changeset 24677:9f81c2175803

overcomplete wavelet denoiser not optimized at all but much cleaner than uspp
author michael
date Wed, 03 Oct 2007 23:50:54 +0000
parents c03afe4ed741
children 9aab9d7b51f0
files libmpcodecs/Makefile libmpcodecs/vf.c libmpcodecs/vf_ow.c
diffstat 3 files changed, 341 insertions(+), 0 deletions(-) [+]
line wrap: on
line diff
--- a/libmpcodecs/Makefile	Wed Oct 03 21:39:01 2007 +0000
+++ b/libmpcodecs/Makefile	Wed Oct 03 23:50:54 2007 +0000
@@ -93,6 +93,7 @@
               vf_yuvcsp.c \
               vf_yuy2.c \
               vf_yvu9.c \
+              vf_ow.c \
 
 # These filters use private headers and do not work with shared libavcodec.
 SRCS_COMMON-$(LIBAVCODEC_A)          += vf_fspp.c \
--- a/libmpcodecs/vf.c	Wed Oct 03 21:39:01 2007 +0000
+++ b/libmpcodecs/vf.c	Wed Oct 03 23:50:54 2007 +0000
@@ -98,6 +98,7 @@
 extern vf_info_t vf_info_yadif;
 extern vf_info_t vf_info_blackframe;
 extern vf_info_t vf_info_geq;
+extern vf_info_t vf_info_ow;
 
 // list of available filters:
 static vf_info_t* filter_list[]={
@@ -191,6 +192,7 @@
 #endif
     &vf_info_yadif,
     &vf_info_blackframe,
+    &vf_info_ow,
     NULL
 };
 
--- /dev/null	Thu Jan 01 00:00:00 1970 +0000
+++ b/libmpcodecs/vf_ow.c	Wed Oct 03 23:50:54 2007 +0000
@@ -0,0 +1,338 @@
+/*
+    Copyright (C) 2007 Michael Niedermayer <michaelni@gmx.at>
+
+    This program is free software; you can redistribute it and/or modify
+    it under the terms of the GNU General Public License as published by
+    the Free Software Foundation; either version 2 of the License, or
+    (at your option) any later version.
+
+    This program is distributed in the hope that it will be useful,
+    but WITHOUT ANY WARRANTY; without even the implied warranty of
+    MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+    GNU General Public License for more details.
+
+    You should have received a copy of the GNU General Public License
+    along with this program; if not, write to the Free Software
+    Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA 02110-1301 USA
+*/
+
+/**
+ * @todo change to float or int
+ * @todo try lifting based implementation
+ * @todo optimize optimize optimize
+ * @todo hard tresholding
+ * @todo use QP to decide filter strength
+ * @todo wavelet normalization / least squares optimal signal vs. noise thresholds
+ */
+ 
+#include <stdio.h>
+#include <string.h>
+#include <inttypes.h>
+#include <math.h>
+
+#include "config.h"
+
+#include "mp_msg.h"
+
+#ifdef HAVE_MALLOC_H
+#include <malloc.h>
+#endif
+
+#include "img_format.h"
+#include "mp_image.h"
+#include "vf.h"
+
+//===========================================================================//
+static const uint8_t  __attribute__((aligned(8))) dither[8][8]={
+{  0,  48,  12,  60,   3,  51,  15,  63, },
+{ 32,  16,  44,  28,  35,  19,  47,  31, },
+{  8,  56,   4,  52,  11,  59,   7,  55, },
+{ 40,  24,  36,  20,  43,  27,  39,  23, },
+{  2,  50,  14,  62,   1,  49,  13,  61, },
+{ 34,  18,  46,  30,  33,  17,  45,  29, },
+{ 10,  58,   6,  54,   9,  57,   5,  53, },
+{ 42,  26,  38,  22,  41,  25,  37,  21, },
+};
+//FIXME the above is duplicated in many filters
+
+struct vf_priv_s {
+    float strength[2];
+    float delta;
+    int mode;
+    int depth;
+    double *plane[16][4];
+    int stride;
+};
+
+#define S 1.41421356237 //sqrt(2)
+
+static const double coeff[2][5]={
+    {
+         0.6029490182363579  *S,
+         0.2668641184428723  *S,
+        -0.07822326652898785 *S,
+        -0.01686411844287495 *S,
+         0.02674875741080976 *S
+    },{
+         1.115087052456994   /S,
+        -0.5912717631142470  /S,
+        -0.05754352622849957 /S,
+         0.09127176311424948 /S
+    }
+};
+
+static const double icoeff[2][5]={
+    {
+         1.115087052456994   /S,
+         0.5912717631142470  /S,
+        -0.05754352622849957 /S,
+        -0.09127176311424948 /S
+    },{
+         0.6029490182363579  *S,
+        -0.2668641184428723  *S,
+        -0.07822326652898785 *S,
+         0.01686411844287495 *S,
+         0.02674875741080976 *S
+    }
+};
+#undef S
+
+static inline int mirror(int x, int w){
+    while((unsigned)x > (unsigned)w){
+        x=-x;
+        if(x<0) x+= 2*w;
+    }
+    return x;
+}
+
+static inline void decompose(double *dstL, double *dstH, double *src, int stride, int w){
+    int x, i;
+    for(x=0; x<w; x++){
+        double sumL= src[x*stride] * coeff[0][0];
+        double sumH= src[x*stride] * coeff[1][0];
+        for(i=1; i<=4; i++){
+            double s= (src[mirror(x-i, w-1)*stride] + src[mirror(x+i, w-1)*stride]);
+
+            sumL+= coeff[0][i]*s;
+            sumH+= coeff[1][i]*s;
+        }
+        dstL[x*stride]= sumL;
+        dstH[x*stride]= sumH;
+    }
+}
+
+static inline void compose(double *dst, double *srcL, double *srcH, int stride, int w){
+    int x, i;
+    for(x=0; x<w; x++){
+        double sumL= srcL[x*stride] * icoeff[0][0];
+        double sumH= srcH[x*stride] * icoeff[1][0];
+        for(i=1; i<=4; i++){
+            int x0= mirror(x-i, w-1)*stride;
+            int x1= mirror(x+i, w-1)*stride;
+
+            sumL+= icoeff[0][i]*(srcL[x0] + srcL[x1]);
+            sumH+= icoeff[1][i]*(srcH[x0] + srcH[x1]);
+        }
+        dst[x*stride]= (sumL + sumH)*0.5;
+    }
+}
+
+static inline void decompose2D(double *dstL, double *dstH, double *src, int xstride, int ystride, int step, int w, int h){
+    int y, x;
+    for(y=0; y<h; y++)
+        for(x=0; x<step; x++)
+            decompose(dstL + ystride*y + xstride*x, dstH + ystride*y + xstride*x, src + ystride*y + xstride*x, step*xstride, (w-x+step-1)/step);
+}
+
+static inline void compose2D(double *dst, double *srcL, double *srcH, int xstride, int ystride, int step, int w, int h){
+    int y, x;
+    for(y=0; y<h; y++)
+        for(x=0; x<step; x++)
+            compose(dst + ystride*y + xstride*x, srcL + ystride*y + xstride*x, srcH + ystride*y + xstride*x, step*xstride, (w-x+step-1)/step);
+}
+
+static void decompose2D2(double *dst[4], double *src, double *temp[2], int stride, int step, int w, int h){
+    decompose2D(temp[0], temp[1], src    , 1, stride, step  , w, h);
+    decompose2D( dst[0],  dst[1], temp[0], stride, 1, step  , h, w);
+    decompose2D( dst[2],  dst[3], temp[1], stride, 1, step  , h, w);
+}
+
+static void compose2D2(double *dst, double *src[4], double *temp[2], int stride, int step, int w, int h){
+    compose2D(temp[0],  src[0],  src[1], stride, 1, step  , h, w);
+    compose2D(temp[1],  src[2],  src[3], stride, 1, step  , h, w);
+    compose2D(dst    , temp[0], temp[1], 1, stride, step  , w, h);
+}
+
+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){
+    int x,y, i, j;
+    double sum=0;
+    double s= p->strength[!is_luma];
+
+    for(y=0; y<height; y++)
+        for(x=0; x<width; x++)
+            p->plane[0][0][x + y*p->stride]= src[x + y*src_stride];
+
+    for(i=0; i<p->depth; i++){
+        decompose2D2(p->plane[i+1], p->plane[i][0], p->plane[0]+1,p->stride, 1<<i, width, height);
+    }
+    for(i=0; i<p->depth; i++){
+        for(j=1; j<4; j++){
+            for(y=0; y<height; y++){
+                for(x=0; x<width; x++){
+                    double v= p->plane[i+1][j][x + y*p->stride];
+                    if     (v> s) v-=s;
+                    else if(v<-s) v+=s;
+                    else          v =0;
+                    p->plane[i+1][j][x + y*p->stride]= v;
+                }
+            }
+        }
+    }
+    for(i=p->depth-1; i>=0; i--){
+        compose2D2(p->plane[i][0], p->plane[i+1], p->plane[0]+1, p->stride, 1<<i, width, height);
+    }
+
+    for(y=0; y<height; y++)
+        for(x=0; x<width; x++){
+            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 :)
+//            double e= i - src[x + y*src_stride];
+//            sum += e*e;
+            if((unsigned)i > 255U) i= ~(i>>31);
+            dst[x + y*dst_stride]= i;
+        }
+
+//    printf("%f\n", sum/height/width);
+}
+
+static int config(struct vf_instance_s* vf, int width, int height, int d_width, int d_height, unsigned int flags, unsigned int outfmt){
+    int h= (height+15)&(~15);
+    int i,j;
+
+    vf->priv->stride= (width+15)&(~15);
+    for(j=0; j<4; j++){
+        for(i=0; i<=vf->priv->depth; i++)
+            vf->priv->plane[i][j]= malloc(vf->priv->stride*h*sizeof(vf->priv->plane[0][0][0]));
+    }
+
+    return vf_next_config(vf,width,height,d_width,d_height,flags,outfmt);
+}
+
+static void get_image(struct vf_instance_s* vf, mp_image_t *mpi){
+    if(mpi->flags&MP_IMGFLAG_PRESERVE) return; // don't change
+    // ok, we can do pp in-place (or pp disabled):
+    vf->dmpi=vf_get_image(vf->next,mpi->imgfmt,
+        mpi->type, mpi->flags | MP_IMGFLAG_READABLE, mpi->width, mpi->height);
+    mpi->planes[0]=vf->dmpi->planes[0];
+    mpi->stride[0]=vf->dmpi->stride[0];
+    mpi->width=vf->dmpi->width;
+    if(mpi->flags&MP_IMGFLAG_PLANAR){
+        mpi->planes[1]=vf->dmpi->planes[1];
+        mpi->planes[2]=vf->dmpi->planes[2];
+        mpi->stride[1]=vf->dmpi->stride[1];
+        mpi->stride[2]=vf->dmpi->stride[2];
+    }
+    mpi->flags|=MP_IMGFLAG_DIRECT;
+}
+
+static int put_image(struct vf_instance_s* vf, mp_image_t *mpi, double pts){
+    mp_image_t *dmpi;
+
+    if(!(mpi->flags&MP_IMGFLAG_DIRECT)){
+        // no DR, so get a new image! hope we'll get DR buffer:
+        dmpi=vf_get_image(vf->next,mpi->imgfmt,
+            MP_IMGTYPE_TEMP,
+            MP_IMGFLAG_ACCEPT_STRIDE|MP_IMGFLAG_PREFER_ALIGNED_STRIDE,
+            mpi->width,mpi->height);
+        vf_clone_mpi_attributes(dmpi, mpi);
+    }else{
+        dmpi=vf->dmpi;
+    }
+
+    filter(vf->priv, dmpi->planes[0], mpi->planes[0], dmpi->stride[0], mpi->stride[0], mpi->w, mpi->h, 1);
+    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);
+    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);
+
+    return vf_next_put_image(vf,dmpi, pts);
+}
+
+static void uninit(struct vf_instance_s* vf){
+    int i,j;
+    if(!vf->priv) return;
+
+    for(j=0; j<4; j++){
+        for(i=0; i<16; i++){
+            free(vf->priv->plane[i][j]);
+            vf->priv->plane[i][j]= NULL;
+        }
+    }
+
+    free(vf->priv);
+    vf->priv=NULL;
+}
+
+//===========================================================================//
+static int query_format(struct vf_instance_s* vf, unsigned int fmt){
+    switch(fmt){
+        case IMGFMT_YVU9:
+        case IMGFMT_IF09:
+        case IMGFMT_YV12:
+        case IMGFMT_I420:
+        case IMGFMT_IYUV:
+        case IMGFMT_CLPL:
+        case IMGFMT_Y800:
+        case IMGFMT_Y8:
+        case IMGFMT_444P:
+        case IMGFMT_422P:
+        case IMGFMT_411P:
+            return vf_next_query_format(vf,fmt);
+    }
+    return 0;
+}
+
+static unsigned int fmt_list[]={
+    IMGFMT_YVU9,
+    IMGFMT_IF09,
+    IMGFMT_YV12,
+    IMGFMT_I420,
+    IMGFMT_IYUV,
+    IMGFMT_CLPL,
+    IMGFMT_Y800,
+    IMGFMT_Y8,
+    IMGFMT_444P,
+    IMGFMT_422P,
+    IMGFMT_411P,
+    0
+};
+
+
+static int open(vf_instance_t *vf, char* args){
+    vf->config=config;
+    vf->put_image=put_image;
+    vf->get_image=get_image;
+    vf->query_format=query_format;
+    vf->uninit=uninit;
+    vf->priv=malloc(sizeof(struct vf_priv_s));
+    memset(vf->priv, 0, sizeof(struct vf_priv_s));
+
+    vf->priv->depth= 8;
+    vf->priv->strength[0]= 1.0;
+    vf->priv->strength[1]= 1.0;
+    vf->priv->delta= 1.0;
+
+    if (args) sscanf(args, "%d:%f:%f:%d:%f", &vf->priv->depth,
+                     &vf->priv->strength[0],
+                     &vf->priv->strength[1],
+                     &vf->priv->mode,
+                     &vf->priv->delta);
+
+    return 1;
+}
+
+vf_info_t vf_info_ow = {
+    "overcomplete wavelet denoiser",
+    "ow",
+    "Michael Niedermayer",
+    "",
+    open,
+    NULL
+};