FFmpeg
 All Data Structures Files Functions Variables Typedefs Enumerations Enumerator Macros Groups Pages
vf_ow.c
Go to the documentation of this file.
1 /*
2  * Copyright (C) 2007 Michael Niedermayer <michaelni@gmx.at>
3  *
4  * This file is part of MPlayer.
5  *
6  * MPlayer is free software; you can redistribute it and/or modify
7  * it under the terms of the GNU General Public License as published by
8  * the Free Software Foundation; either version 2 of the License, or
9  * (at your option) any later version.
10  *
11  * MPlayer 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
14  * GNU General Public License for more details.
15  *
16  * You should have received a copy of the GNU General Public License along
17  * with MPlayer; if not, write to the Free Software Foundation, Inc.,
18  * 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA.
19  */
20 
21 /**
22  * @todo try to change to int
23  * @todo try lifting based implementation
24  * @todo optimize optimize optimize
25  * @todo hard tresholding
26  * @todo use QP to decide filter strength
27  * @todo wavelet normalization / least squares optimal signal vs. noise thresholds
28  */
29 
30 #include <stdio.h>
31 #include <string.h>
32 #include <inttypes.h>
33 #include <math.h>
34 
35 #include "mp_msg.h"
36 #include "img_format.h"
37 #include "mp_image.h"
38 #include "vf.h"
39 
40 //===========================================================================//
41 static const uint8_t __attribute__((aligned(8))) dither[8][8]={
42 { 0, 48, 12, 60, 3, 51, 15, 63, },
43 { 32, 16, 44, 28, 35, 19, 47, 31, },
44 { 8, 56, 4, 52, 11, 59, 7, 55, },
45 { 40, 24, 36, 20, 43, 27, 39, 23, },
46 { 2, 50, 14, 62, 1, 49, 13, 61, },
47 { 34, 18, 46, 30, 33, 17, 45, 29, },
48 { 10, 58, 6, 54, 9, 57, 5, 53, },
49 { 42, 26, 38, 22, 41, 25, 37, 21, },
50 };
51 //FIXME the above is duplicated in many filters
52 
53 struct vf_priv_s {
54  float strength[2];
55  float delta;
56  int mode;
57  int depth;
58  float *plane[16][4];
59  int stride;
60 };
61 
62 #define S 1.41421356237 //sqrt(2)
63 
64 static const double coeff[2][5]={
65  {
66  0.6029490182363579 *S,
67  0.2668641184428723 *S,
68  -0.07822326652898785 *S,
69  -0.01686411844287495 *S,
70  0.02674875741080976 *S
71  },{
72  1.115087052456994 /S,
73  -0.5912717631142470 /S,
74  -0.05754352622849957 /S,
75  0.09127176311424948 /S
76  }
77 };
78 
79 static const double icoeff[2][5]={
80  {
81  1.115087052456994 /S,
82  0.5912717631142470 /S,
83  -0.05754352622849957 /S,
84  -0.09127176311424948 /S
85  },{
86  0.6029490182363579 *S,
87  -0.2668641184428723 *S,
88  -0.07822326652898785 *S,
89  0.01686411844287495 *S,
90  0.02674875741080976 *S
91  }
92 };
93 #undef S
94 
95 static inline int mirror(int x, int w){
96  while((unsigned)x > (unsigned)w){
97  x=-x;
98  if(x<0) x+= 2*w;
99  }
100  return x;
101 }
102 
103 static inline void decompose(float *dstL, float *dstH, float *src, int stride, int w){
104  int x, i;
105  for(x=0; x<w; x++){
106  double sumL= src[x*stride] * coeff[0][0];
107  double sumH= src[x*stride] * coeff[1][0];
108  for(i=1; i<=4; i++){
109  double s= (src[mirror(x-i, w-1)*stride] + src[mirror(x+i, w-1)*stride]);
110 
111  sumL+= coeff[0][i]*s;
112  sumH+= coeff[1][i]*s;
113  }
114  dstL[x*stride]= sumL;
115  dstH[x*stride]= sumH;
116  }
117 }
118 
119 static inline void compose(float *dst, float *srcL, float *srcH, int stride, int w){
120  int x, i;
121  for(x=0; x<w; x++){
122  double sumL= srcL[x*stride] * icoeff[0][0];
123  double sumH= srcH[x*stride] * icoeff[1][0];
124  for(i=1; i<=4; i++){
125  int x0= mirror(x-i, w-1)*stride;
126  int x1= mirror(x+i, w-1)*stride;
127 
128  sumL+= icoeff[0][i]*(srcL[x0] + srcL[x1]);
129  sumH+= icoeff[1][i]*(srcH[x0] + srcH[x1]);
130  }
131  dst[x*stride]= (sumL + sumH)*0.5;
132  }
133 }
134 
135 static inline void decompose2D(float *dstL, float *dstH, float *src, int xstride, int ystride, int step, int w, int h){
136  int y, x;
137  for(y=0; y<h; y++)
138  for(x=0; x<step; x++)
139  decompose(dstL + ystride*y + xstride*x, dstH + ystride*y + xstride*x, src + ystride*y + xstride*x, step*xstride, (w-x+step-1)/step);
140 }
141 
142 static inline void compose2D(float *dst, float *srcL, float *srcH, int xstride, int ystride, int step, int w, int h){
143  int y, x;
144  for(y=0; y<h; y++)
145  for(x=0; x<step; x++)
146  compose(dst + ystride*y + xstride*x, srcL + ystride*y + xstride*x, srcH + ystride*y + xstride*x, step*xstride, (w-x+step-1)/step);
147 }
148 
149 static void decompose2D2(float *dst[4], float *src, float *temp[2], int stride, int step, int w, int h){
150  decompose2D(temp[0], temp[1], src , 1, stride, step , w, h);
151  decompose2D( dst[0], dst[1], temp[0], stride, 1, step , h, w);
152  decompose2D( dst[2], dst[3], temp[1], stride, 1, step , h, w);
153 }
154 
155 static void compose2D2(float *dst, float *src[4], float *temp[2], int stride, int step, int w, int h){
156  compose2D(temp[0], src[0], src[1], stride, 1, step , h, w);
157  compose2D(temp[1], src[2], src[3], stride, 1, step , h, w);
158  compose2D(dst , temp[0], temp[1], 1, stride, step , w, h);
159 }
160 
161 static void filter(struct vf_priv_s *p, uint8_t *dst, uint8_t *src, int dst_stride, int src_stride, int width, int height, int is_luma){
162  int x,y, i, j;
163 // double sum=0;
164  double s= p->strength[!is_luma];
165  int depth= p->depth;
166 
167  while(1<<depth > width || 1<<depth > height)
168  depth--;
169 
170  for(y=0; y<height; y++)
171  for(x=0; x<width; x++)
172  p->plane[0][0][x + y*p->stride]= src[x + y*src_stride];
173 
174  for(i=0; i<depth; i++){
175  decompose2D2(p->plane[i+1], p->plane[i][0], p->plane[0]+1,p->stride, 1<<i, width, height);
176  }
177  for(i=0; i<depth; i++){
178  for(j=1; j<4; j++){
179  for(y=0; y<height; y++){
180  for(x=0; x<width; x++){
181  double v= p->plane[i+1][j][x + y*p->stride];
182  if (v> s) v-=s;
183  else if(v<-s) v+=s;
184  else v =0;
185  p->plane[i+1][j][x + y*p->stride]= v;
186  }
187  }
188  }
189  }
190  for(i=depth-1; i>=0; i--){
191  compose2D2(p->plane[i][0], p->plane[i+1], p->plane[0]+1, p->stride, 1<<i, width, height);
192  }
193 
194  for(y=0; y<height; y++)
195  for(x=0; x<width; x++){
196  i= p->plane[0][0][x + y*p->stride] + dither[x&7][y&7]*(1.0/64) + 1.0/128; //yes the rounding is insane but optimal :)
197 // double e= i - src[x + y*src_stride];
198 // sum += e*e;
199  if((unsigned)i > 255U) i= ~(i>>31);
200  dst[x + y*dst_stride]= i;
201  }
202 
203 // printf("%f\n", sum/height/width);
204 }
205 
206 static int config(struct vf_instance *vf, int width, int height, int d_width, int d_height, unsigned int flags, unsigned int outfmt){
207  int h= (height+15)&(~15);
208  int i,j;
209 
210  vf->priv->stride= (width+15)&(~15);
211  for(j=0; j<4; j++){
212  for(i=0; i<=vf->priv->depth; i++)
213  vf->priv->plane[i][j]= malloc(vf->priv->stride*h*sizeof(vf->priv->plane[0][0][0]));
214  }
215 
216  return ff_vf_next_config(vf,width,height,d_width,d_height,flags,outfmt);
217 }
218 
219 static void get_image(struct vf_instance *vf, mp_image_t *mpi){
220  if(mpi->flags&MP_IMGFLAG_PRESERVE) return; // don't change
221  // ok, we can do pp in-place (or pp disabled):
222  vf->dmpi=ff_vf_get_image(vf->next,mpi->imgfmt,
223  mpi->type, mpi->flags | MP_IMGFLAG_READABLE, mpi->width, mpi->height);
224  mpi->planes[0]=vf->dmpi->planes[0];
225  mpi->stride[0]=vf->dmpi->stride[0];
226  mpi->width=vf->dmpi->width;
227  if(mpi->flags&MP_IMGFLAG_PLANAR){
228  mpi->planes[1]=vf->dmpi->planes[1];
229  mpi->planes[2]=vf->dmpi->planes[2];
230  mpi->stride[1]=vf->dmpi->stride[1];
231  mpi->stride[2]=vf->dmpi->stride[2];
232  }
233  mpi->flags|=MP_IMGFLAG_DIRECT;
234 }
235 
236 static int put_image(struct vf_instance *vf, mp_image_t *mpi, double pts){
237  mp_image_t *dmpi;
238 
239  if(!(mpi->flags&MP_IMGFLAG_DIRECT)){
240  // no DR, so get a new image! hope we'll get DR buffer:
241  dmpi=ff_vf_get_image(vf->next,mpi->imgfmt,
244  mpi->width,mpi->height);
245  ff_vf_clone_mpi_attributes(dmpi, mpi);
246  }else{
247  dmpi=vf->dmpi;
248  }
249 
250  filter(vf->priv, dmpi->planes[0], mpi->planes[0], dmpi->stride[0], mpi->stride[0], mpi->w, mpi->h, 1);
251  filter(vf->priv, dmpi->planes[1], mpi->planes[1], dmpi->stride[1], mpi->stride[1], mpi->w>>mpi->chroma_x_shift, mpi->h>>mpi->chroma_y_shift, 0);
252  filter(vf->priv, dmpi->planes[2], mpi->planes[2], dmpi->stride[2], mpi->stride[2], mpi->w>>mpi->chroma_x_shift, mpi->h>>mpi->chroma_y_shift, 0);
253 
254  return ff_vf_next_put_image(vf,dmpi, pts);
255 }
256 
257 static void uninit(struct vf_instance *vf){
258  int i,j;
259  if(!vf->priv) return;
260 
261  for(j=0; j<4; j++){
262  for(i=0; i<16; i++){
263  free(vf->priv->plane[i][j]);
264  vf->priv->plane[i][j]= NULL;
265  }
266  }
267 
268  free(vf->priv);
269  vf->priv=NULL;
270 }
271 
272 //===========================================================================//
273 static int query_format(struct vf_instance *vf, unsigned int fmt){
274  switch(fmt){
275  case IMGFMT_YVU9:
276  case IMGFMT_IF09:
277  case IMGFMT_YV12:
278  case IMGFMT_I420:
279  case IMGFMT_IYUV:
280  case IMGFMT_CLPL:
281  case IMGFMT_Y800:
282  case IMGFMT_Y8:
283  case IMGFMT_444P:
284  case IMGFMT_422P:
285  case IMGFMT_411P:
286  return ff_vf_next_query_format(vf,fmt);
287  }
288  return 0;
289 }
290 
291 
292 static int vf_open(vf_instance_t *vf, char *args){
293  vf->config=config;
294  vf->put_image=put_image;
295  vf->get_image=get_image;
297  vf->uninit=uninit;
298  vf->priv=malloc(sizeof(struct vf_priv_s));
299  memset(vf->priv, 0, sizeof(struct vf_priv_s));
300 
301  vf->priv->depth= 8;
302  vf->priv->strength[0]= 1.0;
303  vf->priv->strength[1]= 1.0;
304  vf->priv->delta= 1.0;
305 
306  if (args) sscanf(args, "%d:%f:%f:%d:%f", &vf->priv->depth,
307  &vf->priv->strength[0],
308  &vf->priv->strength[1],
309  &vf->priv->mode,
310  &vf->priv->delta);
311 
312  return 1;
313 }
314 
316  "overcomplete wavelet denoiser",
317  "ow",
318  "Michael Niedermayer",
319  "",
320  vf_open,
321  NULL
322 };