annotate libmpcodecs/vf_expand.c @ 17588:79081ba52e00

Move the v{Y,C}CoeffsBank vectors into the SwsContext, filling them in just once when the scaler is initialized, instead of building them and freeing them over and over. This gives massive performance improvements. patch by Alan Curry, pacman*at*TheWorld*dot*com

#define OSD_SUPPORT

#include <stdio.h>
#include <stdlib.h>
#include <string.h>

#include "config.h"
#include "mp_msg.h"

#include "img_format.h"
#include "mp_image.h"
#include "vf.h"

#include "libvo/fastmemcpy.h"

#ifdef OSD_SUPPORT
#include "libvo/sub.h"
#include "libvo/osd.h"
#endif

#include "m_option.h"
#include "m_struct.h"

#define MAX(a,b) ((a) > (b) ? (a) : (b))

static struct vf_priv_s {
    int exp_w,exp_h;
    int exp_x,exp_y;
    int osd;
    double aspect;
    int round;
    unsigned char* fb_ptr;
    int first_slice;
} vf_priv_dflt = {
  -1,-1,
  -1,-1,
  0,
  0.,
  1,
  NULL,
  0
};

extern int opt_screen_size_x;
extern int opt_screen_size_y;

//===========================================================================//
#ifdef OSD_SUPPORT

static struct vf_instance_s* vf=NULL; // fixme (needs sub.c changes)
static int orig_w,orig_h;

static void remove_func_2(int x0,int y0, int w,int h){
    // TODO: let's cleanup the place
    //printf("OSD clear: %d;%d %dx%d  \n",x0,y0,w,h);
    vf_mpi_clear(vf->dmpi,x0,y0,w,h);
}

static void remove_func(int x0,int y0, int w,int h){
    if(!vo_osd_changed_flag) return;
    // split it to 4 parts:
    if(y0<vf->priv->exp_y){
	// it has parts above the image:
	int y=y0+h;
	if(y>vf->priv->exp_y) y=vf->priv->exp_y;
	remove_func_2(x0,y0,w,y-y0);
	if(y0+h<=vf->priv->exp_y) return;
	h-=y-y0;y0=y;
    }
    if(y0+h>vf->priv->exp_y+orig_h){
	// it has parts under the image:
	int y=y0;
	if(y<vf->priv->exp_y+orig_h) y=vf->priv->exp_y+orig_h;
	remove_func_2(x0,y,w,y0+h-y);
	if(y0>=vf->priv->exp_y+orig_h) return;
	h=y-y0;
    }
    if(x0>=vf->priv->exp_x || x0+w<=vf->priv->exp_x+orig_w) return;
    // TODO  clear left and right side of the image if needed
}

static void draw_func(int x0,int y0, int w,int h,unsigned char* src, unsigned char *srca, int stride){
    unsigned char* dst;
    if(!vo_osd_changed_flag && vf->dmpi->planes[0]==vf->priv->fb_ptr){
	// ok, enough to update the area inside the video, leave the black bands
	// untouched!
	if(x0<vf->priv->exp_x){
	    int tmp=vf->priv->exp_x-x0;
	    w-=tmp; src+=tmp; srca+=tmp; x0+=tmp;
	}
	if(y0<vf->priv->exp_y){
	    int tmp=vf->priv->exp_y-y0;
	    h-=tmp; src+=tmp*stride; srca+=tmp*stride; y0+=tmp;
	}
	if(x0+w>vf->priv->exp_x+orig_w){
	    w=vf->priv->exp_x+orig_w-x0;
	}
	if(y0+h>vf->priv->exp_y+orig_h){
	    h=vf->priv->exp_y+orig_h-y0;
	}
    }
    if(w<=0 || h<=0) return; // nothing to do...
//    printf("OSD redraw: %d;%d %dx%d  \n",x0,y0,w,h);
    dst=vf->dmpi->planes[0]+
			vf->dmpi->stride[0]*y0+
			(vf->dmpi->bpp>>3)*x0;
    switch(vf->dmpi->imgfmt){
    case IMGFMT_BGR15:
    case IMGFMT_RGB15:
	vo_draw_alpha_rgb15(w,h,src,srca,stride,dst,vf->dmpi->stride[0]);
	break;
    case IMGFMT_BGR16:
    case IMGFMT_RGB16:
	vo_draw_alpha_rgb16(w,h,src,srca,stride,dst,vf->dmpi->stride[0]);
	break;
    case IMGFMT_BGR24:
    case IMGFMT_RGB24:
	vo_draw_alpha_rgb24(w,h,src,srca,stride,dst,vf->dmpi->stride[0]);
	break;
    case IMGFMT_BGR32:
    case IMGFMT_RGB32:
	vo_draw_alpha_rgb32(w,h,src,srca,stride,dst,vf->dmpi->stride[0]);
	break;
    case IMGFMT_YV12:
    case IMGFMT_I420:
    case IMGFMT_IYUV:
    case IMGFMT_YVU9:
    case IMGFMT_IF09:
    case IMGFMT_Y800:
    case IMGFMT_Y8:
	vo_draw_alpha_yv12(w,h,src,srca,stride,dst,vf->dmpi->stride[0]);
	break;
    case IMGFMT_YUY2:
	vo_draw_alpha_yuy2(w,h,src,srca,stride,dst,vf->dmpi->stride[0]);
	break;
    case IMGFMT_UYVY:
	vo_draw_alpha_yuy2(w,h,src,srca,stride,dst+1,vf->dmpi->stride[0]);
	break;
    }
}

static void draw_osd(struct vf_instance_s* vf_,int w,int h){
    vf=vf_;orig_w=w;orig_h=h;
//    printf("======================================\n");
    if(vf->priv->exp_w!=w || vf->priv->exp_h!=h ||
	vf->priv->exp_x || vf->priv->exp_y){
	// yep, we're expanding image, not just copy.
	if(vf->dmpi->planes[0]!=vf->priv->fb_ptr){
	    // double buffering, so we need full clear :(
	    if (vf->priv->exp_y > 0)
		remove_func_2(0,0,vf->priv->exp_w,vf->priv->exp_y);
	    if (vf->priv->exp_y+h < vf->priv->exp_h)
		remove_func_2(0,vf->priv->exp_y+h,vf->priv->exp_w,vf->priv->exp_h-h-vf->priv->exp_y);
	} else {
	    // partial clear:
	    vo_remove_text(vf->priv->exp_w,vf->priv->exp_h,remove_func);
	}
    }
    vo_draw_text(vf->priv->exp_w,vf->priv->exp_h,draw_func);
    // save buffer pointer for double buffering detection - yes, i know it's
    // ugly method, but note that codecs with DR support does the same...
    if(vf->dmpi)
      vf->priv->fb_ptr=vf->dmpi->planes[0];
}

#endif
//===========================================================================//

static int config(struct vf_instance_s* vf,
        int width, int height, int d_width, int d_height,
	unsigned int flags, unsigned int outfmt){
    if (outfmt == IMGFMT_IF09) return 0;
    // calculate the missing parameters:
#if 0
    if(vf->priv->exp_w<width) vf->priv->exp_w=width;
    if(vf->priv->exp_h<height) vf->priv->exp_h=height;
#else
    if ( vf->priv->exp_w == -1 ) vf->priv->exp_w=width;
      else if (vf->priv->exp_w < -1 ) vf->priv->exp_w=width - vf->priv->exp_w;
        else if ( vf->priv->exp_w<width ) vf->priv->exp_w=width;
    if ( vf->priv->exp_h == -1 ) vf->priv->exp_h=height;
      else if ( vf->priv->exp_h < -1 ) vf->priv->exp_h=height - vf->priv->exp_h;
        else if( vf->priv->exp_h<height ) vf->priv->exp_h=height;
#endif
    if (vf->priv->aspect) {
        vf->priv->aspect *= ((double)width/height) / ((double)d_width/d_height);
        if (vf->priv->exp_h < vf->priv->exp_w / vf->priv->aspect) {
            vf->priv->exp_h = vf->priv->exp_w / vf->priv->aspect + 0.5;
        } else {
            vf->priv->exp_w = vf->priv->exp_h * vf->priv->aspect + 0.5;
        }
    }
    if (vf->priv->round > 1) { // round up.
        vf->priv->exp_w = (1 + (vf->priv->exp_w - 1) / vf->priv->round) * vf->priv->round;
        vf->priv->exp_h = (1 + (vf->priv->exp_h - 1) / vf->priv->round) * vf->priv->round;
    }

    if(vf->priv->exp_x<0 || vf->priv->exp_x+width>vf->priv->exp_w) vf->priv->exp_x=(vf->priv->exp_w-width)/2;
    if(vf->priv->exp_y<0 || vf->priv->exp_y+height>vf->priv->exp_h) vf->priv->exp_y=(vf->priv->exp_h-height)/2;
    vf->priv->fb_ptr=NULL;

    if(!opt_screen_size_x && !opt_screen_size_y){
	d_width=d_width*vf->priv->exp_w/width;
	d_height=d_height*vf->priv->exp_h/height;
    }
    return vf_next_config(vf,vf->priv->exp_w,vf->priv->exp_h,d_width,d_height,flags,outfmt);
}

// there are 4 cases:
// codec --DR--> expand --DR--> vo
// codec --DR--> expand -copy-> vo
// codec -copy-> expand --DR--> vo
// codec -copy-> expand -copy-> vo (worst case)

static void get_image(struct vf_instance_s* vf, mp_image_t *mpi){
//    if(mpi->type==MP_IMGTYPE_IPB) return; // not yet working
#ifdef OSD_SUPPORT
    if(vf->priv->osd && (mpi->flags&MP_IMGFLAG_PRESERVE)){
	// check if we have to render osd!
	vo_update_osd(vf->priv->exp_w, vf->priv->exp_h);
	if(vo_osd_check_range_update(vf->priv->exp_x,vf->priv->exp_y,
	    vf->priv->exp_x+mpi->w,vf->priv->exp_y+mpi->h)) return;
    }
#endif
    if(vf->priv->exp_w==mpi->width ||
       (mpi->flags&(MP_IMGFLAG_ACCEPT_STRIDE|MP_IMGFLAG_ACCEPT_WIDTH)) ){
	// try full DR !
	mpi->priv=vf->dmpi=vf_get_image(vf->next,mpi->imgfmt,
	    mpi->type, mpi->flags, 
            MAX(vf->priv->exp_w, mpi->width +vf->priv->exp_x), 
            MAX(vf->priv->exp_h, mpi->height+vf->priv->exp_y));
#if 1
	if((vf->dmpi->flags & MP_IMGFLAG_DRAW_CALLBACK) &&
	  !(vf->dmpi->flags & MP_IMGFLAG_DIRECT)){
	    printf("Full DR not possible, trying SLICES instead!\n");
	    return;
	}
#endif
	// set up mpi as a cropped-down image of dmpi:
	if(mpi->flags&MP_IMGFLAG_PLANAR){
	    mpi->planes[0]=vf->dmpi->planes[0]+
		vf->priv->exp_y*vf->dmpi->stride[0]+vf->priv->exp_x;
	    mpi->planes[1]=vf->dmpi->planes[1]+
		(vf->priv->exp_y>>mpi->chroma_y_shift)*vf->dmpi->stride[1]+(vf->priv->exp_x>>mpi->chroma_x_shift);
	    mpi->planes[2]=vf->dmpi->planes[2]+
		(vf->priv->exp_y>>mpi->chroma_y_shift)*vf->dmpi->stride[2]+(vf->priv->exp_x>>mpi->chroma_x_shift);
	    mpi->stride[1]=vf->dmpi->stride[1];
	    mpi->stride[2]=vf->dmpi->stride[2];
	} else {
	    mpi->planes[0]=vf->dmpi->planes[0]+
		vf->priv->exp_y*vf->dmpi->stride[0]+
		vf->priv->exp_x*(vf->dmpi->bpp/8);
	}
	mpi->stride[0]=vf->dmpi->stride[0];
	mpi->width=vf->dmpi->width;
	mpi->flags|=MP_IMGFLAG_DIRECT;
	mpi->flags&=~MP_IMGFLAG_DRAW_CALLBACK;
//	vf->dmpi->flags&=~MP_IMGFLAG_DRAW_CALLBACK;
    }
}

static void start_slice(struct vf_instance_s* vf, mp_image_t *mpi){
//    printf("start_slice called! flag=%d\n",mpi->flags&MP_IMGFLAG_DRAW_CALLBACK);
    if(!vf->next->draw_slice){
	mpi->flags&=~MP_IMGFLAG_DRAW_CALLBACK;
	return;
    }
    // they want slices!!! allocate the buffer.
    if(!mpi->priv)
	mpi->priv=vf->dmpi=vf_get_image(vf->next,mpi->imgfmt,
//	MP_IMGTYPE_TEMP, MP_IMGFLAG_ACCEPT_STRIDE | MP_IMGFLAG_PREFER_ALIGNED_STRIDE,
	    MP_IMGTYPE_TEMP, mpi->flags, 
            MAX(vf->priv->exp_w, mpi->width +vf->priv->exp_x), 
            MAX(vf->priv->exp_h, mpi->height+vf->priv->exp_y));
    if(!(vf->dmpi->flags&MP_IMGFLAG_DRAW_CALLBACK))
	printf("WARNING! next filter doesn't support SLICES, get ready for sig11...\n"); // shouldn't happen.
    vf->priv->first_slice = 1;
}

static void draw_top_blackbar_slice(struct vf_instance_s* vf,
				    unsigned char** src, int* stride, int w,int h, int x, int y){
    if(vf->priv->exp_y>0 && y == 0) {
	vf_next_draw_slice(vf, vf->dmpi->planes, vf->dmpi->stride,
			   vf->dmpi->w,vf->priv->exp_y,0,0);
    }
    
}

static void draw_bottom_blackbar_slice(struct vf_instance_s* vf,
				    unsigned char** src, int* stride, int w,int h, int x, int y){
    if(vf->priv->exp_y+vf->h<vf->dmpi->h && y+h == vf->h) {
	unsigned char *src2[MP_MAX_PLANES];
	src2[0] = vf->dmpi->planes[0]
		+ (vf->priv->exp_y+vf->h)*vf->dmpi->stride[0];
	if(vf->dmpi->flags&MP_IMGFLAG_PLANAR){
	    src2[1] = vf->dmpi->planes[1]
		+ ((vf->priv->exp_y+vf->h)>>vf->dmpi->chroma_y_shift)*vf->dmpi->stride[1];
	    src2[2] = vf->dmpi->planes[2]
		+ ((vf->priv->exp_y+vf->h)>>vf->dmpi->chroma_y_shift)*vf->dmpi->stride[2];
	} else {
	    src2[1] = vf->dmpi->planes[1]; // passthrough rgb8 palette
	}
	vf_next_draw_slice(vf, src2, vf->dmpi->stride,
			   vf->dmpi->w,vf->dmpi->h-(vf->priv->exp_y+vf->h),
			   0,vf->priv->exp_y+vf->h);
    }
}

static void draw_slice(struct vf_instance_s* vf,
        unsigned char** src, int* stride, int w,int h, int x, int y){
//    printf("draw_slice() called %d at %d\n",h,y);
    
    if (y == 0 && y+h == vf->h) {
	// special case - only one slice
	draw_top_blackbar_slice(vf, src, stride, w, h, x, y);
	vf_next_draw_slice(vf,src,stride,w,h,x+vf->priv->exp_x,y+vf->priv->exp_y);
	draw_bottom_blackbar_slice(vf, src, stride, w, h, x, y);
	return;
    }
    if (vf->priv->first_slice) {
	draw_top_blackbar_slice(vf, src, stride, w, h, x, y);
	draw_bottom_blackbar_slice(vf, src, stride, w, h, x, y);
    }
    vf_next_draw_slice(vf,src,stride,w,h,x+vf->priv->exp_x,y+vf->priv->exp_y);
    if (!vf->priv->first_slice) {
	draw_top_blackbar_slice(vf, src, stride, w, h, x, y);
	draw_bottom_blackbar_slice(vf, src, stride, w, h, x, y);
    }
    vf->priv->first_slice = 0;
}

static int put_image(struct vf_instance_s* vf, mp_image_t *mpi){
    if(mpi->flags&MP_IMGFLAG_DIRECT || mpi->flags&MP_IMGFLAG_DRAW_CALLBACK){
	vf->dmpi=mpi->priv;
	if(!vf->dmpi) { printf("Why do we get NULL \n"); return 0; }
	mpi->priv=NULL;
#ifdef OSD_SUPPORT
	if(vf->priv->osd) draw_osd(vf,mpi->w,mpi->h);
#endif
	// we've used DR, so we're ready...
	if(!(mpi->flags&MP_IMGFLAG_PLANAR))
	    vf->dmpi->planes[1] = mpi->planes[1]; // passthrough rgb8 palette
	return vf_next_put_image(vf,vf->dmpi);
    }

    // hope we'll get DR buffer:
    vf->dmpi=vf_get_image(vf->next,mpi->imgfmt,
	MP_IMGTYPE_TEMP, MP_IMGFLAG_ACCEPT_STRIDE,
	vf->priv->exp_w, vf->priv->exp_h);
    
    // copy mpi->dmpi...
    if(mpi->flags&MP_IMGFLAG_PLANAR){
	memcpy_pic(vf->dmpi->planes[0]+
	        vf->priv->exp_y*vf->dmpi->stride[0]+vf->priv->exp_x,
		mpi->planes[0], mpi->w, mpi->h,
		vf->dmpi->stride[0],mpi->stride[0]);
	memcpy_pic(vf->dmpi->planes[1]+
		(vf->priv->exp_y>>mpi->chroma_y_shift)*vf->dmpi->stride[1]+(vf->priv->exp_x>>mpi->chroma_x_shift),
		mpi->planes[1], mpi->chroma_width, mpi->chroma_height,
		vf->dmpi->stride[1],mpi->stride[1]);
	memcpy_pic(vf->dmpi->planes[2]+
		(vf->priv->exp_y>>mpi->chroma_y_shift)*vf->dmpi->stride[2]+(vf->priv->exp_x>>mpi->chroma_x_shift),
		mpi->planes[2], mpi->chroma_width, mpi->chroma_height,
		vf->dmpi->stride[2],mpi->stride[2]);
    } else {
	memcpy_pic(vf->dmpi->planes[0]+
	        vf->priv->exp_y*vf->dmpi->stride[0]+vf->priv->exp_x*(vf->dmpi->bpp/8),
		mpi->planes[0], mpi->w*(vf->dmpi->bpp/8), mpi->h,
		vf->dmpi->stride[0],mpi->stride[0]);
	vf->dmpi->planes[1] = mpi->planes[1]; // passthrough rgb8 palette
    }
#ifdef OSD_SUPPORT
    if(vf->priv->osd) draw_osd(vf,mpi->w,mpi->h);
#endif
    return vf_next_put_image(vf,vf->dmpi);
}

//===========================================================================//

static int control(struct vf_instance_s* vf, int request, void* data){
#ifdef OSD_SUPPORT
    switch(request){
    case VFCTRL_DRAW_OSD:
	if(vf->priv->osd) return CONTROL_TRUE;
    }
#endif
    return vf_next_control(vf,request,data);
}

static int open(vf_instance_t *vf, char* args){
    vf->config=config;
    vf->control=control;
    vf->start_slice=start_slice;
    vf->draw_slice=draw_slice;
    vf->get_image=get_image;
    vf->put_image=put_image;
    mp_msg(MSGT_VFILTER, MSGL_INFO, "Expand: %d x %d, %d ; %d, osd: %d, aspect: %lf, round: %d\n",
    vf->priv->exp_w,
    vf->priv->exp_h,
    vf->priv->exp_x,
    vf->priv->exp_y,
    vf->priv->osd,
    vf->priv->aspect,
    vf->priv->round);
    return 1;
}

#define ST_OFF(f) M_ST_OFF(struct vf_priv_s,f)
static m_option_t vf_opts_fields[] = {
  {"w", ST_OFF(exp_w), CONF_TYPE_INT, 0, 0 ,0, NULL},
  {"h", ST_OFF(exp_h), CONF_TYPE_INT, 0, 0 ,0, NULL},
  {"x", ST_OFF(exp_x), CONF_TYPE_INT, M_OPT_MIN, -1, 0, NULL},
  {"y", ST_OFF(exp_y), CONF_TYPE_INT, M_OPT_MIN, -1, 0, NULL},
  {"osd", ST_OFF(osd), CONF_TYPE_FLAG, 0 , 0, 1, NULL},
  {"aspect", ST_OFF(aspect), CONF_TYPE_DOUBLE, M_OPT_MIN, 0, 0, NULL},
  {"round", ST_OFF(round), CONF_TYPE_INT, M_OPT_MIN, 1, 0, NULL},
  { NULL, NULL, 0, 0, 0, 0,  NULL }
};

static m_struct_t vf_opts = {
  "expand",
  sizeof(struct vf_priv_s),
  &vf_priv_dflt,
  vf_opts_fields
};



vf_info_t vf_info_expand = {
#ifdef OSD_SUPPORT
    "expanding & osd",
#else
    "expanding",
#endif
    "expand",
    "A'rpi",
    "",
    open,
    &vf_opts
};

//===========================================================================//