annotate libmpcodecs/vf_scale.c @ 12925:2abd8e24eb69

10l to the author of this longstanding and obscure bug. Each language should be removed only once from the list. Thanks to Chris White for pointing out that there was a problem.

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

#include "../config.h"
#include "../mp_msg.h"
#include "../cpudetect.h"

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

#include "../libvo/fastmemcpy.h"
#include "../postproc/swscale.h"
#include "vf_scale.h"

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

static struct vf_priv_s {
    int w,h;
    int v_chr_drop;
    int param;
    unsigned int fmt;
    struct SwsContext *ctx;
    struct SwsContext *ctx2; //for interlaced slices only
    unsigned char* palette;
    int interlaced;
    int query_format_cache[64];
} vf_priv_dflt = {
  -1,-1,
  0,
  0,
  0,
  NULL,
  NULL,
  NULL
};

extern int opt_screen_size_x;
extern int opt_screen_size_y;

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

void sws_getFlagsAndFilterFromCmdLine(int *flags, SwsFilter **srcFilterParam, SwsFilter **dstFilterParam);

static unsigned int outfmt_list[]={
// YUV:
    IMGFMT_444P,
    IMGFMT_422P,
    IMGFMT_YV12,
    IMGFMT_I420,
    IMGFMT_IYUV,
    IMGFMT_YVU9,
    IMGFMT_IF09,
    IMGFMT_411P,
    IMGFMT_YUY2,
    IMGFMT_UYVY,
// RGB and grayscale (Y8 and Y800):
    IMGFMT_BGR32,
    IMGFMT_RGB32,
    IMGFMT_BGR24,
    IMGFMT_RGB24,
    IMGFMT_BGR16,
    IMGFMT_RGB16,
    IMGFMT_BGR15,
    IMGFMT_RGB15,
    IMGFMT_Y800,
    IMGFMT_Y8,
    IMGFMT_BGR8,
    IMGFMT_RGB8,
    IMGFMT_BGR4,
    IMGFMT_RGB4,
    IMGFMT_BG4B,
    IMGFMT_RG4B,
    IMGFMT_BGR1,
    IMGFMT_RGB1,
    0
};

static unsigned int find_best_out(vf_instance_t *vf){
    unsigned int best=0;
    int i;

    // find the best outfmt:
    for(i=0; i<sizeof(outfmt_list)/sizeof(int)-1; i++){
        const int format= outfmt_list[i];
        int ret= vf->priv->query_format_cache[i]-1;
        if(ret == -1){
            ret= vf_next_query_format(vf, outfmt_list[i]);
            vf->priv->query_format_cache[i]= ret+1;
        }
        
	mp_msg(MSGT_VFILTER,MSGL_DBG2,"scale: query(%s) -> %d\n",vo_format_name(format),ret&3);
	if(ret&VFCAP_CSP_SUPPORTED_BY_HW){
            best=format; // no conversion -> bingo!
            break;
        } 
	if(ret&VFCAP_CSP_SUPPORTED && !best) 
            best=format; // best with conversion
    }
    return best;
}

static int config(struct vf_instance_s* vf,
        int width, int height, int d_width, int d_height,
	unsigned int flags, unsigned int outfmt){
    unsigned int best=find_best_out(vf);
    int vo_flags;
    int int_sws_flags=0;
    SwsFilter *srcFilter, *dstFilter;
    
    if(!best){
	mp_msg(MSGT_VFILTER,MSGL_WARN,"SwScale: no supported outfmt found :(\n");
	return 0;
    }
    
    vo_flags=vf->next->query_format(vf->next,best);
    
    // scaling to dwidth*d_height, if all these TRUE:
    // - option -zoom
    // - no other sw/hw up/down scaling avail.
    // - we're after postproc
    // - user didn't set w:h
    if(!(vo_flags&VFCAP_POSTPROC) && (flags&4) && 
	    vf->priv->w<0 && vf->priv->h<0){	// -zoom
	int x=(vo_flags&VFCAP_SWSCALE) ? 0 : 1;
	if(d_width<width || d_height<height){
	    // downscale!
	    if(vo_flags&VFCAP_HWSCALE_DOWN) x=0;
	} else {
	    // upscale:
	    if(vo_flags&VFCAP_HWSCALE_UP) x=0;
	}
	if(x){
	    // user wants sw scaling! (-zoom)
	    vf->priv->w=d_width;
	    vf->priv->h=d_height;
	}
    }

    // calculate the missing parameters:
    switch(best) {
    case IMGFMT_YUY2:		/* YUY2 needs w rounded to 2 */
    case IMGFMT_UYVY:
	if(vf->priv->w==-3) vf->priv->w=(vf->priv->h*width/height+1)&~1; else
	if(vf->priv->w==-2) vf->priv->w=(vf->priv->h*d_width/d_height+1)&~1;
	if(vf->priv->w<0) vf->priv->w=width; else
	if(vf->priv->w==0) vf->priv->w=d_width;
	if(vf->priv->h==-3) vf->priv->h=vf->priv->w*height/width; else
	if(vf->priv->h==-2) vf->priv->h=vf->priv->w*d_height/d_width;
	break;
    case IMGFMT_YV12:		/* YV12 needs w & h rounded to 2 */
    case IMGFMT_I420:
    case IMGFMT_IYUV:
	if(vf->priv->w==-3) vf->priv->w=(vf->priv->h*width/height+1)&~1; else
	if(vf->priv->w==-2) vf->priv->w=(vf->priv->h*d_width/d_height+1)&~1;
	if(vf->priv->w<0) vf->priv->w=width; else
	if(vf->priv->w==0) vf->priv->w=d_width;
	if(vf->priv->h==-3) vf->priv->h=(vf->priv->w*height/width+1)&~1; else
	if(vf->priv->h==-2) vf->priv->h=(vf->priv->w*d_height/d_width+1)&~1;
	break;
    default:
    if(vf->priv->w==-3) vf->priv->w=vf->priv->h*width/height; else
    if(vf->priv->w==-2) vf->priv->w=vf->priv->h*d_width/d_height;
    if(vf->priv->w<0) vf->priv->w=width; else
    if(vf->priv->w==0) vf->priv->w=d_width;
    if(vf->priv->h==-3) vf->priv->h=vf->priv->w*height/width; else
    if(vf->priv->h==-2) vf->priv->h=vf->priv->w*d_height/d_width;
    break;
    }
    
    if(vf->priv->h<0) vf->priv->h=height; else
    if(vf->priv->h==0) vf->priv->h=d_height;
    
    mp_msg(MSGT_VFILTER,MSGL_DBG2,"SwScale: scaling %dx%d %s to %dx%d %s  \n",
	width,height,vo_format_name(outfmt),
	vf->priv->w,vf->priv->h,vo_format_name(best));

    // free old ctx:
    if(vf->priv->ctx) sws_freeContext(vf->priv->ctx);
    if(vf->priv->ctx2)sws_freeContext(vf->priv->ctx2);
    
    // new swscaler:
    sws_getFlagsAndFilterFromCmdLine(&int_sws_flags, &srcFilter, &dstFilter);
    int_sws_flags|= vf->priv->v_chr_drop << SWS_SRC_V_CHR_DROP_SHIFT;
    int_sws_flags|= vf->priv->param      << SWS_PARAM_SHIFT;
    vf->priv->ctx=sws_getContext(width, height >> vf->priv->interlaced,
	    outfmt,
		  vf->priv->w, vf->priv->h >> vf->priv->interlaced,
	    best,
	    int_sws_flags | get_sws_cpuflags(), srcFilter, dstFilter);
    if(vf->priv->interlaced){
        vf->priv->ctx2=sws_getContext(width, height >> 1,
	    outfmt,
		  vf->priv->w, vf->priv->h >> 1,
	    best,
	    int_sws_flags | get_sws_cpuflags(), srcFilter, dstFilter);
    }
    if(!vf->priv->ctx){
	// error...
	mp_msg(MSGT_VFILTER,MSGL_WARN,"Couldn't init SwScaler for this setup\n");
	return 0;
    }
    vf->priv->fmt=best;

    if(vf->priv->palette){
	free(vf->priv->palette);
	vf->priv->palette=NULL;
    }
    switch(best){
    case IMGFMT_BGR8: {
      /* set 332 palette for 8 bpp */
	int i;
	vf->priv->palette=malloc(4*256);
	for(i=0; i<256; i++){
	    vf->priv->palette[4*i+0]=4*(i&3)*21;
	    vf->priv->palette[4*i+1]=4*((i>>2)&7)*9;
	    vf->priv->palette[4*i+2]=4*((i>>5)&7)*9;
	}
	break; }
    case IMGFMT_BGR4: 
    case IMGFMT_BG4B: {
	int i;
	vf->priv->palette=malloc(4*16);
	for(i=0; i<16; i++){
	    vf->priv->palette[4*i+0]=4*(i&1)*63;
	    vf->priv->palette[4*i+1]=4*((i>>1)&3)*21;
	    vf->priv->palette[4*i+2]=4*((i>>3)&1)*63;
	}
	break; }
    }

    if(!opt_screen_size_x && !opt_screen_size_y){
	// Compute new d_width and d_height, preserving aspect
	// while ensuring that both are >= output size in pixels.
	if (vf->priv->h * d_width > vf->priv->w * d_height) {
		d_width = vf->priv->h * d_width / d_height;
		d_height = vf->priv->h;
	} else {
		d_height = vf->priv->w * d_height / d_width;
		d_width = vf->priv->w;
	}
	//d_width=d_width*vf->priv->w/width;
	//d_height=d_height*vf->priv->h/height;
    }
    return vf_next_config(vf,vf->priv->w,vf->priv->h,d_width,d_height,flags,best);
}

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(!(mpi->flags&MP_IMGFLAG_DRAW_CALLBACK)) return; // shouldn't happen
    // they want slices!!! allocate the buffer.
    mpi->priv=vf->dmpi=vf_get_image(vf->next,vf->priv->fmt,
//	mpi->type, mpi->flags & (~MP_IMGFLAG_DRAW_CALLBACK),
	MP_IMGTYPE_TEMP, MP_IMGFLAG_ACCEPT_STRIDE | MP_IMGFLAG_PREFER_ALIGNED_STRIDE,
	vf->priv->w, vf->priv->h);
}

static void scale(struct SwsContext *sws1, struct SwsContext *sws2, uint8_t *src[3], int src_stride[3], int y, int h, 
                  uint8_t *dst[3], int dst_stride[3], int interlaced){
    if(interlaced){
        int i;
        uint8_t *src2[3]={src[0], src[1], src[2]};
        uint8_t *dst2[3]={dst[0], dst[1], dst[2]};
        int src_stride2[3]={2*src_stride[0], 2*src_stride[1], 2*src_stride[2]};
        int dst_stride2[3]={2*dst_stride[0], 2*dst_stride[1], 2*dst_stride[2]};

        sws_scale_ordered(sws1, src2, src_stride2, y>>1, h>>1, dst2, dst_stride2);
        for(i=0; i<3; i++){
            src2[i] += src_stride[i];
            dst2[i] += dst_stride[i];
        }
        sws_scale_ordered(sws2, src2, src_stride2, y>>1, h>>1, dst2, dst_stride2);
    }else{
        sws_scale_ordered(sws1, src, src_stride, y, h, dst, dst_stride);
    }                  
}

static void draw_slice(struct vf_instance_s* vf,
        unsigned char** src, int* stride, int w,int h, int x, int y){
    mp_image_t *dmpi=vf->dmpi;
    if(!dmpi){
	mp_msg(MSGT_VFILTER,MSGL_FATAL,"vf_scale: draw_slice() called with dmpi=NULL (no get_image??)\n");
	return;
    }
//    printf("vf_scale::draw_slice() y=%d h=%d\n",y,h);
    scale(vf->priv->ctx, vf->priv->ctx2, src, stride, y, h, dmpi->planes, dmpi->stride, vf->priv->interlaced);
}

static int put_image(struct vf_instance_s* vf, mp_image_t *mpi){
    mp_image_t *dmpi=mpi->priv;

//    printf("vf_scale::put_image(): processing whole frame! dmpi=%p flag=%d\n",
//	dmpi, (mpi->flags&MP_IMGFLAG_DRAW_CALLBACK));
    
  if(!(mpi->flags&MP_IMGFLAG_DRAW_CALLBACK && dmpi)){
  
    // hope we'll get DR buffer:
    dmpi=vf_get_image(vf->next,vf->priv->fmt,
	MP_IMGTYPE_TEMP, MP_IMGFLAG_ACCEPT_STRIDE | MP_IMGFLAG_PREFER_ALIGNED_STRIDE,
	vf->priv->w, vf->priv->h);
    
      scale(vf->priv->ctx, vf->priv->ctx, mpi->planes,mpi->stride,0,mpi->h,dmpi->planes,dmpi->stride, vf->priv->interlaced);
  }

    if(vf->priv->w==mpi->w && vf->priv->h==mpi->h){
	// just conversion, no scaling -> keep postprocessing data
	// this way we can apply pp filter to non-yv12 source using scaler
        vf_clone_mpi_attributes(dmpi, mpi);
    }

    if(vf->priv->palette) dmpi->planes[1]=vf->priv->palette; // export palette!
    
    return vf_next_put_image(vf,dmpi);
}

static int control(struct vf_instance_s* vf, int request, void* data){
    int *table;
    int *inv_table;
    int r;
    int brightness, contrast, saturation, srcRange, dstRange;
    vf_equalizer_t *eq;

  if(vf->priv->ctx)
    switch(request){
    case VFCTRL_GET_EQUALIZER:
	r= sws_getColorspaceDetails(vf->priv->ctx, &inv_table, &srcRange, &table, &dstRange, &brightness, &contrast, &saturation);
	if(r<0) break;

	eq = data;
	if (!strcmp(eq->item,"brightness")) {
		eq->value =  ((brightness*100) + (1<<15))>>16;
	}
	else if (!strcmp(eq->item,"contrast")) {
		eq->value = (((contrast  *100) + (1<<15))>>16) - 100;
	}
	else if (!strcmp(eq->item,"saturation")) {
		eq->value = (((saturation*100) + (1<<15))>>16) - 100;
	}
	else
		break;
	return CONTROL_TRUE;
    case VFCTRL_SET_EQUALIZER:
	r= sws_getColorspaceDetails(vf->priv->ctx, &inv_table, &srcRange, &table, &dstRange, &brightness, &contrast, &saturation);
	if(r<0) break;
//printf("set %f %f %f\n", brightness/(float)(1<<16), contrast/(float)(1<<16), saturation/(float)(1<<16));
	eq = data;

	if (!strcmp(eq->item,"brightness")) {
		brightness = (( eq->value     <<16) + 50)/100;
	}
	else if (!strcmp(eq->item,"contrast")) {
		contrast   = (((eq->value+100)<<16) + 50)/100;
	}
	else if (!strcmp(eq->item,"saturation")) {
		saturation = (((eq->value+100)<<16) + 50)/100;
	}
	else
		break;

	r= sws_setColorspaceDetails(vf->priv->ctx, inv_table, srcRange, table, dstRange, brightness, contrast, saturation);
	if(r<0) break;
	if(vf->priv->ctx2){
            r= sws_setColorspaceDetails(vf->priv->ctx2, inv_table, srcRange, table, dstRange, brightness, contrast, saturation);
            if(r<0) break;
        }

	return CONTROL_TRUE;
    default:
	break;
    }
    
    return vf_next_control(vf,request,data);
}

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

//  supported Input formats: YV12, I420, IYUV, YUY2, UYVY, BGR32, BGR24, BGR16, BGR15, RGB32, RGB24, Y8, Y800

static int query_format(struct vf_instance_s* vf, unsigned int fmt){
    switch(fmt){
    case IMGFMT_YV12:
    case IMGFMT_I420:
    case IMGFMT_IYUV:
    case IMGFMT_UYVY:
    case IMGFMT_YUY2:
    case IMGFMT_BGR32:
    case IMGFMT_BGR24:
    case IMGFMT_BGR16:
    case IMGFMT_BGR15:
    case IMGFMT_RGB32:
    case IMGFMT_RGB24:
    case IMGFMT_Y800: 
    case IMGFMT_Y8: 
    case IMGFMT_YVU9: 
    case IMGFMT_IF09: 
    case IMGFMT_444P: 
    case IMGFMT_422P: 
    case IMGFMT_411P: 
    {
	unsigned int best=find_best_out(vf);
	int flags;
	if(!best) return 0;	 // no matching out-fmt
	flags=vf_next_query_format(vf,best);
	if(!(flags&3)) return 0; // huh?
	if(fmt!=best) flags&=~VFCAP_CSP_SUPPORTED_BY_HW;
	// do not allow scaling, if we are before the PP fliter!
	if(!(flags&VFCAP_POSTPROC)) flags|=VFCAP_SWSCALE;
	return flags;
      }
    }
    return 0;	// nomatching in-fmt
}

static void uninit(struct vf_instance_s *vf){
    if(vf->priv->ctx) sws_freeContext(vf->priv->ctx);
    if(vf->priv->ctx2) sws_freeContext(vf->priv->ctx2);
    if(vf->priv->palette) free(vf->priv->palette);
    free(vf->priv);
}

static int open(vf_instance_t *vf, char* args){
    vf->config=config;
    vf->start_slice=start_slice;
    vf->draw_slice=draw_slice;
    vf->put_image=put_image;
    vf->query_format=query_format;
    vf->control= control;
    vf->uninit=uninit;
    if(!vf->priv) {
    vf->priv=malloc(sizeof(struct vf_priv_s));
    // TODO: parse args ->
    vf->priv->ctx=NULL;
    vf->priv->ctx2=NULL;
    vf->priv->w=
    vf->priv->h=-1;
    vf->priv->v_chr_drop=0;
    vf->priv->param=0;
    vf->priv->palette=NULL;
    } // if(!vf->priv)
    if(args) sscanf(args, "%d:%d:%d:%d",
    &vf->priv->w,
    &vf->priv->h,
    &vf->priv->v_chr_drop,
    &vf->priv->param);
    mp_msg(MSGT_VFILTER,MSGL_V,"SwScale params: %d x %d (-1=no scaling)\n",
    vf->priv->w,
    vf->priv->h);
    
    return 1;
}

//global sws_flags from the command line
int sws_flags=2;

//global srcFilter
static SwsFilter *src_filter= NULL;

float sws_lum_gblur= 0.0;
float sws_chr_gblur= 0.0;
int sws_chr_vshift= 0;
int sws_chr_hshift= 0;
float sws_chr_sharpen= 0.0;
float sws_lum_sharpen= 0.0;

int get_sws_cpuflags(){
    return 
          (gCpuCaps.hasMMX   ? SWS_CPU_CAPS_MMX   : 0)
	| (gCpuCaps.hasMMX2  ? SWS_CPU_CAPS_MMX2  : 0)
	| (gCpuCaps.has3DNow ? SWS_CPU_CAPS_3DNOW : 0)
        | (gCpuCaps.hasAltiVec ? SWS_CPU_CAPS_ALTIVEC : 0);
}

void sws_getFlagsAndFilterFromCmdLine(int *flags, SwsFilter **srcFilterParam, SwsFilter **dstFilterParam)
{
	static int firstTime=1;
	*flags=0;

#ifdef ARCH_X86
	if(gCpuCaps.hasMMX)
		asm volatile("emms\n\t"::: "memory"); //FIXME this shouldnt be required but it IS (even for non mmx versions)
#endif
	if(firstTime)
	{
		firstTime=0;
		*flags= SWS_PRINT_INFO;
	}
	else if(verbose>1) *flags= SWS_PRINT_INFO;

	if(src_filter) sws_freeFilter(src_filter);

	src_filter= sws_getDefaultFilter(
		sws_lum_gblur, sws_chr_gblur,
		sws_lum_sharpen, sws_chr_sharpen,
		sws_chr_hshift, sws_chr_vshift, verbose>1);
        
	switch(sws_flags)
	{
		case 0: *flags|= SWS_FAST_BILINEAR; break;
		case 1: *flags|= SWS_BILINEAR; break;
		case 2: *flags|= SWS_BICUBIC; break;
		case 3: *flags|= SWS_X; break;
		case 4: *flags|= SWS_POINT; break;
		case 5: *flags|= SWS_AREA; break;
		case 6: *flags|= SWS_BICUBLIN; break;
		case 7: *flags|= SWS_GAUSS; break;
		case 8: *flags|= SWS_SINC; break;
		case 9: *flags|= SWS_LANCZOS; break;
		case 10:*flags|= SWS_SPLINE; break;
		default:*flags|= SWS_BILINEAR; break;
	}
	
	*srcFilterParam= src_filter;
	*dstFilterParam= NULL;
}

// will use sws_flags & src_filter (from cmd line)
struct SwsContext *sws_getContextFromCmdLine(int srcW, int srcH, int srcFormat, int dstW, int dstH, int dstFormat)
{
	int flags;
	SwsFilter *dstFilterParam, *srcFilterParam;
	sws_getFlagsAndFilterFromCmdLine(&flags, &srcFilterParam, &dstFilterParam);

	return sws_getContext(srcW, srcH, srcFormat, dstW, dstH, dstFormat, flags | get_sws_cpuflags(), srcFilterParam, dstFilterParam);
}

/// An example of presets usage
static struct size_preset {
  char* name;
  int w, h;
} vf_size_presets_defs[] = {
  // TODO add more 'standard' resolutions
  { "qntsc", 352, 240 },
  { "qpal", 352, 288 },
  { "ntsc", 720, 480 },
  { "pal", 720, 576 },
  { "sntsc", 640, 480 },
  { "spal", 768, 576 },
  { NULL, 0, 0}
};

#define ST_OFF(f) M_ST_OFF(struct size_preset,f)
static m_option_t vf_size_preset_fields[] = {
  {"w", ST_OFF(w), CONF_TYPE_INT, M_OPT_MIN,1 ,0, NULL},
  {"h", ST_OFF(h), CONF_TYPE_INT, M_OPT_MIN,1 ,0, NULL},
  { NULL, NULL, 0, 0, 0, 0,  NULL }
};

static m_struct_t vf_size_preset = {
  "scale_size_preset",
  sizeof(struct size_preset),
  NULL,
  vf_size_preset_fields
};

static m_struct_t vf_opts;
static m_obj_presets_t size_preset = {
  &vf_size_preset, // Input struct desc
  &vf_opts, // Output struct desc
  vf_size_presets_defs, // The list of presets
  ST_OFF(name) // At wich offset is the name field in the preset struct
};

/// Now the options
#undef ST_OFF
#define ST_OFF(f) M_ST_OFF(struct vf_priv_s,f)
static m_option_t vf_opts_fields[] = {
  {"w", ST_OFF(w), CONF_TYPE_INT, M_OPT_MIN,-3 ,0, NULL},
  {"h", ST_OFF(h), CONF_TYPE_INT, M_OPT_MIN,-3 ,0, NULL},
  {"interlaced", ST_OFF(interlaced), CONF_TYPE_INT, M_OPT_RANGE, 0, 1, NULL},
  {"chr-drop", ST_OFF(v_chr_drop), CONF_TYPE_INT, M_OPT_RANGE, 0, 3, NULL},
  {"param", ST_OFF(param), CONF_TYPE_INT, M_OPT_RANGE, 0, 100, NULL},
  // Note that here the 2 field is NULL (ie 0)
  // As we want this option to act on the option struct itself
  {"presize", 0, CONF_TYPE_OBJ_PRESETS, 0, 0, 0, &size_preset},
  { NULL, NULL, 0, 0, 0, 0,  NULL }
};

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

vf_info_t vf_info_scale = {
    "software scaling",
    "scale",
    "A'rpi",
    "",
    open,
    &vf_opts
};

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