) - (src[8] + src[15]) + round_add) >> 5), 0, 255);  
         dst[12] = CLIP(((20 * (src[12] + src[13]) - 3 * (((src[11] + src[14])<<1) - (src[10] + src[15])) - (src[9] + src[16]) + round_add) >> 5), 0, 255);  
   
         dst[13] = CLIP(((20 * (src[13] + src[14]) + (src[16]<<1) + 3 * (src[11] - ((src[12] + src[15]) << 1)) - src[10] + round_add) >> 5), 0, 255);  
         dst[14] = CLIP(((19 * src[15] + 20 * src[14] + 3 * (src[12] - src[16] - (src[13] << 1)) - src[11] + round_add) >> 5), 0, 255);  
         dst[15] = CLIP(((23 * src[15] + 7 * ((src[16]<<1) - src[14]) + 3 * src[13] - src[12] + round_add) >> 5), 0, 255);  
155    
156          dst += stride;  #define CLIP(X,A,B) (X < A) ? (A) : ((X > B) ? (B) : (X))
         src += stride;  
     }  
 }  
157    
158  void interpolate8x8_lowpass_h_c(uint8_t *dst, uint8_t *src, int32_t stride, int32_t rounding)  void interpolate8x8_lowpass_h(uint8_t *dst, uint8_t *src, int32_t dst_stride, int32_t src_stride, int32_t rounding)
159  {  {
160      int32_t i;      int32_t i;
         uint8_t round_add = 16 - rounding;  
161    
162      for(i = 0; i < 9; i++)      for(i = 0; i < 8; i++)
163      {      {
164            dst[0] = CLIP((((src[0] + src[1]) * 160 - (src[0] + src[2]) * 48 + (src[1] + src[3]) * 24 - (src[2] + src[4]) * 8 + (128 - rounding)) / 256), 0, 255);
165            dst[1] = CLIP((((src[1] + src[2]) * 160 - (src[0] + src[3]) * 48 + (src[0] + src[4]) * 24 - (src[1] + src[5]) * 8 + (128 - rounding)) / 256), 0, 255);
166            dst[2] = CLIP((((src[2] + src[3]) * 160 - (src[1] + src[4]) * 48 + (src[0] + src[5]) * 24 - (src[0] + src[6]) * 8 + (128 - rounding)) / 256), 0, 255);
167            dst[3] = CLIP((((src[3] + src[4]) * 160 - (src[2] + src[5]) * 48 + (src[1] + src[6]) * 24 - (src[0] + src[7]) * 8 + (128 - rounding)) / 256), 0, 255);
168            dst[4] = CLIP((((src[4] + src[5]) * 160 - (src[3] + src[6]) * 48 + (src[2] + src[7]) * 24 - (src[1] + src[8]) * 8 + (128 - rounding)) / 256), 0, 255);
169            dst[5] = CLIP((((src[5] + src[6]) * 160 - (src[4] + src[7]) * 48 + (src[3] + src[8]) * 24 - (src[2] + src[8]) * 8 + (128 - rounding)) / 256), 0, 255);
170            dst[6] = CLIP((((src[6] + src[7]) * 160 - (src[5] + src[8]) * 48 + (src[4] + src[8]) * 24 - (src[3] + src[7]) * 8 + (128 - rounding)) / 256), 0, 255);
171            dst[7] = CLIP((((src[7] + src[8]) * 160 - (src[6] + src[8]) * 48 + (src[5] + src[7]) * 24 - (src[4] + src[6]) * 8 + (128 - rounding)) / 256), 0, 255);
172    
173          dst[0] = CLIP(((7 * ((src[0]<<1) - src[2]) + 23 * src[1] + 3 * src[3] - src[4] + round_add) >> 5), 0, 255);          dst += dst_stride;
174          dst[1] = CLIP(((19 * src[1] + 20 * src[2] - src[5] + 3 * (src[4] - src[0] - (src[3]<<1)) + round_add) >> 5), 0, 255);          src += src_stride;
         dst[2] = CLIP(((20 * (src[2] + src[3]) + (src[0]<<1) + 3 * (src[5] - ((src[1] + src[4])<<1)) - src[6] + round_add) >> 5), 0, 255);  
         dst[3] = CLIP(((20 * (src[3] + src[4]) + 3 * ((src[6] + src[1]) - ((src[2] + src[5])<<1)) - (src[0] + src[7]) + round_add) >> 5), 0, 255);  
         dst[4] = CLIP(((20 * (src[4] + src[5]) - 3 * (((src[3] + src[6])<<1) - (src[2] + src[7])) - (src[1] + src[8]) + round_add) >> 5), 0, 255);  
         dst[5] = CLIP(((20 * (src[5] + src[6]) + (src[8]<<1) + 3 * (src[3] - ((src[4] + src[7]) << 1)) - src[2] + round_add) >> 5), 0, 255);  
         dst[6] = CLIP(((19 * src[7] + 20 * src[6] + 3 * (src[4] - src[8] - (src[5] << 1)) - src[3] + round_add) >> 5), 0, 255);  
         dst[7] = CLIP(((23 * src[7] + 7 * ((src[8]<<1) - src[6]) + 3 * src[5] - src[4] + round_add) >> 5), 0, 255);  
   
         dst += stride;  
         src += stride;  
175      }      }
176  }  }
177    
178  void interpolate8x8_6tap_lowpass_v_c(uint8_t *dst, uint8_t *src, int32_t stride, int32_t rounding)  void interpolate8x8_lowpass_v(uint8_t *dst, uint8_t *src, int32_t dst_stride, int32_t src_stride, int32_t rounding)
179  {  {
180      int32_t i;      int32_t i;
         uint8_t round_add = 16 - rounding;  
181    
182      for(i = 0; i < 8; i++)      for(i = 0; i < 8; i++)
183      {      {
         int32_t src_2 = src[-2*stride];  
         int32_t src_1 = src[-stride];  
184          int32_t src0 = src[0];          int32_t src0 = src[0];
185          int32_t src1 = src[stride];          int32_t src1 = src[src_stride];
186          int32_t src2 = src[2 * stride];          int32_t src2 = src[2 * src_stride];
187          int32_t src3 = src[3 * stride];          int32_t src3 = src[3 * src_stride];
188          int32_t src4 = src[4 * stride];          int32_t src4 = src[4 * src_stride];
189          int32_t src5 = src[5 * stride];          int32_t src5 = src[5 * src_stride];
190          int32_t src6 = src[6 * stride];          int32_t src6 = src[6 * src_stride];
191          int32_t src7 = src[7 * stride];          int32_t src7 = src[7 * src_stride];
192          int32_t src8 = src[8 * stride];          int32_t src8 = src[8 * src_stride];
193          int32_t src9 = src[9 * stride];  
194          int32_t src10 = src[10 * stride];                  dst[0] = CLIP((((src0 + src1) * 160 - (src0 + src2) * 48 + (src1 + src3) * 24 - (src2 + src4) * 8 + (128 - rounding)) / 256), 0, 255);
195            dst[dst_stride] = CLIP((((src1 + src2) * 160 - (src0 + src3) * 48 + (src0 + src4) * 24 - (src1 + src5) * 8 + (128 - rounding)) / 256), 0, 255);
196          dst[0]                  = CLIP((((src_2 + src3) + 5 * (((src0 + src1)<<2) - (src_1 + src2)) + round_add) >> 5), 0, 255);          dst[2 * dst_stride] = CLIP((((src2 + src3) * 160 - (src1 + src4) * 48 + (src0 + src5) * 24 - (src0 + src6) * 8 + (128 - rounding)) / 256), 0, 255);
197          dst[stride]             = CLIP((((src_1 + src4) + 5 * (((src1 + src2)<<2) - (src0 + src3)) + round_add) >> 5), 0, 255);          dst[3 * dst_stride] = CLIP((((src3 + src4) * 160 - (src2 + src5) * 48 + (src1 + src6) * 24 - (src0 + src7) * 8 + (128 - rounding)) / 256), 0, 255);
198          dst[2 * stride] = CLIP((((src0 + src5) + 5 * (((src2 + src3)<<2) - (src1 + src4)) + round_add) >> 5), 0, 255);          dst[4 * dst_stride] = CLIP((((src4 + src5) * 160 - (src3 + src6) * 48 + (src2 + src7) * 24 - (src1 + src8) * 8 + (128 - rounding)) / 256), 0, 255);
199          dst[3 * stride] = CLIP((((src1 + src6) + 5 * (((src3 + src4)<<2) - (src2 + src5)) + round_add) >> 5), 0, 255);          dst[5 * dst_stride] = CLIP((((src5 + src6) * 160 - (src4 + src7) * 48 + (src3 + src8) * 24 - (src2 + src8) * 8 + (128 - rounding)) / 256), 0, 255);
200          dst[4 * stride] = CLIP((((src2 + src7) + 5 * (((src4 + src5)<<2) - (src3 + src6)) + round_add) >> 5), 0, 255);          dst[6 * dst_stride] = CLIP((((src6 + src7) * 160 - (src5 + src8) * 48 + (src4 + src8) * 24 - (src3 + src7) * 8 + (128 - rounding)) / 256), 0, 255);
201          dst[5 * stride] = CLIP((((src3 + src8) + 5 * (((src5 + src6)<<2) - (src4 + src7)) + round_add) >> 5), 0, 255);          dst[7 * dst_stride] = CLIP((((src7 + src8) * 160 - (src6 + src8) * 48 + (src5 + src7) * 24 - (src4 + src6) * 8 + (128 - rounding)) / 256), 0, 255);
         dst[6 * stride] = CLIP((((src4 + src9) + 5 * (((src6 + src7)<<2) - (src5 + src8)) + round_add) >> 5), 0, 255);  
         dst[7 * stride] = CLIP((((src5 + src10) + 5 * (((src7 + src8)<<2) - (src6 + src9)) + round_add) >> 5), 0, 255);  
202    
203                  dst++;                  dst++;
204          src++;          src++;
205      }      }
206  }  }
207    
208  void interpolate16x16_lowpass_v_c(uint8_t *dst, uint8_t *src, int32_t stride, int32_t rounding)  void interpolate8x8_lowpass_hv(uint8_t *dst1, uint8_t *dst2, uint8_t *src, int32_t dst1_stride, int32_t dst2_stride, int32_t src_stride, int32_t rounding)
209  {  {
210        uint8_t data[72];
211    
212      int32_t i;      int32_t i;
         uint8_t round_add = 16 - rounding;  
213    
214      for(i = 0; i < 17; i++)  
215        for(i = 0; i < 9; i++)
216      {      {
217          int32_t src0 = src[0];          dst2[0] = data[8 * i + 0] = CLIP((((src[0] + src[1]) * 160 - (src[0] + src[2]) * 48 + (src[1] + src[3]) * 24 - (src[2] + src[4]) * 8 + (128 - rounding)) / 256), 0, 255);
218          int32_t src1 = src[stride];          dst2[1] = data[8 * i + 1] = CLIP((((src[1] + src[2]) * 160 - (src[0] + src[3]) * 48 + (src[0] + src[4]) * 24 - (src[1] + src[5]) * 8 + (128 - rounding)) / 256), 0, 255);
219          int32_t src2 = src[2 * stride];          dst2[2] = data[8 * i + 2] = CLIP((((src[2] + src[3]) * 160 - (src[1] + src[4]) * 48 + (src[0] + src[5]) * 24 - (src[0] + src[6]) * 8 + (128 - rounding)) / 256), 0, 255);
220          int32_t src3 = src[3 * stride];          dst2[3] = data[8 * i + 3] = CLIP((((src[3] + src[4]) * 160 - (src[2] + src[5]) * 48 + (src[1] + src[6]) * 24 - (src[0] + src[7]) * 8 + (128 - rounding)) / 256), 0, 255);
221          int32_t src4 = src[4 * stride];          dst2[4] = data[8 * i + 4] = CLIP((((src[4] + src[5]) * 160 - (src[3] + src[6]) * 48 + (src[2] + src[7]) * 24 - (src[1] + src[8]) * 8 + (128 - rounding)) / 256), 0, 255);
222          int32_t src5 = src[5 * stride];          dst2[5] = data[8 * i + 5] = CLIP((((src[5] + src[6]) * 160 - (src[4] + src[7]) * 48 + (src[3] + src[8]) * 24 - (src[2] + src[8]) * 8 + (128 - rounding)) / 256), 0, 255);
223          int32_t src6 = src[6 * stride];          dst2[6] = data[8 * i + 6] = CLIP((((src[6] + src[7]) * 160 - (src[5] + src[8]) * 48 + (src[4] + src[8]) * 24 - (src[3] + src[7]) * 8 + (128 - rounding)) / 256), 0, 255);
224          int32_t src7 = src[7 * stride];          dst2[7] = data[8 * i + 7] = CLIP((((src[7] + src[8]) * 160 - (src[6] + src[8]) * 48 + (src[5] + src[7]) * 24 - (src[4] + src[6]) * 8 + (128 - rounding)) / 256), 0, 255);
         int32_t src8 = src[8 * stride];  
         int32_t src9 = src[9 * stride];  
         int32_t src10 = src[10 * stride];  
         int32_t src11 = src[11 * stride];  
         int32_t src12 = src[12 * stride];  
         int32_t src13 = src[13 * stride];  
         int32_t src14 = src[14 * stride];  
         int32_t src15 = src[15 * stride];  
         int32_t src16 = src[16 * stride];  
   
   
         dst[0] = CLIP(((7 * ((src0<<1) - src2) +  23 * src1 + 3 * src3 - src4 + round_add) >> 5), 0, 255);  
         dst[stride] = CLIP(((19 * src1 + 20 * src2 - src5 + 3 * (src4 - src0 - (src3<<1)) + round_add) >> 5), 0, 255);  
         dst[2*stride] = CLIP(((20 * (src2 + src3) + (src0<<1) + 3 * (src5 - ((src1 + src4)<<1)) - src6 + round_add) >> 5), 0, 255);  
   
         dst[3*stride] = CLIP(((20 * (src3 + src4) + 3 * ((src6 + src1) - ((src2 + src5)<<1)) - (src0 + src7) + round_add) >> 5), 0, 255);  
         dst[4*stride] = CLIP(((20 * (src4 + src5) - 3 * (((src3 + src6)<<1) - (src2 + src7)) - (src1 + src8) + round_add) >> 5), 0, 255);  
         dst[5*stride] = CLIP(((20 * (src5 + src6) - 3 * (((src4 + src7)<<1) - (src3 + src8)) - (src2 + src9) + round_add) >> 5), 0, 255);  
         dst[6*stride] = CLIP(((20 * (src6 + src7) - 3 * (((src5 + src8)<<1) - (src4 + src9)) - (src3 + src10) + round_add) >> 5), 0, 255);  
         dst[7*stride] = CLIP(((20 * (src7 + src8) - 3 * (((src6 + src9)<<1) - (src5 + src10)) - (src4 + src11) + round_add) >> 5), 0, 255);  
         dst[8*stride] = CLIP(((20 * (src8 + src9) - 3 * (((src7 + src10)<<1) - (src6 + src11)) - (src5 + src12) + round_add) >> 5), 0, 255);  
         dst[9*stride] = CLIP(((20 * (src9 + src10) - 3 * (((src8 + src11)<<1) - (src7 + src12)) - (src6 + src13) + round_add) >> 5), 0, 255);  
         dst[10*stride] = CLIP(((20 * (src10 + src11) - 3 * (((src9 + src12)<<1) - (src8 + src13)) - (src7 + src14) + round_add) >> 5), 0, 255);  
         dst[11*stride] = CLIP(((20 * (src11 + src12) - 3 * (((src10 + src13)<<1) - (src9 + src14)) - (src8 + src15) + round_add) >> 5), 0, 255);  
         dst[12*stride] = CLIP(((20 * (src12 + src13) - 3 * (((src11 + src14)<<1) - (src10 + src15)) - (src9 + src16) + round_add) >> 5), 0, 255);  
   
         dst[13*stride] = CLIP(((20 * (src13 + src14) + (src16<<1) + 3 * (src11 - ((src12 + src15) << 1)) - src10 + round_add) >> 5), 0, 255);  
         dst[14*stride] = CLIP(((19 * src15 + 20 * src14 + 3 * (src12 - src16 - (src13 << 1)) - src11 + round_add) >> 5), 0, 255);  
         dst[15*stride] = CLIP(((23 * src15 + 7 * ((src16<<1) - src14) + 3 * src13 - src12 + round_add) >> 5), 0, 255);  
225    
226                  dst++;          src += src_stride;
227          src++;                  dst2 += dst2_stride;
     }  
228  }  }
229    
230  void interpolate8x8_lowpass_v_c(uint8_t *dst, uint8_t *src, int32_t stride, int32_t rounding)      for(i = 0; i < 8; i++)
 {  
     int32_t i;  
         uint8_t round_add = 16 - rounding;  
   
     for(i = 0; i < 9; i++)  
231      {      {
232          int32_t src0 = src[0];          int32_t src0 = data[i];
233          int32_t src1 = src[stride];          int32_t src1 = data[8 + i];
234          int32_t src2 = src[2 * stride];          int32_t src2 = data[2 * 8 + i];
235          int32_t src3 = src[3 * stride];          int32_t src3 = data[3 * 8 + i];
236          int32_t src4 = src[4 * stride];          int32_t src4 = data[4 * 8 + i];
237          int32_t src5 = src[5 * stride];          int32_t src5 = data[5 * 8 + i];
238          int32_t src6 = src[6 * stride];          int32_t src6 = data[6 * 8 + i];
239          int32_t src7 = src[7 * stride];          int32_t src7 = data[7 * 8 + i];
240          int32_t src8 = src[8 * stride];          int32_t src8 = data[8 * 8 + i];
241    
242          dst[0]                  = CLIP(((7 * ((src0<<1) - src2) + 23 * src1 + 3 * src3 - src4 + round_add) >> 5), 0, 255);                  dst1[0] = CLIP((((src0 + src1) * 160 - (src0 + src2) * 48 + (src1 + src3) * 24 - (src2 + src4) * 8 + (128 - rounding)) / 256), 0, 255);
243          dst[stride]             = CLIP(((19 * src1 + 20 * src2 - src5 + 3 * (src4 - src0 - (src3 << 1)) + round_add) >> 5), 0, 255);          dst1[dst1_stride] = CLIP((((src1 + src2) * 160 - (src0 + src3) * 48 + (src0 + src4) * 24 - (src1 + src5) * 8 + (128 - rounding)) / 256), 0, 255);
244          dst[2 * stride] = CLIP(((20 * (src2 + src3) + (src0<<1) + 3 * (src5 - ((src1 + src4) <<1 )) - src6 + round_add) >> 5), 0, 255);          dst1[2 * dst1_stride] = CLIP((((src2 + src3) * 160 - (src1 + src4) * 48 + (src0 + src5) * 24 - (src0 + src6) * 8 + (128 - rounding)) / 256), 0, 255);
245          dst[3 * stride] = CLIP(((20 * (src3 + src4) + 3 * ((src6 + src1) - ((src2 + src5)<<1)) - (src0 + src7) + round_add) >> 5), 0, 255);          dst1[3 * dst1_stride] = CLIP((((src3 + src4) * 160 - (src2 + src5) * 48 + (src1 + src6) * 24 - (src0 + src7) * 8 + (128 - rounding)) / 256), 0, 255);
246          dst[4 * stride] = CLIP(((20 * (src4 + src5) + 3 * ((src2 + src7) - ((src3 + src6)<<1)) - (src1 + src8) + round_add) >> 5), 0, 255);          dst1[4 * dst1_stride] = CLIP((((src4 + src5) * 160 - (src3 + src6) * 48 + (src2 + src7) * 24 - (src1 + src8) * 8 + (128 - rounding)) / 256), 0, 255);
247          dst[5 * stride] = CLIP(((20 * (src5 + src6) + (src8<<1) + 3 * (src3 - ((src4 + src7) << 1)) - src2 + round_add) >> 5), 0, 255);          dst1[5 * dst1_stride] = CLIP((((src5 + src6) * 160 - (src4 + src7) * 48 + (src3 + src8) * 24 - (src2 + src8) * 8 + (128 - rounding)) / 256), 0, 255);
248          dst[6 * stride] = CLIP(((19 * src7 + 20 * src6 - src3 + 3 * (src4 - src8 - (src5 << 1)) + round_add) >> 5), 0, 255);          dst1[6 * dst1_stride] = CLIP((((src6 + src7) * 160 - (src5 + src8) * 48 + (src4 + src8) * 24 - (src3 + src7) * 8 + (128 - rounding)) / 256), 0, 255);
249          dst[7 * stride] = CLIP(((7 * ((src8<<1) - src6) + 23 * src7 + 3 * src5 - src4 + round_add) >> 5), 0, 255);          dst1[7 * dst1_stride] = CLIP((((src7 + src8) * 160 - (src6 + src8) * 48 + (src5 + src7) * 24 - (src4 + src6) * 8 + (128 - rounding)) / 256), 0, 255);
250    
251                  dst++;                  dst1++;
         src++;  
252      }      }
253  }  }
254    
255  void interpolate16x16_lowpass_hv_c(uint8_t *dst1, uint8_t *dst2, uint8_t *src, int32_t stride, int32_t rounding)  void interpolate8x8_bilinear2(uint8_t *dst, uint8_t *src1, uint8_t *src2, int32_t dst_stride, int32_t src_stride, int32_t rounding)
256  {  {
257          int32_t i;          int32_t i;
         uint8_t round_add = 16 - rounding;  
         uint8_t *h_ptr = dst2;  
258    
259      for(i = 0; i < 17; i++)      for(i = 0; i < 8; i++)
260      {      {
261            dst[0] = (src1[0] + src2[0] + (1 - rounding)) >> 1;
262          h_ptr[0] = CLIP(((7 * ((src[0]<<1) - src[2]) +  23 * src[1] + 3 * src[3] - src[4] + round_add) >> 5), 0, 255);          dst[1] = (src1[1] + src2[1] + (1 - rounding)) >> 1;
263          h_ptr[1] = CLIP(((19 * src[1] + 20 * src[2] - src[5] + 3 * (src[4] - src[0] - (src[3]<<1)) + round_add) >> 5), 0, 255);          dst[2] = (src1[2] + src2[2] + (1 - rounding)) >> 1;
264          h_ptr[2] = CLIP(((20 * (src[2] + src[3]) + (src[0]<<1) + 3 * (src[5] - ((src[1] + src[4])<<1)) - src[6] + round_add) >> 5), 0, 255);          dst[3] = (src1[3] + src2[3] + (1 - rounding)) >> 1;
265            dst[4] = (src1[4] + src2[4] + (1 - rounding)) >> 1;
266          h_ptr[3] = CLIP(((20 * (src[3] + src[4]) + 3 * ((src[6] + src[1]) - ((src[2] + src[5])<<1)) - (src[0] + src[7]) + round_add) >> 5), 0, 255);          dst[5] = (src1[5] + src2[5] + (1 - rounding)) >> 1;
267          h_ptr[4] = CLIP(((20 * (src[4] + src[5]) - 3 * (((src[3] + src[6])<<1) - (src[2] + src[7])) - (src[1] + src[8]) + round_add) >> 5), 0, 255);          dst[6] = (src1[6] + src2[6] + (1 - rounding)) >> 1;
268          h_ptr[5] = CLIP(((20 * (src[5] + src[6]) - 3 * (((src[4] + src[7])<<1) - (src[3] + src[8])) - (src[2] + src[9]) + round_add) >> 5), 0, 255);          dst[7] = (src1[7] + src2[7] + (1 - rounding)) >> 1;
269          h_ptr[6] = CLIP(((20 * (src[6] + src[7]) - 3 * (((src[5] + src[8])<<1) - (src[4] + src[9])) - (src[3] + src[10]) + round_add) >> 5), 0, 255);  
270          h_ptr[7] = CLIP(((20 * (src[7] + src[8]) - 3 * (((src[6] + src[9])<<1) - (src[5] + src[10])) - (src[4] + src[11]) + round_add) >> 5), 0, 255);          dst += dst_stride;
271          h_ptr[8] = CLIP(((20 * (src[8] + src[9]) - 3 * (((src[7] + src[10])<<1) - (src[6] + src[11])) - (src[5] + src[12]) + round_add) >> 5), 0, 255);          src1 += src_stride;
272          h_ptr[9] = CLIP(((20 * (src[9] + src[10]) - 3 * (((src[8] + src[11])<<1) - (src[7] + src[12])) - (src[6] + src[13]) + round_add) >> 5), 0, 255);          src2 += 8;
         h_ptr[10] = CLIP(((20 * (src[10] + src[11]) - 3 * (((src[9] + src[12])<<1) - (src[8] + src[13])) - (src[7] + src[14]) + round_add) >> 5), 0, 255);  
         h_ptr[11] = CLIP(((20 * (src[11] + src[12]) - 3 * (((src[10] + src[13])<<1) - (src[9] + src[14])) - (src[8] + src[15]) + round_add) >> 5), 0, 255);  
         h_ptr[12] = CLIP(((20 * (src[12] + src[13]) - 3 * (((src[11] + src[14])<<1) - (src[10] + src[15])) - (src[9] + src[16]) + round_add) >> 5), 0, 255);  
   
         h_ptr[13] = CLIP(((20 * (src[13] + src[14]) + (src[16]<<1) + 3 * (src[11] - ((src[12] + src[15]) << 1)) - src[10] + round_add) >> 5), 0, 255);  
         h_ptr[14] = CLIP(((19 * src[15] + 20 * src[14] + 3 * (src[12] - src[16] - (src[13] << 1)) - src[11] + round_add) >> 5), 0, 255);  
         h_ptr[15] = CLIP(((23 * src[15] + 7 * ((src[16]<<1) - src[14]) + 3 * src[13] - src[12] + round_add) >> 5), 0, 255);  
   
         h_ptr += stride;  
         src += stride;  
273      }      }
   
         interpolate16x16_lowpass_v_c(dst1, dst2, stride, rounding);  
   
274  }  }
275    
276  void interpolate8x8_lowpass_hv_c(uint8_t *dst1, uint8_t *dst2, uint8_t *src, int32_t stride, int32_t rounding)  void interpolate8x8_bilinear4(uint8_t *dst, uint8_t *src1, uint8_t *src2, uint8_t *src3, uint8_t *src4, int32_t stride, int32_t rounding)
277  {  {
278          int32_t i;          int32_t i;
         uint8_t round_add = 16 - rounding;  
         uint8_t *h_ptr = dst2;  
279    
280      for(i = 0; i < 9; i++)      for(i = 0; i < 8; i++)
281      {      {
282            dst[0] = (src1[0] + src2[0] + src3[0] + src4[0] + (2 - rounding)) >> 2;
283            dst[1] = (src1[1] + src2[1] + src3[1] + src4[1] + (2 - rounding)) >> 2;
284            dst[2] = (src1[2] + src2[2] + src3[2] + src4[2] + (2 - rounding)) >> 2;
285            dst[3] = (src1[3] + src2[3] + src3[3] + src4[3] + (2 - rounding)) >> 2;
286            dst[4] = (src1[4] + src2[4] + src3[4] + src4[4] + (2 - rounding)) >> 2;
287            dst[5] = (src1[5] + src2[5] + src3[5] + src4[5] + (2 - rounding)) >> 2;
288            dst[6] = (src1[6] + src2[6] + src3[6] + src4[6] + (2 - rounding)) >> 2;
289            dst[7] = (src1[7] + src2[7] + src3[7] + src4[7] + (2 - rounding)) >> 2;
290    
291          h_ptr[0] = CLIP(((7 * ((src[0]<<1) - src[2]) + 23 * src[1] + 3 * src[3] - src[4] + round_add) >> 5), 0, 255);                  dst += stride;
292          h_ptr[1] = CLIP(((19 * src[1] + 20 * src[2] - src[5] + 3 * (src[4] - src[0] - (src[3]<<1)) + round_add) >> 5), 0, 255);          src1 += stride;
293          h_ptr[2] = CLIP(((20 * (src[2] + src[3]) + (src[0]<<1) + 3 * (src[5] - ((src[1] + src[4])<<1)) - src[6] + round_add) >> 5), 0, 255);          src2 += 8;
294          h_ptr[3] = CLIP(((20 * (src[3] + src[4]) + 3 * ((src[6] + src[1]) - ((src[2] + src[5])<<1)) - (src[0] + src[7]) + round_add) >> 5), 0, 255);          src3 += 8;
295          h_ptr[4] = CLIP(((20 * (src[4] + src[5]) - 3 * (((src[3] + src[6])<<1) - (src[2] + src[7])) - (src[1] + src[8]) + round_add) >> 5), 0, 255);          src4 += 8;
         h_ptr[5] = CLIP(((20 * (src[5] + src[6]) + (src[8]<<1) + 3 * (src[3] - ((src[4] + src[7]) << 1)) - src[2] + round_add) >> 5), 0, 255);  
         h_ptr[6] = CLIP(((19 * src[7] + 20 * src[6] + 3 * (src[4] - src[8] - (src[5] << 1)) - src[3] + round_add) >> 5), 0, 255);  
         h_ptr[7] = CLIP(((23 * src[7] + 7 * ((src[8]<<1) - src[6]) + 3 * src[5] - src[4] + round_add) >> 5), 0, 255);  
   
         h_ptr += stride;  
         src += stride;  
296      }      }
   
         interpolate8x8_lowpass_v_c(dst1, dst2, stride, rounding);  
   
297  }  }
Legend:
Removed from v.1.4.2.4  
changed lines
  Added in v.1.6
No admin address has been configured
 ViewVC Help
Powered by ViewVC 1.0.4