mirror of
https://code.videolan.org/videolan/vlc
synced 2024-09-16 16:02:54 +02:00
1995 lines
95 KiB
C
1995 lines
95 KiB
C
/*****************************************************************************
|
|
* panoramix.c : Wall panoramic video with edge blending plugin for vlc
|
|
*****************************************************************************
|
|
* Copyright (C) 2000, 2001, 2002, 2003 VideoLAN
|
|
* $Id: panoramix.c 2006-08-29 16:20:15Z ced $
|
|
*
|
|
* Authors: Cedric Cocquebert <cedric.cocquebert@supelec.fr>
|
|
* based on Samuel Hocevar <sam@zoy.org>
|
|
*
|
|
* 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., 59 Temple Place - Suite 330, Boston, MA 02111, USA.
|
|
*****************************************************************************/
|
|
|
|
/*****************************************************************************
|
|
* Preamble
|
|
*****************************************************************************/
|
|
|
|
#include <vlc/vlc.h>
|
|
#include <vlc_vout.h>
|
|
|
|
#include "filter_common.h"
|
|
|
|
// add by cedric.cocquebert@supelec.fr
|
|
#define OVERLAP 2350
|
|
#ifdef OVERLAP
|
|
#include <math.h>
|
|
// OS CODE DEPENDANT to get display dimensions
|
|
#ifdef SYS_MINGW32
|
|
#include <windows.h>
|
|
#else
|
|
#include <X11/Xlib.h>
|
|
#endif
|
|
#define GAMMA 1
|
|
// #define PACKED_YUV 1
|
|
#define F2(a) ((a)*(a))
|
|
#define F4(a,b,x) ((a)*(F2(x))+((b)*(x)))
|
|
#define ACCURACY 1000
|
|
#define RATIO_MAX 2500
|
|
#define CLIP_01(a) (a < 0.0 ? 0.0 : (a > 1.0 ? 1.0 : a))
|
|
// #define CLIP_0A(a) (a < 0.0 ? 0.0 : (a > ACCURACY ? ACCURACY : a))
|
|
#endif
|
|
|
|
/*****************************************************************************
|
|
* Local prototypes
|
|
*****************************************************************************/
|
|
static int Create ( vlc_object_t * );
|
|
static void Destroy ( vlc_object_t * );
|
|
|
|
static int Init ( vout_thread_t * );
|
|
static void End ( vout_thread_t * );
|
|
#ifdef PACKED_YUV
|
|
static void RenderPackedYUV ( vout_thread_t *, picture_t * );
|
|
#endif
|
|
static void RenderPlanarYUV ( vout_thread_t *, picture_t * );
|
|
static void RenderPackedRGB ( vout_thread_t *, picture_t * );
|
|
|
|
static void RemoveAllVout ( vout_thread_t *p_vout );
|
|
|
|
static int SendEvents( vlc_object_t *, char const *,
|
|
vlc_value_t, vlc_value_t, void * );
|
|
|
|
/*****************************************************************************
|
|
* Module descriptor
|
|
*****************************************************************************/
|
|
#define COLS_TEXT N_("Number of columns")
|
|
#define COLS_LONGTEXT N_("Select the number of horizontal video windows in " \
|
|
"which to split the video")
|
|
|
|
#define ROWS_TEXT N_("Number of rows")
|
|
#define ROWS_LONGTEXT N_("Select the number of vertical video windows in " \
|
|
"which to split the video")
|
|
|
|
#define ACTIVE_TEXT N_("Active windows")
|
|
#define ACTIVE_LONGTEXT N_("Comma separated list of active windows, " \
|
|
"defaults to all")
|
|
|
|
#define CFG_PREFIX "panoramix-"
|
|
|
|
vlc_module_begin();
|
|
set_description( N_("Panoramix: wall with overlap video filter") );
|
|
set_shortname( _("Panoramix" ));
|
|
set_capability( "video filter", 0 );
|
|
set_category( CAT_VIDEO );
|
|
set_subcategory( SUBCAT_VIDEO_VFILTER );
|
|
|
|
add_integer( CFG_PREFIX "cols", -1, NULL,
|
|
COLS_TEXT, COLS_LONGTEXT, VLC_TRUE );
|
|
add_integer( CFG_PREFIX "rows", -1, NULL,
|
|
ROWS_TEXT, ROWS_LONGTEXT, VLC_TRUE );
|
|
|
|
#ifdef OVERLAP
|
|
#define OFFSET_X_TEXT N_("Offset X offset (automatic compensation)")
|
|
#define OFFSET_X_LONGTEXT N_("Select if you want an automatic offset in horizontal (in case of misalignment due to autoratio control)")
|
|
add_bool( CFG_PREFIX "offset-x", 1, NULL, OFFSET_X_TEXT, OFFSET_X_LONGTEXT, VLC_TRUE );
|
|
|
|
#define LENGTH_TEXT N_("length of the overlapping area (in %)")
|
|
#define LENGTH_LONGTEXT N_("Select in percent the length of the blended zone")
|
|
add_integer_with_range( CFG_PREFIX "bz-length", 100, 0, 100, NULL, LENGTH_TEXT, LENGTH_LONGTEXT, VLC_TRUE );
|
|
|
|
#define HEIGHT_TEXT N_("height of the overlapping area (in %)")
|
|
#define HEIGHT_LONGTEXT N_("Select in percent the height of the blended zone (case of 2x2 wall)")
|
|
add_integer_with_range( CFG_PREFIX "bz-height", 100, 0, 100, NULL, HEIGHT_TEXT, HEIGHT_LONGTEXT, VLC_TRUE );
|
|
|
|
#define ATTENUATION_TEXT N_("Attenuation")
|
|
#define ATTENUATION_LONGTEXT N_("Check this option if you want attenuate blended zone by this plug-in (if option is unchecked, attenuate is made by opengl)")
|
|
add_bool( CFG_PREFIX "attenuate", 1, NULL, ATTENUATION_TEXT, ATTENUATION_LONGTEXT, VLC_FALSE );
|
|
|
|
#define BEGIN_TEXT N_("Attenuation, begin (in %)")
|
|
#define BEGIN_LONGTEXT N_("Select in percent the Lagrange coeff of the beginning blended zone")
|
|
add_integer_with_range( CFG_PREFIX "bz-begin", 0, 0, 100, NULL, BEGIN_TEXT, BEGIN_LONGTEXT, VLC_TRUE );
|
|
|
|
#define MIDDLE_TEXT N_("Attenuation, middle (in %)")
|
|
#define MIDDLE_LONGTEXT N_("Select in percent the Lagrange coeff of the middle of blended zone")
|
|
add_integer_with_range( CFG_PREFIX "bz-middle", 50, 0, 100, NULL, MIDDLE_TEXT, MIDDLE_LONGTEXT, VLC_FALSE );
|
|
|
|
#define END_TEXT N_("Attenuation, end (in %)")
|
|
#define END_LONGTEXT N_("Select in percent the Lagrange coeff of the end of blended zone")
|
|
add_integer_with_range( CFG_PREFIX "bz-end", 100, 0, 100, NULL, END_TEXT, END_LONGTEXT, VLC_TRUE );
|
|
|
|
#define MIDDLE_POS_TEXT N_("middle position (in %)")
|
|
#define MIDDLE_POS_LONGTEXT N_("Select in percent (50 is center) the position of the middle point (Lagrange) of blended zone")
|
|
add_integer_with_range( CFG_PREFIX "bz-middle-pos", 50, 1, 99, NULL, MIDDLE_POS_TEXT, MIDDLE_POS_LONGTEXT, VLC_FALSE );
|
|
#ifdef GAMMA
|
|
#define RGAMMA_TEXT N_("Gamma (Red) correction")
|
|
#define RGAMMA_LONGTEXT N_("Select the gamma for the correction of blended zone (Red or Y component)")
|
|
add_float_with_range( CFG_PREFIX "bz-gamma-red", 1, 0, 5, NULL, RGAMMA_TEXT, RGAMMA_LONGTEXT, VLC_TRUE );
|
|
|
|
#define GGAMMA_TEXT N_("Gamma (Green) correction")
|
|
#define GGAMMA_LONGTEXT N_("Select the gamma for the correction of blended zone (Green or U component)")
|
|
add_float_with_range( CFG_PREFIX "bz-gamma-green", 1, 0, 5, NULL, GGAMMA_TEXT, GGAMMA_LONGTEXT, VLC_TRUE );
|
|
|
|
#define BGAMMA_TEXT N_("Gamma (Blue) correction")
|
|
#define BGAMMA_LONGTEXT N_("Select the gamma for the correction of blended zone (Blue or V component)")
|
|
add_float_with_range( CFG_PREFIX "bz-gamma-blue", 1, 0, 5, NULL, BGAMMA_TEXT, BGAMMA_LONGTEXT, VLC_TRUE );
|
|
#endif
|
|
#define RGAMMA_BC_TEXT N_("Black Crush for Red")
|
|
#define RGAMMA_BC_LONGTEXT N_("Select the Black Crush of blended zone (Red or Y component)")
|
|
#define GGAMMA_BC_TEXT N_("Black Crush for Green")
|
|
#define GGAMMA_BC_LONGTEXT N_("Select the Black Crush of blended zone (Green or U component)")
|
|
#define BGAMMA_BC_TEXT N_("Black Crush for Blue")
|
|
#define BGAMMA_BC_LONGTEXT N_("Select the Black Crush of blended zone (Blue or V component)")
|
|
|
|
#define RGAMMA_WC_TEXT N_("White Crush for Red")
|
|
#define RGAMMA_WC_LONGTEXT N_("Select the White Crush of blended zone (Red or Y component)")
|
|
#define GGAMMA_WC_TEXT N_("White Crush for Green")
|
|
#define GGAMMA_WC_LONGTEXT N_("Select the White Crush of blended zone (Green or U component)")
|
|
#define BGAMMA_WC_TEXT N_("White Crush for Blue")
|
|
#define BGAMMA_WC_LONGTEXT N_("Select the White Crush of blended zone (Blue or V component)")
|
|
|
|
#define RGAMMA_BL_TEXT N_("Black Level for Red")
|
|
#define RGAMMA_BL_LONGTEXT N_("Select the Black Level of blended zone (Red or Y component)")
|
|
#define GGAMMA_BL_TEXT N_("Black Level for Green")
|
|
#define GGAMMA_BL_LONGTEXT N_("Select the Black Level of blended zone (Green or U component)")
|
|
#define BGAMMA_BL_TEXT N_("Black Level for Blue")
|
|
#define BGAMMA_BL_LONGTEXT N_("Select the Black Level of blended zone (Blue or V component)")
|
|
|
|
#define RGAMMA_WL_TEXT N_("White Level for Red")
|
|
#define RGAMMA_WL_LONGTEXT N_("Select the White Level of blended zone (Red or Y component)")
|
|
#define GGAMMA_WL_TEXT N_("White Level for Green")
|
|
#define GGAMMA_WL_LONGTEXT N_("Select the White Level of blended zone (Green or U component)")
|
|
#define BGAMMA_WL_TEXT N_("White Level for Blue")
|
|
#define BGAMMA_WL_LONGTEXT N_("Select the White Level of blended zone (Blue or V component)")
|
|
add_integer_with_range( CFG_PREFIX "bz-blackcrush-red", 140, 0, 255, NULL, RGAMMA_BC_TEXT, RGAMMA_BC_LONGTEXT, VLC_TRUE );
|
|
add_integer_with_range( CFG_PREFIX "bz-blackcrush-green", 140, 0, 255, NULL, GGAMMA_BC_TEXT, GGAMMA_BC_LONGTEXT, VLC_TRUE );
|
|
add_integer_with_range( CFG_PREFIX "bz-blackcrush-blue", 140, 0, 255, NULL, BGAMMA_BC_TEXT, BGAMMA_BC_LONGTEXT, VLC_TRUE );
|
|
add_integer_with_range( CFG_PREFIX "bz-whitecrush-red", 200, 0, 255, NULL, RGAMMA_WC_TEXT, RGAMMA_WC_LONGTEXT, VLC_TRUE );
|
|
add_integer_with_range( CFG_PREFIX "bz-whitecrush-green", 200, 0, 255, NULL, GGAMMA_WC_TEXT, GGAMMA_WC_LONGTEXT, VLC_TRUE );
|
|
add_integer_with_range( CFG_PREFIX "bz-whitecrush-blue", 200, 0, 255, NULL, BGAMMA_WC_TEXT, BGAMMA_WC_LONGTEXT, VLC_TRUE );
|
|
add_integer_with_range( CFG_PREFIX "bz-blacklevel-red", 150, 0, 255, NULL, RGAMMA_BL_TEXT, RGAMMA_BL_LONGTEXT, VLC_TRUE );
|
|
add_integer_with_range( CFG_PREFIX "bz-blacklevel-green", 150, 0, 255, NULL, GGAMMA_BL_TEXT, GGAMMA_BL_LONGTEXT, VLC_TRUE );
|
|
add_integer_with_range( CFG_PREFIX "bz-blacklevel-blue", 150, 0, 255, NULL, BGAMMA_BL_TEXT, BGAMMA_BL_LONGTEXT, VLC_TRUE );
|
|
add_integer_with_range( CFG_PREFIX "bz-whitelevel-red", 0, 0, 255, NULL, RGAMMA_WL_TEXT, RGAMMA_WL_LONGTEXT, VLC_TRUE );
|
|
add_integer_with_range( CFG_PREFIX "bz-whitelevel-green", 0, 0, 255, NULL, GGAMMA_WL_TEXT, GGAMMA_WL_LONGTEXT, VLC_TRUE );
|
|
add_integer_with_range( CFG_PREFIX "bz-whitelevel-blue", 0, 0, 255, NULL, BGAMMA_WL_TEXT, BGAMMA_WL_LONGTEXT, VLC_TRUE );
|
|
#ifndef SYS_MINGW32
|
|
#define XINERAMA_TEXT N_("Xinerama option")
|
|
#define XINERAMA_LONGTEXT N_("Uncheck if you have not used xinerama")
|
|
add_bool( CFG_PREFIX "xinerama", 1, NULL, XINERAMA_TEXT, XINERAMA_LONGTEXT, VLC_TRUE );
|
|
#endif
|
|
#endif
|
|
|
|
add_string( CFG_PREFIX "active", NULL, NULL, ACTIVE_TEXT, ACTIVE_LONGTEXT, VLC_TRUE );
|
|
|
|
add_shortcut( "panoramix" );
|
|
set_callbacks( Create, Destroy );
|
|
vlc_module_end();
|
|
|
|
static const char *ppsz_filter_options[] = {
|
|
"cols", "rows", "offset-x", "bz-length", "bz-height", "attenuate",
|
|
"bz-begin", "bz-middle", "bz-end", "bz-middle-pos", "bz-gamma-red",
|
|
"bz-gamma-green", "bz-gamma-blue", "bz-blackcrush-red",
|
|
"bz-blackcrush-green", "bz-blackcrush-blue", "bz-whitecrush-red",
|
|
"bz-whitecrush-green", "bz-whitecrush-blue", "bz-blacklevel-red",
|
|
"bz-blacklevel-green", "bz-blacklevel-blue", "bz-whitelevel-red",
|
|
"bz-whitelevel-green", "bz-whitelevel-blue", "xinerama", "active",
|
|
NULL
|
|
};
|
|
|
|
/*****************************************************************************
|
|
* vout_sys_t: Wall video output method descriptor
|
|
*****************************************************************************
|
|
* This structure is part of the video output thread descriptor.
|
|
* It describes the Wall specific properties of an output thread.
|
|
*****************************************************************************/
|
|
struct vout_sys_t
|
|
{
|
|
#ifdef OVERLAP
|
|
vlc_bool_t b_autocrop;
|
|
vlc_bool_t b_attenuate;
|
|
unsigned int bz_length, bz_height, bz_begin, bz_middle, bz_end, bz_middle_pos;
|
|
unsigned int i_ratio_max;
|
|
unsigned int i_ratio;
|
|
unsigned int a_0, a_1, a_2;
|
|
vlc_bool_t b_has_changed;
|
|
int lambda[2][VOUT_MAX_PLANES][500];
|
|
int cstYUV[2][VOUT_MAX_PLANES][500];
|
|
int lambda2[2][VOUT_MAX_PLANES][500];
|
|
int cstYUV2[2][VOUT_MAX_PLANES][500];
|
|
unsigned int i_halfLength;
|
|
unsigned int i_halfHeight;
|
|
int i_offset_x;
|
|
int i_offset_y;
|
|
#ifdef GAMMA
|
|
float f_gamma_red, f_gamma_green, f_gamma_blue;
|
|
float f_gamma[VOUT_MAX_PLANES];
|
|
uint8_t LUT[VOUT_MAX_PLANES][ACCURACY + 1][256];
|
|
#ifdef PACKED_YUV
|
|
uint8_t LUT2[VOUT_MAX_PLANES][256][500];
|
|
#endif
|
|
#endif
|
|
#ifndef SYS_MINGW32
|
|
vlc_bool_t b_xinerama;
|
|
#endif
|
|
#endif
|
|
int i_col;
|
|
int i_row;
|
|
int i_vout;
|
|
struct vout_list_t
|
|
{
|
|
vlc_bool_t b_active;
|
|
int i_width;
|
|
int i_height;
|
|
vout_thread_t *p_vout;
|
|
} *pp_vout;
|
|
};
|
|
|
|
|
|
|
|
/*****************************************************************************
|
|
* Control: control facility for the vout (forwards to child vout)
|
|
*****************************************************************************/
|
|
static int Control( vout_thread_t *p_vout, int i_query, va_list args )
|
|
{
|
|
int i_row, i_col, i_vout = 0;
|
|
|
|
for( i_row = 0; i_row < p_vout->p_sys->i_row; i_row++ )
|
|
{
|
|
for( i_col = 0; i_col < p_vout->p_sys->i_col; i_col++ )
|
|
{
|
|
vout_vaControl( p_vout->p_sys->pp_vout[ i_vout ].p_vout,
|
|
i_query, args );
|
|
i_vout++;
|
|
}
|
|
}
|
|
return VLC_SUCCESS;
|
|
}
|
|
|
|
/*****************************************************************************
|
|
* Create: allocates Wall video thread output method
|
|
*****************************************************************************
|
|
* This function allocates and initializes a Wall vout method.
|
|
*****************************************************************************/
|
|
static int Create( vlc_object_t *p_this )
|
|
{
|
|
vout_thread_t *p_vout = (vout_thread_t *)p_this;
|
|
char *psz_method, *psz_tmp, *psz_method_tmp;
|
|
int i_vout;
|
|
|
|
/* Allocate structure */
|
|
p_vout->p_sys = malloc( sizeof( vout_sys_t ) );
|
|
if( p_vout->p_sys == NULL )
|
|
{
|
|
msg_Err( p_vout, "out of memory" );
|
|
return VLC_ENOMEM;
|
|
}
|
|
|
|
p_vout->pf_init = Init;
|
|
p_vout->pf_end = End;
|
|
p_vout->pf_manage = NULL;
|
|
/* Color Format not supported
|
|
// Planar Y, packed UV
|
|
case VLC_FOURCC('Y','M','G','A'):
|
|
// Packed YUV 4:2:2, U:Y:V:Y, interlaced
|
|
case VLC_FOURCC('I','U','Y','V'): // packed by 2
|
|
// Packed YUV 2:1:1, Y:U:Y:V
|
|
case VLC_FOURCC('Y','2','1','1'): // packed by 4
|
|
// Packed YUV Reverted
|
|
case VLC_FOURCC('c','y','u','v'): // packed by 2
|
|
*/
|
|
switch (p_vout->render.i_chroma)
|
|
{
|
|
// planar YUV
|
|
case VLC_FOURCC('I','4','4','4'):
|
|
case VLC_FOURCC('I','4','2','2'):
|
|
case VLC_FOURCC('I','4','2','0'):
|
|
case VLC_FOURCC('Y','V','1','2'):
|
|
case VLC_FOURCC('I','Y','U','V'):
|
|
case VLC_FOURCC('I','4','1','1'):
|
|
case VLC_FOURCC('I','4','1','0'):
|
|
case VLC_FOURCC('Y','V','U','9'):
|
|
case VLC_FOURCC('Y','U','V','A'):
|
|
p_vout->pf_render = RenderPlanarYUV;
|
|
break;
|
|
// packed RGB
|
|
case VLC_FOURCC('R','G','B','2'): // packed by 1
|
|
case VLC_FOURCC('R','V','1','5'): // packed by 2
|
|
case VLC_FOURCC('R','V','1','6'): // packed by 2
|
|
case VLC_FOURCC('R','V','2','4'): // packed by 3
|
|
case VLC_FOURCC('R','V','3','2'): // packed by 4
|
|
p_vout->pf_render = RenderPackedRGB;
|
|
break;
|
|
#ifdef PACKED_YUV
|
|
// packed YUV
|
|
case VLC_FOURCC('Y','U','Y','2'): // packed by 2
|
|
case VLC_FOURCC('Y','U','N','V'): // packed by 2
|
|
case VLC_FOURCC('U','Y','V','Y'): // packed by 2
|
|
case VLC_FOURCC('U','Y','N','V'): // packed by 2
|
|
case VLC_FOURCC('Y','4','2','2'): // packed by 2
|
|
p_vout->pf_render = RenderPackedYUV;
|
|
break;
|
|
#endif
|
|
default:
|
|
msg_Err( p_vout, "colorspace not supported by plug-in !!!");
|
|
free( p_vout->p_sys );
|
|
return VLC_ENOMEM;
|
|
}
|
|
p_vout->pf_display = NULL;
|
|
p_vout->pf_control = Control;
|
|
|
|
config_ChainParse( p_vout, CFG_PREFIX, ppsz_filter_options,
|
|
p_vout->p_cfg );
|
|
|
|
/* Look what method was requested */
|
|
p_vout->p_sys->i_col = var_CreateGetInteger( p_vout, CFG_PREFIX "cols" );
|
|
p_vout->p_sys->i_row = var_CreateGetInteger( p_vout, CFG_PREFIX "rows" );
|
|
|
|
// OS dependant code : Autodetect number of displays in wall
|
|
#ifdef SYS_MINGW32
|
|
if ((p_vout->p_sys->i_col < 0) || (p_vout->p_sys->i_row < 0) )
|
|
{
|
|
int nbMonitors = GetSystemMetrics(SM_CMONITORS);
|
|
if (nbMonitors == 1)
|
|
{
|
|
nbMonitors = 5; // 1 display => 5x1 simulation
|
|
p_vout->p_sys->i_col = nbMonitors;
|
|
p_vout->p_sys->i_row = 1;
|
|
}
|
|
else
|
|
{
|
|
p_vout->p_sys->i_col = GetSystemMetrics( SM_CXVIRTUALSCREEN ) / GetSystemMetrics( SM_CXSCREEN );
|
|
p_vout->p_sys->i_row = GetSystemMetrics( SM_CYVIRTUALSCREEN ) / GetSystemMetrics( SM_CYSCREEN );
|
|
if (p_vout->p_sys->i_col * p_vout->p_sys->i_row != nbMonitors)
|
|
{
|
|
p_vout->p_sys->i_col = nbMonitors;
|
|
p_vout->p_sys->i_row = 1;
|
|
}
|
|
}
|
|
var_SetInteger( p_vout, CFG_PREFIX "cols", p_vout->p_sys->i_col);
|
|
var_SetInteger( p_vout, CFG_PREFIX "rows", p_vout->p_sys->i_row);
|
|
}
|
|
#endif
|
|
|
|
#ifdef OVERLAP
|
|
p_vout->p_sys->i_offset_x = var_CreateGetInteger( p_vout, CFG_PREFIX "offset-x" );
|
|
if (p_vout->p_sys->i_col > 2) p_vout->p_sys->i_offset_x = 0; // offset-x is used in case of 2x1 wall & autocrop
|
|
p_vout->p_sys->b_autocrop = !(var_CreateGetInteger( p_vout, "crop-ratio" ) == 0);
|
|
if (!p_vout->p_sys->b_autocrop) p_vout->p_sys->b_autocrop = var_CreateGetInteger( p_vout, "autocrop" );
|
|
p_vout->p_sys->b_attenuate = var_CreateGetInteger( p_vout, CFG_PREFIX "attenuate");
|
|
p_vout->p_sys->bz_length = var_CreateGetInteger( p_vout, CFG_PREFIX "bz-length" );
|
|
if (p_vout->p_sys->i_row > 1)
|
|
p_vout->p_sys->bz_height = var_CreateGetInteger( p_vout, CFG_PREFIX "bz-height" );
|
|
else
|
|
p_vout->p_sys->bz_height = 100;
|
|
p_vout->p_sys->bz_begin = var_CreateGetInteger( p_vout, CFG_PREFIX "bz-begin" );
|
|
p_vout->p_sys->bz_middle = var_CreateGetInteger( p_vout, CFG_PREFIX "bz-middle" );
|
|
p_vout->p_sys->bz_end = var_CreateGetInteger( p_vout, CFG_PREFIX "bz-end" );
|
|
p_vout->p_sys->bz_middle_pos = var_CreateGetInteger( p_vout, CFG_PREFIX "bz-middle-pos" );
|
|
double d_p = 100.0 / p_vout->p_sys->bz_middle_pos;
|
|
p_vout->p_sys->i_ratio_max = var_CreateGetInteger( p_vout, "autocrop-ratio-max" ); // in crop module with autocrop ...
|
|
p_vout->p_sys->i_ratio = var_CreateGetInteger( p_vout, "crop-ratio" ); // in crop module with manual ratio ...
|
|
|
|
p_vout->p_sys->a_2 = d_p * p_vout->p_sys->bz_begin - (double)(d_p * d_p / (d_p - 1)) * p_vout->p_sys->bz_middle + (double)(d_p / (d_p - 1)) * p_vout->p_sys->bz_end;
|
|
p_vout->p_sys->a_1 = -(d_p + 1) * p_vout->p_sys->bz_begin + (double)(d_p * d_p / (d_p - 1)) * p_vout->p_sys->bz_middle - (double)(1 / (d_p - 1)) * p_vout->p_sys->bz_end;
|
|
p_vout->p_sys->a_0 = p_vout->p_sys->bz_begin;
|
|
|
|
#ifdef GAMMA
|
|
p_vout->p_sys->f_gamma_red = var_CreateGetFloat( p_vout, CFG_PREFIX "bz-gamma-red" );
|
|
p_vout->p_sys->f_gamma_green = var_CreateGetFloat( p_vout, CFG_PREFIX "bz-gamma-green" );
|
|
p_vout->p_sys->f_gamma_blue = var_CreateGetFloat( p_vout, CFG_PREFIX "bz-gamma-blue" );
|
|
#endif
|
|
#ifndef SYS_MINGW32
|
|
p_vout->p_sys->b_xinerama= var_CreateGetInteger( p_vout, CFG_PREFIX "xinerama" );
|
|
#endif
|
|
#else
|
|
p_vout->p_sys->i_col = __MAX( 1, __MIN( 15, p_vout->p_sys->i_col ) );
|
|
p_vout->p_sys->i_row = __MAX( 1, __MIN( 15, p_vout->p_sys->i_row ) );
|
|
#endif
|
|
|
|
msg_Dbg( p_vout, "opening a %i x %i wall",
|
|
p_vout->p_sys->i_col, p_vout->p_sys->i_row );
|
|
|
|
p_vout->p_sys->pp_vout = malloc( p_vout->p_sys->i_row *
|
|
p_vout->p_sys->i_col *
|
|
sizeof(struct vout_list_t) );
|
|
if( p_vout->p_sys->pp_vout == NULL )
|
|
{
|
|
msg_Err( p_vout, "out of memory" );
|
|
free( p_vout->p_sys );
|
|
return VLC_ENOMEM;
|
|
}
|
|
|
|
psz_method_tmp =
|
|
psz_method = var_CreateGetNonEmptyString( p_vout, CFG_PREFIX "active" );
|
|
|
|
/* If no trailing vout are specified, take them all */
|
|
if( psz_method == NULL )
|
|
{
|
|
for( i_vout = p_vout->p_sys->i_row * p_vout->p_sys->i_col;
|
|
i_vout--; )
|
|
{
|
|
p_vout->p_sys->pp_vout[i_vout].b_active = 1;
|
|
}
|
|
}
|
|
/* If trailing vout are specified, activate only the requested ones */
|
|
else
|
|
{
|
|
for( i_vout = p_vout->p_sys->i_row * p_vout->p_sys->i_col;
|
|
i_vout--; )
|
|
{
|
|
p_vout->p_sys->pp_vout[i_vout].b_active = 0;
|
|
}
|
|
|
|
while( *psz_method )
|
|
{
|
|
psz_tmp = psz_method;
|
|
while( *psz_tmp && *psz_tmp != ',' )
|
|
{
|
|
psz_tmp++;
|
|
}
|
|
|
|
if( *psz_tmp )
|
|
{
|
|
*psz_tmp = '\0';
|
|
i_vout = atoi( psz_method );
|
|
psz_method = psz_tmp + 1;
|
|
}
|
|
else
|
|
{
|
|
i_vout = atoi( psz_method );
|
|
psz_method = psz_tmp;
|
|
}
|
|
|
|
if( i_vout >= 0 &&
|
|
i_vout < p_vout->p_sys->i_row * p_vout->p_sys->i_col )
|
|
{
|
|
p_vout->p_sys->pp_vout[i_vout].b_active = 1;
|
|
}
|
|
}
|
|
}
|
|
|
|
free( psz_method_tmp );
|
|
|
|
return VLC_SUCCESS;
|
|
}
|
|
|
|
|
|
#ifdef OVERLAP
|
|
/*****************************************************************************
|
|
* CLIP_0A: clip between 0 and ACCURACY
|
|
*****************************************************************************/
|
|
inline static int CLIP_0A( int a )
|
|
{
|
|
return (a > ACCURACY) ? ACCURACY : (a < 0) ? 0 : a;
|
|
}
|
|
|
|
#ifdef GAMMA
|
|
/*****************************************************************************
|
|
* Gamma: Gamma correction
|
|
*****************************************************************************/
|
|
static double Gamma_Correction(int i_plane, float f_component, float f_BlackCrush[VOUT_MAX_PLANES], float f_WhiteCrush[VOUT_MAX_PLANES], float f_BlackLevel[VOUT_MAX_PLANES], float f_WhiteLevel[VOUT_MAX_PLANES], float f_Gamma[VOUT_MAX_PLANES])
|
|
{
|
|
float f_Input;
|
|
|
|
f_Input = (f_component * f_BlackLevel[i_plane]) / (f_BlackCrush[i_plane]) + (1.0 - f_BlackLevel[i_plane]);
|
|
if (f_component <= f_BlackCrush[i_plane])
|
|
return pow(f_Input, 1.0 / f_Gamma[i_plane]);
|
|
else if (f_component >= f_WhiteCrush[i_plane])
|
|
{
|
|
f_Input = (f_component * (1.0 - (f_WhiteLevel[i_plane] + 1.0)) + (f_WhiteLevel[i_plane] + 1.0) * f_WhiteCrush[i_plane] - 1.0) / (f_WhiteCrush[i_plane] - 1.0);
|
|
return pow(f_Input, 1.0 / f_Gamma[i_plane]);
|
|
}
|
|
else
|
|
return 1.0;
|
|
}
|
|
|
|
#ifdef PACKED_YUV
|
|
|
|
/*****************************************************************************
|
|
* F: Function to calculate Gamma correction
|
|
*****************************************************************************/
|
|
static uint8_t F(uint8_t i, float gamma)
|
|
{
|
|
double input = (double) i / 255.0;
|
|
|
|
// return clip(255 * pow(input, 1.0 / gamma));
|
|
|
|
if (input < 0.5)
|
|
return clip_uint8((255 * pow(2 * input, gamma)) / 2);
|
|
else
|
|
return clip_uint8(255 * (1 - pow(2 * (1 - input), gamma) / 2));
|
|
|
|
}
|
|
#endif
|
|
#endif
|
|
|
|
/*****************************************************************************
|
|
* AdjustHeight: ajust p_sys->i_height to have same BZ width for any ratio
|
|
*****************************************************************************/
|
|
static int AdjustHeight( vout_thread_t *p_vout )
|
|
{
|
|
vlc_bool_t b_fullscreen = var_CreateGetInteger( p_vout, "fullscreen" );
|
|
int i_window_width = p_vout->i_window_width;
|
|
int i_window_height = p_vout->i_window_height;
|
|
double d_halfLength = 0;
|
|
double d_halfLength_crop;
|
|
double d_halfLength_calculated;
|
|
int i_offset = 0;
|
|
|
|
// OS DEPENDANT CODE to get display dimensions
|
|
if (b_fullscreen)
|
|
{
|
|
#ifdef SYS_MINGW32
|
|
i_window_width = GetSystemMetrics(SM_CXSCREEN);
|
|
i_window_height = GetSystemMetrics(SM_CYSCREEN);
|
|
#else
|
|
Display *p_display = XOpenDisplay( "" );
|
|
if (p_vout->p_sys->b_xinerama)
|
|
{
|
|
i_window_width = DisplayWidth(p_display, 0) / p_vout->p_sys->i_col;
|
|
i_window_height = DisplayHeight(p_display, 0) / p_vout->p_sys->i_row;
|
|
}
|
|
else
|
|
{
|
|
i_window_width = DisplayWidth(p_display, 0);
|
|
i_window_height = DisplayHeight(p_display, 0);
|
|
}
|
|
XCloseDisplay( p_display );
|
|
free(p_display);
|
|
#endif
|
|
var_SetInteger( p_vout, "width", i_window_width);
|
|
var_SetInteger( p_vout, "height", i_window_height);
|
|
p_vout->i_window_width = i_window_width;
|
|
p_vout->i_window_height = i_window_height;
|
|
}
|
|
|
|
if (p_vout->p_sys->bz_length)
|
|
if ((!p_vout->p_sys->b_autocrop) && (!p_vout->p_sys->i_ratio))
|
|
{
|
|
if ((p_vout->p_sys->i_row > 1) || (p_vout->p_sys->i_col > 1))
|
|
{
|
|
while ((d_halfLength <= 0) || (d_halfLength > p_vout->render.i_width / (2 * p_vout->p_sys->i_col)))
|
|
{
|
|
if (p_vout->p_sys->bz_length >= 50)
|
|
d_halfLength = i_window_width * p_vout->render.i_height / (2 * i_window_height * p_vout->p_sys->i_row) - p_vout->render.i_width / (2 * p_vout->p_sys->i_col);
|
|
else
|
|
{
|
|
d_halfLength = (p_vout->render.i_width * p_vout->p_sys->bz_length) / (100.0 * p_vout->p_sys->i_col);
|
|
d_halfLength = __MAX(i_window_width * p_vout->render.i_height / (2 * i_window_height * p_vout->p_sys->i_row) - p_vout->render.i_width / (2 * p_vout->p_sys->i_col), d_halfLength);
|
|
}
|
|
if ((d_halfLength <= 0) || (d_halfLength > p_vout->render.i_width / (2 * p_vout->p_sys->i_col))) p_vout->p_sys->i_row--;
|
|
if (p_vout->p_sys->i_row < 1 )
|
|
{
|
|
p_vout->p_sys->i_row = 1;
|
|
break;
|
|
}
|
|
}
|
|
p_vout->p_sys->i_halfLength = (d_halfLength + 0.5);
|
|
p_vout->p_sys->bz_length = (p_vout->p_sys->i_halfLength * 100.0 * p_vout->p_sys->i_col) / p_vout->render.i_width;
|
|
var_SetInteger( p_vout, "bz-length", p_vout->p_sys->bz_length);
|
|
var_SetInteger( p_vout, "panoramix-rows", p_vout->p_sys->i_row);
|
|
}
|
|
}
|
|
else
|
|
{
|
|
d_halfLength = ((2 * (double)i_window_width - (double)(p_vout->p_sys->i_ratio_max * i_window_height) / 1000.0 ) * (double)p_vout->p_sys->bz_length) / 200.0;
|
|
d_halfLength_crop = d_halfLength * VOUT_ASPECT_FACTOR * (double)p_vout->output.i_width
|
|
/ (double)i_window_height / (double)p_vout->render.i_aspect;
|
|
p_vout->p_sys->i_halfLength = (d_halfLength_crop + 0.5);
|
|
d_halfLength_calculated = p_vout->p_sys->i_halfLength * (double)i_window_height *
|
|
(double)p_vout->render.i_aspect / VOUT_ASPECT_FACTOR / (double)p_vout->output.i_width;
|
|
|
|
if (!p_vout->p_sys->b_attenuate)
|
|
{
|
|
double d_bz_length = (p_vout->p_sys->i_halfLength * p_vout->p_sys->i_col * 100.0) / p_vout->render.i_width;
|
|
// F(2x) != 2F(x) in opengl module
|
|
if (p_vout->p_sys->i_col == 2) d_bz_length = (100.0 * d_bz_length) / (100.0 - d_bz_length) ;
|
|
var_SetInteger( p_vout, "bz-length", (int)(d_bz_length + 0.5));
|
|
}
|
|
i_offset = (int)d_halfLength - (int)
|
|
(p_vout->p_sys->i_halfLength * (double)i_window_height *
|
|
(double)p_vout->render.i_aspect / VOUT_ASPECT_FACTOR / (double)p_vout->output.i_width);
|
|
}
|
|
else
|
|
d_halfLength = 0;
|
|
|
|
return i_offset;
|
|
}
|
|
#endif
|
|
|
|
|
|
/*****************************************************************************
|
|
* Init: initialize Wall video thread output method
|
|
*****************************************************************************/
|
|
static int Init( vout_thread_t *p_vout )
|
|
{
|
|
int i_index, i_row, i_col, i_width, i_height;
|
|
picture_t *p_pic;
|
|
|
|
I_OUTPUTPICTURES = 0;
|
|
|
|
/* Initialize the output structure */
|
|
p_vout->output.i_chroma = p_vout->render.i_chroma;
|
|
p_vout->output.i_width = p_vout->render.i_width;
|
|
p_vout->output.i_height = p_vout->render.i_height;
|
|
p_vout->output.i_aspect = p_vout->render.i_aspect;
|
|
#ifdef OVERLAP
|
|
p_vout->p_sys->b_has_changed = p_vout->p_sys->b_attenuate;
|
|
int i_video_x = var_GetInteger( p_vout, "video-x");
|
|
int i_video_y = var_GetInteger( p_vout, "video-y");
|
|
#ifdef GAMMA
|
|
if (p_vout->p_sys->b_attenuate)
|
|
{
|
|
int i_index2, i_plane;
|
|
int constantYUV[3] = {0,128,128};
|
|
float f_BlackCrush[VOUT_MAX_PLANES];
|
|
float f_BlackLevel[VOUT_MAX_PLANES];
|
|
float f_WhiteCrush[VOUT_MAX_PLANES];
|
|
float f_WhiteLevel[VOUT_MAX_PLANES];
|
|
p_vout->p_sys->f_gamma[0] = var_CreateGetFloat( p_vout, CFG_PREFIX "bz-gamma-red" );
|
|
p_vout->p_sys->f_gamma[1] = var_CreateGetFloat( p_vout, CFG_PREFIX "bz-gamma-green" );
|
|
p_vout->p_sys->f_gamma[2] = var_CreateGetFloat( p_vout, CFG_PREFIX "bz-gamma-blue" );
|
|
f_BlackCrush[0] = var_CreateGetInteger( p_vout, CFG_PREFIX "bz-blackcrush-red" ) / 255.0;
|
|
f_BlackCrush[1] = var_CreateGetInteger( p_vout, CFG_PREFIX "bz-blackcrush-green" ) / 255.0;
|
|
f_BlackCrush[2] = var_CreateGetInteger( p_vout, CFG_PREFIX "bz-blackcrush-blue" ) / 255.0;
|
|
f_WhiteCrush[0] = var_CreateGetInteger( p_vout, CFG_PREFIX "bz-whitecrush-red" ) / 255.0;
|
|
f_WhiteCrush[1] = var_CreateGetInteger( p_vout, CFG_PREFIX "bz-whitecrush-green" ) / 255.0;
|
|
f_WhiteCrush[2] = var_CreateGetInteger( p_vout, CFG_PREFIX "bz-whitecrush-blue" ) / 255.0;
|
|
f_BlackLevel[0] = var_CreateGetInteger( p_vout, CFG_PREFIX "bz-blacklevel-red" ) / 255.0;
|
|
f_BlackLevel[1] = var_CreateGetInteger( p_vout, CFG_PREFIX "bz-blacklevel-green" ) / 255.0;
|
|
f_BlackLevel[2] = var_CreateGetInteger( p_vout, CFG_PREFIX "bz-blacklevel-blue" ) / 255.0;
|
|
f_WhiteLevel[0] = var_CreateGetInteger( p_vout, CFG_PREFIX "bz-whitelevel-red" ) / 255.0;
|
|
f_WhiteLevel[1] = var_CreateGetInteger( p_vout, CFG_PREFIX "bz-whitelevel-green" ) / 255.0;
|
|
f_WhiteLevel[2] = var_CreateGetInteger( p_vout, CFG_PREFIX "bz-whitelevel-blue" ) / 255.0;
|
|
switch (p_vout->render.i_chroma)
|
|
{
|
|
// planar YVU
|
|
case VLC_FOURCC('Y','V','1','2'):
|
|
case VLC_FOURCC('Y','V','U','9'):
|
|
// packed UYV
|
|
case VLC_FOURCC('U','Y','V','Y'): // packed by 2
|
|
case VLC_FOURCC('U','Y','N','V'): // packed by 2
|
|
case VLC_FOURCC('Y','4','2','2'): // packed by 2
|
|
// case VLC_FOURCC('c','y','u','v'): // packed by 2
|
|
p_vout->p_sys->f_gamma[2] = var_CreateGetFloat( p_vout, CFG_PREFIX "bz-gamma-green" );
|
|
p_vout->p_sys->f_gamma[1] = var_CreateGetFloat( p_vout, CFG_PREFIX "bz-gamma-blue" );
|
|
f_BlackCrush[2] = var_CreateGetInteger( p_vout, CFG_PREFIX "bz-blackcrush-green" ) / 255.0;
|
|
f_BlackCrush[1] = var_CreateGetInteger( p_vout, CFG_PREFIX "bz-blackcrush-blue" ) / 255.0;
|
|
f_WhiteCrush[2] = var_CreateGetInteger( p_vout, CFG_PREFIX "bz-whitecrush-green" ) / 255.0;
|
|
f_WhiteCrush[1] = var_CreateGetInteger( p_vout, CFG_PREFIX "bz-whitecrush-blue" ) / 255.0;
|
|
f_BlackLevel[2] = var_CreateGetInteger( p_vout, CFG_PREFIX "bz-blacklevel-green" ) / 255.0;
|
|
f_BlackLevel[1] = var_CreateGetInteger( p_vout, CFG_PREFIX "bz-blacklevel-blue" ) / 255.0;
|
|
f_WhiteLevel[2] = var_CreateGetInteger( p_vout, CFG_PREFIX "bz-whitelevel-green" ) / 255.0;
|
|
f_WhiteLevel[1] = var_CreateGetInteger( p_vout, CFG_PREFIX "bz-whitelevel-blue" ) / 255.0;
|
|
// planar YUV
|
|
case VLC_FOURCC('I','4','4','4'):
|
|
case VLC_FOURCC('I','4','2','2'):
|
|
case VLC_FOURCC('I','4','2','0'):
|
|
case VLC_FOURCC('I','4','1','1'):
|
|
case VLC_FOURCC('I','4','1','0'):
|
|
case VLC_FOURCC('I','Y','U','V'):
|
|
case VLC_FOURCC('Y','U','V','A'):
|
|
// packed YUV
|
|
case VLC_FOURCC('Y','U','Y','2'): // packed by 2
|
|
case VLC_FOURCC('Y','U','N','V'): // packed by 2
|
|
for (i_index = 0; i_index < 256; i_index++)
|
|
for (i_index2 = 0; i_index2 <= ACCURACY; i_index2++)
|
|
for (i_plane = 0; i_plane < VOUT_MAX_PLANES; i_plane++)
|
|
{
|
|
float f_lut = CLIP_01(1.0 -
|
|
((ACCURACY - (float)i_index2)
|
|
* Gamma_Correction(i_plane, (float)i_index / 255.0, f_BlackCrush, f_WhiteCrush, f_BlackLevel, f_WhiteLevel, p_vout->p_sys->f_gamma)
|
|
/ (ACCURACY - 1)));
|
|
p_vout->p_sys->LUT[i_plane][i_index2][i_index] = f_lut * i_index + (int)((1.0 - f_lut) * (float)constantYUV[i_plane]);
|
|
}
|
|
break;
|
|
// packed RGB
|
|
case VLC_FOURCC('R','G','B','2'): // packed by 1
|
|
case VLC_FOURCC('R','V','1','5'): // packed by 2
|
|
case VLC_FOURCC('R','V','1','6'): // packed by 2
|
|
case VLC_FOURCC('R','V','2','4'): // packed by 3
|
|
case VLC_FOURCC('R','V','3','2'): // packed by 4
|
|
for (i_index = 0; i_index < 256; i_index++)
|
|
for (i_index2 = 0; i_index2 <= ACCURACY; i_index2++)
|
|
for (i_plane = 0; i_plane < VOUT_MAX_PLANES; i_plane++)
|
|
{
|
|
float f_lut = CLIP_01(1.0 -
|
|
((ACCURACY - (float)i_index2)
|
|
* Gamma_Correction(i_plane, (float)i_index / 255.0, f_BlackCrush, f_WhiteCrush, f_BlackLevel, f_WhiteLevel, p_vout->p_sys->f_gamma)
|
|
/ (ACCURACY - 1)));
|
|
p_vout->p_sys->LUT[i_plane][i_index2][i_index] = f_lut * i_index;
|
|
}
|
|
break;
|
|
default:
|
|
msg_Err( p_vout, "colorspace not supported by plug-in !!!");
|
|
free( p_vout->p_sys );
|
|
return VLC_ENOMEM;
|
|
}
|
|
}
|
|
#endif
|
|
if (p_vout->p_sys->i_offset_x)
|
|
p_vout->p_sys->i_offset_x = AdjustHeight(p_vout);
|
|
else
|
|
AdjustHeight(p_vout);
|
|
#endif
|
|
|
|
/* Try to open the real video output */
|
|
msg_Dbg( p_vout, "spawning the real video outputs" );
|
|
|
|
p_vout->p_sys->i_vout = 0;
|
|
|
|
/* FIXME: use bresenham instead of those ugly divisions */
|
|
for( i_row = 0; i_row < p_vout->p_sys->i_row; i_row++ )
|
|
{
|
|
for( i_col = 0; i_col < p_vout->p_sys->i_col; i_col++ )
|
|
{
|
|
video_format_t fmt;
|
|
|
|
memset( &fmt, 0, sizeof(video_format_t) );
|
|
|
|
if( i_col + 1 < p_vout->p_sys->i_col )
|
|
{
|
|
i_width = ( p_vout->render.i_width
|
|
/ p_vout->p_sys->i_col ) & ~0x1;
|
|
}
|
|
else
|
|
{
|
|
i_width = p_vout->render.i_width
|
|
- ( ( p_vout->render.i_width
|
|
/ p_vout->p_sys->i_col ) & ~0x1 ) * i_col;
|
|
|
|
}
|
|
#ifdef OVERLAP
|
|
i_width += p_vout->p_sys->i_halfLength;
|
|
if (p_vout->p_sys->i_col > 2 ) i_width += p_vout->p_sys->i_halfLength;
|
|
i_width -= i_width % 2;
|
|
#endif
|
|
if( i_row + 1 < p_vout->p_sys->i_row )
|
|
{
|
|
i_height = ( p_vout->render.i_height
|
|
/ p_vout->p_sys->i_row ) & ~0x3;
|
|
}
|
|
else
|
|
{
|
|
i_height = p_vout->render.i_height
|
|
- ( ( p_vout->render.i_height
|
|
/ p_vout->p_sys->i_row ) & ~0x3 ) * i_row;
|
|
}
|
|
|
|
#ifdef OVERLAP
|
|
if (p_vout->p_sys->i_row >= 2)
|
|
{
|
|
p_vout->p_sys->i_halfHeight = (p_vout->p_sys->i_halfLength * p_vout->p_sys->bz_height) / 100;
|
|
p_vout->p_sys->i_halfHeight -= (p_vout->p_sys->i_halfHeight % 2);
|
|
i_height += p_vout->p_sys->i_halfHeight;
|
|
if (p_vout->p_sys->i_row > 2) i_height += p_vout->p_sys->i_halfHeight;
|
|
}
|
|
i_height -= i_height % 2;
|
|
#endif
|
|
p_vout->p_sys->pp_vout[ p_vout->p_sys->i_vout ].i_width = i_width;
|
|
p_vout->p_sys->pp_vout[ p_vout->p_sys->i_vout ].i_height = i_height;
|
|
|
|
if( !p_vout->p_sys->pp_vout[ p_vout->p_sys->i_vout ].b_active )
|
|
{
|
|
p_vout->p_sys->i_vout++;
|
|
continue;
|
|
}
|
|
|
|
fmt.i_width = fmt.i_visible_width = p_vout->render.i_width;
|
|
fmt.i_height = fmt.i_visible_height = p_vout->render.i_height;
|
|
fmt.i_x_offset = fmt.i_y_offset = 0;
|
|
fmt.i_chroma = p_vout->render.i_chroma;
|
|
fmt.i_aspect = p_vout->render.i_aspect;
|
|
fmt.i_sar_num = p_vout->render.i_aspect * fmt.i_height / fmt.i_width;
|
|
fmt.i_sar_den = VOUT_ASPECT_FACTOR;
|
|
fmt.i_width = fmt.i_visible_width = i_width;
|
|
fmt.i_height = fmt.i_visible_height = i_height;
|
|
fmt.i_aspect = p_vout->render.i_aspect
|
|
* p_vout->render.i_height / i_height
|
|
* i_width / p_vout->render.i_width;
|
|
#ifdef OVERLAP
|
|
if (p_vout->p_sys->i_offset_x < 0)
|
|
{
|
|
var_SetInteger(p_vout, "video-x", -p_vout->p_sys->i_offset_x);
|
|
p_vout->p_sys->i_offset_x = 0;
|
|
}
|
|
#endif
|
|
p_vout->p_sys->pp_vout[ p_vout->p_sys->i_vout ].p_vout =
|
|
vout_Create( p_vout, &fmt);
|
|
|
|
if( p_vout->p_sys->pp_vout[ p_vout->p_sys->i_vout ].p_vout == NULL )
|
|
{
|
|
msg_Err( p_vout, "failed to get %ix%i vout threads",
|
|
p_vout->p_sys->i_col, p_vout->p_sys->i_row );
|
|
RemoveAllVout( p_vout );
|
|
return VLC_EGENERIC;
|
|
}
|
|
ADD_CALLBACKS(
|
|
p_vout->p_sys->pp_vout[ p_vout->p_sys->i_vout ].p_vout,
|
|
SendEvents );
|
|
#ifdef OVERLAP
|
|
p_vout->p_sys->pp_vout[ p_vout->p_sys->i_vout ].p_vout->i_alignment = 0;
|
|
if (i_col == 0) p_vout->p_sys->pp_vout[ p_vout->p_sys->i_vout ].p_vout->i_alignment |= VOUT_ALIGN_RIGHT;
|
|
else if (i_col == p_vout->p_sys->i_col -1) p_vout->p_sys->pp_vout[ p_vout->p_sys->i_vout ].p_vout->i_alignment |= VOUT_ALIGN_LEFT;
|
|
if (p_vout->p_sys->i_row > 1)
|
|
{
|
|
if (i_row == 0) p_vout->p_sys->pp_vout[ p_vout->p_sys->i_vout ].p_vout->i_alignment |= VOUT_ALIGN_BOTTOM;
|
|
else if (i_row == p_vout->p_sys->i_row -1) p_vout->p_sys->pp_vout[ p_vout->p_sys->i_vout ].p_vout->i_alignment |= VOUT_ALIGN_TOP;
|
|
}
|
|
// i_n : number of active pp_vout
|
|
int i_index, i_n = p_vout->p_sys->i_vout;
|
|
for (i_index = p_vout->p_sys->i_vout; i_index >= 0; i_index--) if (!p_vout->p_sys->pp_vout[i_index].b_active) i_n -= 1;
|
|
var_SetInteger( p_vout, "align", p_vout->p_sys->pp_vout[ p_vout->p_sys->i_vout ].p_vout->i_alignment );
|
|
var_SetInteger( p_vout, "video-x",i_video_x + p_vout->p_sys->i_offset_x + ((i_n + 1) % p_vout->p_sys->i_col) * p_vout->i_window_width);
|
|
var_SetInteger( p_vout, "video-y",i_video_y + ((i_n + 1) / p_vout->p_sys->i_col) * p_vout->i_window_height);
|
|
#endif
|
|
p_vout->p_sys->i_vout++;
|
|
}
|
|
}
|
|
|
|
ALLOCATE_DIRECTBUFFERS( VOUT_MAX_PICTURES );
|
|
|
|
ADD_PARENT_CALLBACKS( SendEventsToChild );
|
|
|
|
return VLC_SUCCESS;
|
|
}
|
|
|
|
/*****************************************************************************
|
|
* End: terminate Wall video thread output method
|
|
*****************************************************************************/
|
|
static void End( vout_thread_t *p_vout )
|
|
{
|
|
int i_index;
|
|
|
|
#ifdef OVERLAP
|
|
var_SetInteger( p_vout, "bz-length", p_vout->p_sys->bz_length);
|
|
#endif
|
|
/* Free the fake output buffers we allocated */
|
|
for( i_index = I_OUTPUTPICTURES ; i_index ; )
|
|
{
|
|
i_index--;
|
|
free( PP_OUTPUTPICTURE[ i_index ]->p_data_orig );
|
|
}
|
|
}
|
|
|
|
/*****************************************************************************
|
|
* Destroy: destroy Wall video thread output method
|
|
*****************************************************************************
|
|
* Terminate an output method created by WallCreateOutputMethod
|
|
*****************************************************************************/
|
|
static void Destroy( vlc_object_t *p_this )
|
|
{
|
|
vout_thread_t *p_vout = (vout_thread_t *)p_this;
|
|
|
|
#ifdef GLOBAL_OUTPUT
|
|
DEL_CALLBACKS( p_vout->p_sys->p_vout, SendEvents);
|
|
vlc_object_detach( p_vout->p_sys->p_vout );
|
|
vout_Destroy( p_vout->p_sys->p_vout );
|
|
DEL_PARENT_CALLBACKS( SendEventsToChild);
|
|
#endif
|
|
|
|
RemoveAllVout( p_vout );
|
|
DEL_PARENT_CALLBACKS( SendEventsToChild );
|
|
|
|
free( p_vout->p_sys->pp_vout );
|
|
free( p_vout->p_sys );
|
|
|
|
}
|
|
|
|
/*****************************************************************************
|
|
* RenderPlanarYUV: displays previously rendered output
|
|
*****************************************************************************
|
|
* This function send the currently rendered image to Wall image, waits
|
|
* until it is displayed and switch the two rendering buffers, preparing next
|
|
* frame.
|
|
*****************************************************************************/
|
|
static void RenderPlanarYUV( vout_thread_t *p_vout, picture_t *p_pic )
|
|
{
|
|
picture_t *p_outpic = NULL;
|
|
int i_col, i_row, i_vout, i_plane;
|
|
int pi_left_skip[VOUT_MAX_PLANES], pi_top_skip[VOUT_MAX_PLANES];
|
|
#ifdef OVERLAP
|
|
int LeftOffset, TopOffset;
|
|
int constantYUV[3] = {0,128,128};
|
|
int Denom;
|
|
int a_2;
|
|
int a_1;
|
|
int a_0;
|
|
int i_index, i_index2;
|
|
#endif
|
|
|
|
|
|
i_vout = 0;
|
|
|
|
for( i_plane = 0 ; i_plane < p_pic->i_planes ; i_plane++ )
|
|
{
|
|
pi_top_skip[i_plane] = 0;
|
|
}
|
|
|
|
for( i_row = 0; i_row < p_vout->p_sys->i_row; i_row++ )
|
|
{
|
|
for( i_plane = 0 ; i_plane < p_pic->i_planes ; i_plane++ )
|
|
{
|
|
pi_left_skip[i_plane] = 0;
|
|
}
|
|
|
|
for( i_col = 0; i_col < p_vout->p_sys->i_col; i_col++ )
|
|
{
|
|
if( !p_vout->p_sys->pp_vout[ i_vout ].b_active )
|
|
{
|
|
for( i_plane = 0 ; i_plane < p_pic->i_planes ; i_plane++ )
|
|
{
|
|
pi_left_skip[i_plane] +=
|
|
p_vout->p_sys->pp_vout[ i_vout ].i_width
|
|
* p_pic->p[i_plane].i_pitch / p_vout->output.i_width;
|
|
}
|
|
i_vout++;
|
|
continue;
|
|
}
|
|
|
|
while( ( p_outpic =
|
|
vout_CreatePicture( p_vout->p_sys->pp_vout[ i_vout ].p_vout,
|
|
0, 0, 0 )
|
|
) == NULL )
|
|
{
|
|
if( p_vout->b_die || p_vout->b_error )
|
|
{
|
|
vout_DestroyPicture(
|
|
p_vout->p_sys->pp_vout[ i_vout ].p_vout, p_outpic );
|
|
return;
|
|
}
|
|
|
|
msleep( VOUT_OUTMEM_SLEEP );
|
|
}
|
|
|
|
vout_DatePicture( p_vout->p_sys->pp_vout[ i_vout ].p_vout,
|
|
p_outpic, p_pic->date );
|
|
vout_LinkPicture( p_vout->p_sys->pp_vout[ i_vout ].p_vout,
|
|
p_outpic );
|
|
|
|
for( i_plane = 0 ; i_plane < p_pic->i_planes ; i_plane++ )
|
|
{
|
|
uint8_t *p_in, *p_in_end, *p_out;
|
|
int i_in_pitch = p_pic->p[i_plane].i_pitch;
|
|
int i_out_pitch = p_outpic->p[i_plane].i_pitch;
|
|
int i_copy_pitch = p_outpic->p[i_plane].i_visible_pitch;
|
|
int i_lines = p_outpic->p[i_plane].i_visible_lines;
|
|
#ifdef OVERLAP
|
|
if (i_col) pi_left_skip[i_plane] -= (2 * p_vout->p_sys->i_halfLength ) / (p_vout->p_sys->pp_vout[i_vout].i_width / i_copy_pitch);
|
|
if ((i_row) && (!i_col)) pi_top_skip[i_plane] -= (2 * p_vout->p_sys->i_halfHeight * p_pic->p[i_plane].i_pitch) / (p_vout->p_sys->pp_vout[i_vout].i_width / i_copy_pitch);
|
|
if ((p_vout->p_sys->i_row > 2) && (i_row == 1) && (!i_col)) pi_top_skip[i_plane] -= (2 * p_vout->p_sys->i_halfHeight * p_pic->p[i_plane].i_pitch) / (p_vout->p_sys->pp_vout[i_vout].i_width / i_copy_pitch);
|
|
if ((!p_vout->p_sys->pp_vout[p_vout->p_sys->i_col].b_active))
|
|
pi_top_skip[i_plane] -= (2 * p_vout->p_sys->i_halfHeight * i_row * p_pic->p[i_plane].i_pitch) / (p_vout->p_sys->pp_vout[i_vout].i_width / i_copy_pitch);
|
|
// i_n : previous inactive pp_vout
|
|
int i_n=0;
|
|
while ((!p_vout->p_sys->pp_vout[i_row * p_vout->p_sys->i_col + i_col - 1 - i_n].b_active) && (i_col - i_n > 1)) i_n++;
|
|
if ((i_col > 1) && i_n)
|
|
pi_left_skip[i_plane] -= i_n*(2 * p_vout->p_sys->i_halfLength ) / (p_vout->p_sys->pp_vout[i_vout].i_width / i_copy_pitch);
|
|
|
|
p_in = p_pic->p[i_plane].p_pixels
|
|
/* Wall proprities */
|
|
+ pi_top_skip[i_plane] + pi_left_skip[i_plane];
|
|
|
|
if ((p_vout->p_sys->i_row > 2) &&
|
|
((!i_row) || (i_row + 1 == p_vout->p_sys->i_row)))
|
|
i_lines -= (2 * p_vout->p_sys->i_halfHeight) / (p_vout->p_sys->pp_vout[i_vout].i_width / i_copy_pitch);
|
|
|
|
// 1088 lines bug in a mpeg2 stream of 1080 lines
|
|
if ((p_vout->p_sys->i_row - 1 == i_row) &&
|
|
(p_pic->p[i_plane].i_lines == 1088))
|
|
i_lines -= 8 / (p_vout->p_sys->pp_vout[i_vout].i_width / i_copy_pitch);
|
|
|
|
p_in_end = p_in + i_lines * p_pic->p[i_plane].i_pitch;
|
|
#else
|
|
p_in = p_pic->p[i_plane].p_pixels
|
|
+ pi_top_skip[i_plane] + pi_left_skip[i_plane];
|
|
|
|
p_in_end = p_in + i_lines * p_pic->p[i_plane].i_pitch;
|
|
#endif
|
|
p_out = p_outpic->p[i_plane].p_pixels;
|
|
#ifdef OVERLAP
|
|
if ((p_vout->p_sys->i_row > 2) && (!i_row))
|
|
p_out += (p_outpic->p[i_plane].i_pitch * (2 * p_vout->p_sys->i_halfHeight) / (p_vout->p_sys->pp_vout[i_vout].i_width / i_copy_pitch));
|
|
|
|
int length;
|
|
int i_col_mod;
|
|
length = 2 * p_vout->p_sys->i_halfLength / (p_vout->p_sys->pp_vout[i_vout].i_width / i_copy_pitch);
|
|
|
|
if (p_vout->p_sys->b_has_changed)
|
|
{
|
|
Denom = F2(length);
|
|
a_2 = p_vout->p_sys->a_2 * (ACCURACY / 100);
|
|
a_1 = p_vout->p_sys->a_1 * length * (ACCURACY / 100);
|
|
a_0 = p_vout->p_sys->a_0 * Denom * (ACCURACY / 100);
|
|
for(i_col_mod = 0; i_col_mod < 2; i_col_mod++)
|
|
for (i_index = 0; i_index < length; i_index++)
|
|
{
|
|
p_vout->p_sys->lambda[i_col_mod][i_plane][i_index] = CLIP_0A(!i_col_mod ? ACCURACY - (F4(a_2, a_1, i_index) + a_0) / Denom : ACCURACY - (F4(a_2, a_1,length - i_index) + a_0) / Denom);
|
|
p_vout->p_sys->cstYUV[i_col_mod][i_plane][i_index] = ((ACCURACY - p_vout->p_sys->lambda[i_col_mod][i_plane][i_index]) * constantYUV[i_plane]) / ACCURACY;
|
|
}
|
|
}
|
|
#endif
|
|
while( p_in < p_in_end )
|
|
{
|
|
#ifndef OVERLAP
|
|
p_vout->p_libvlc->pf_memcpy( p_out , p_in, i_copy_pitch);
|
|
#else
|
|
if (p_vout->p_sys->i_col > 2)
|
|
{
|
|
length /= 2;
|
|
if (i_col == 0)
|
|
p_vout->p_libvlc->pf_memcpy( p_out + length , p_in, i_copy_pitch - length);
|
|
else if (i_col + 1 == p_vout->p_sys->i_col)
|
|
p_vout->p_libvlc->pf_memcpy( p_out, p_in - length, i_copy_pitch - length);
|
|
else
|
|
p_vout->p_libvlc->pf_memcpy( p_out, p_in - length, i_copy_pitch);
|
|
|
|
if ((i_col == 0))
|
|
// black bar
|
|
{
|
|
LeftOffset = 0;
|
|
p_out += LeftOffset;
|
|
memset(p_out, constantYUV[i_plane], length);
|
|
p_out -= LeftOffset;
|
|
}
|
|
else if ((i_col + 1 == p_vout->p_sys->i_col ))
|
|
// black bar
|
|
{
|
|
LeftOffset = i_copy_pitch - length;
|
|
p_out += LeftOffset;
|
|
memset(p_out, constantYUV[i_plane], length);
|
|
p_out -= LeftOffset;
|
|
}
|
|
length *= 2;
|
|
}
|
|
else
|
|
p_vout->p_libvlc->pf_memcpy( p_out , p_in, i_copy_pitch);
|
|
|
|
if (p_vout->p_sys->b_attenuate)
|
|
{
|
|
// vertical blend
|
|
// first blended zone
|
|
if (i_col)
|
|
{
|
|
LeftOffset = 0;
|
|
p_out += LeftOffset;
|
|
for (i_index = 0; i_index < length; i_index++)
|
|
{
|
|
#ifndef GAMMA
|
|
*(p_out + i_index) = (p_vout->p_sys->lambda[1][i_plane][i_index] *
|
|
(*(p_out + i_index))) / ACCURACY +
|
|
p_vout->p_sys->cstYUV[1][i_plane][i_index];
|
|
#else
|
|
*(p_out + i_index) = p_vout->p_sys->LUT[i_plane][p_vout->p_sys->lambda[1][i_plane][i_index]][*(p_out + i_index)];
|
|
#endif
|
|
}
|
|
p_out -= LeftOffset;
|
|
}
|
|
// second blended zone
|
|
if (i_col + 1 < p_vout->p_sys->i_col)
|
|
{
|
|
LeftOffset = i_copy_pitch - length;
|
|
p_out += LeftOffset;
|
|
for (i_index = 0; i_index < length; i_index++)
|
|
{
|
|
#ifndef GAMMA
|
|
*(p_out + i_index) = (p_vout->p_sys->lambda[0][i_plane][i_index] *
|
|
(*(p_out + i_index))) / ACCURACY +
|
|
p_vout->p_sys->cstYUV[0][i_plane][i_index];
|
|
#else
|
|
|
|
*(p_out + i_index) = p_vout->p_sys->LUT[i_plane][p_vout->p_sys->lambda[0][i_plane][i_index]][*(p_out + i_index)];
|
|
#endif
|
|
}
|
|
p_out -= LeftOffset;
|
|
}
|
|
// end blended zone
|
|
}
|
|
#endif
|
|
p_in += i_in_pitch;
|
|
p_out += i_out_pitch;
|
|
}
|
|
#ifdef OVERLAP
|
|
// horizontal blend
|
|
if (!p_vout->p_sys->b_attenuate)
|
|
{
|
|
if ((i_row == 0) && (p_vout->p_sys->i_row > 2))
|
|
// black bar
|
|
{
|
|
int height = 2 * p_vout->p_sys->i_halfHeight / (p_vout->p_sys->pp_vout[i_vout].i_width / i_copy_pitch);
|
|
TopOffset = i_lines + (2 * p_vout->p_sys->i_halfHeight) / (p_vout->p_sys->pp_vout[i_vout].i_width / i_copy_pitch);
|
|
p_out -= TopOffset * p_outpic->p[i_plane].i_pitch;
|
|
for (i_index = 0; i_index < height; i_index++)
|
|
for (i_index2 = 0; i_index2 < i_copy_pitch; i_index2++)
|
|
*(p_out + (i_index * p_outpic->p[i_plane].i_pitch) + i_index2) = constantYUV[i_plane];
|
|
p_out += TopOffset * p_outpic->p[i_plane].i_pitch;
|
|
}
|
|
else if ((i_row + 1 == p_vout->p_sys->i_row) && (p_vout->p_sys->i_row > 2))
|
|
// black bar
|
|
{
|
|
int height = 2 * p_vout->p_sys->i_halfHeight / (p_vout->p_sys->pp_vout[i_vout].i_width / i_copy_pitch);
|
|
TopOffset = height - (2 * p_vout->p_sys->i_halfHeight) / (p_vout->p_sys->pp_vout[i_vout].i_width / i_copy_pitch);
|
|
p_out -= TopOffset * p_outpic->p[i_plane].i_pitch;
|
|
for (i_index = 0; i_index < height; i_index++)
|
|
for (i_index2 = 0; i_index2 < i_copy_pitch; i_index2++)
|
|
*(p_out + (i_index * p_outpic->p[i_plane].i_pitch) + i_index2) = constantYUV[i_plane];
|
|
p_out += TopOffset * p_outpic->p[i_plane].i_pitch;
|
|
}
|
|
}
|
|
else
|
|
{
|
|
if (p_vout->p_sys->i_row >= 2)
|
|
{
|
|
length = 2 * p_vout->p_sys->i_halfHeight / (p_vout->p_sys->pp_vout[i_vout].i_width / i_copy_pitch);
|
|
if (p_vout->p_sys->b_has_changed)
|
|
{
|
|
Denom = F2(length);
|
|
a_2 = p_vout->p_sys->a_2 * (ACCURACY / 100);
|
|
a_1 = p_vout->p_sys->a_1 * length * (ACCURACY / 100);
|
|
a_0 = p_vout->p_sys->a_0 * Denom * (ACCURACY / 100);
|
|
for(i_col_mod = 0; i_col_mod < 2; i_col_mod++)
|
|
for (i_index = 0; i_index < length; i_index++)
|
|
{
|
|
p_vout->p_sys->lambda2[i_col_mod][i_plane][i_index] = CLIP_0A(!i_col_mod ? ACCURACY - (F4(a_2, a_1, i_index) + a_0) / Denom : ACCURACY - (F4(a_2, a_1,length - i_index) + a_0) / Denom);
|
|
p_vout->p_sys->cstYUV2[i_col_mod][i_plane][i_index] = ((ACCURACY - p_vout->p_sys->lambda2[i_col_mod][i_plane][i_index]) * constantYUV[i_plane]) / ACCURACY;
|
|
}
|
|
}
|
|
// first blended zone
|
|
|
|
if (i_row)
|
|
{
|
|
TopOffset = i_lines;
|
|
p_out -= TopOffset * p_outpic->p[i_plane].i_pitch;
|
|
for (i_index = 0; i_index < length; i_index++)
|
|
for (i_index2 = 0; i_index2 < i_copy_pitch; i_index2++)
|
|
#ifndef GAMMA
|
|
*(p_out + (i_index * p_outpic->p[i_plane].i_pitch) + i_index2) = (p_vout->p_sys->lambda2[1][i_plane][i_index] *
|
|
(*(p_out + (i_index * p_outpic->p[i_plane].i_pitch) + i_index2))) / ACCURACY +
|
|
p_vout->p_sys->cstYUV2[1][i_plane][i_index];
|
|
#else
|
|
|
|
*(p_out + (i_index * p_outpic->p[i_plane].i_pitch) + i_index2) = p_vout->p_sys->LUT[i_plane][p_vout->p_sys->lambda2[1][i_plane][i_index]][*(p_out + (i_index * p_outpic->p[i_plane].i_pitch) + i_index2)];
|
|
#endif
|
|
p_out += TopOffset * p_outpic->p[i_plane].i_pitch;
|
|
}
|
|
else if (p_vout->p_sys->i_row > 2)
|
|
// black bar
|
|
{
|
|
TopOffset = i_lines + (2 * p_vout->p_sys->i_halfHeight) / (p_vout->p_sys->pp_vout[i_vout].i_width / i_copy_pitch);
|
|
p_out -= TopOffset * p_outpic->p[i_plane].i_pitch;
|
|
for (i_index = 0; i_index < length; i_index++)
|
|
for (i_index2 = 0; i_index2 < i_copy_pitch; i_index2++)
|
|
*(p_out + (i_index * p_outpic->p[i_plane].i_pitch) + i_index2) = constantYUV[i_plane];
|
|
p_out += TopOffset * p_outpic->p[i_plane].i_pitch;
|
|
}
|
|
|
|
// second blended zone
|
|
|
|
if (i_row + 1 < p_vout->p_sys->i_row)
|
|
{
|
|
TopOffset = length;
|
|
p_out -= TopOffset * p_outpic->p[i_plane].i_pitch;
|
|
for (i_index = 0; i_index < length; i_index++)
|
|
for (i_index2 = 0; i_index2 < i_copy_pitch; i_index2++)
|
|
#ifndef GAMMA
|
|
*(p_out + (i_index * p_outpic->p[i_plane].i_pitch) + i_index2) = (p_vout->p_sys->lambda2[0][i_plane][i_index] *
|
|
(*(p_out + (i_index * p_outpic->p[i_plane].i_pitch) + i_index2))) / ACCURACY +
|
|
p_vout->p_sys->cstYUV2[0][i_plane][i_index];
|
|
#else
|
|
|
|
*(p_out + (i_index * p_outpic->p[i_plane].i_pitch) + i_index2) = p_vout->p_sys->LUT[i_plane][p_vout->p_sys->lambda2[0][i_plane][i_index]][*(p_out + (i_index * p_outpic->p[i_plane].i_pitch) + i_index2)];
|
|
|
|
#endif
|
|
p_out += TopOffset * p_outpic->p[i_plane].i_pitch;
|
|
}
|
|
else if (p_vout->p_sys->i_row > 2)
|
|
// black bar
|
|
{
|
|
TopOffset = length - (2 * p_vout->p_sys->i_halfHeight) / (p_vout->p_sys->pp_vout[i_vout].i_width / i_copy_pitch);
|
|
p_out -= TopOffset * p_outpic->p[i_plane].i_pitch;
|
|
for (i_index = 0; i_index < length; i_index++)
|
|
for (i_index2 = 0; i_index2 < i_copy_pitch; i_index2++)
|
|
*(p_out + (i_index * p_outpic->p[i_plane].i_pitch) + i_index2) = constantYUV[i_plane];
|
|
p_out += TopOffset * p_outpic->p[i_plane].i_pitch;
|
|
}
|
|
// end blended zone
|
|
}
|
|
}
|
|
#endif
|
|
// bug for wall filter : fix by CC
|
|
// pi_left_skip[i_plane] += i_out_pitch;
|
|
pi_left_skip[i_plane] += i_copy_pitch;
|
|
}
|
|
|
|
vout_UnlinkPicture( p_vout->p_sys->pp_vout[ i_vout ].p_vout,
|
|
p_outpic );
|
|
vout_DisplayPicture( p_vout->p_sys->pp_vout[ i_vout ].p_vout,
|
|
p_outpic );
|
|
i_vout++;
|
|
}
|
|
|
|
for( i_plane = 0 ; i_plane < p_pic->i_planes ; i_plane++ )
|
|
{
|
|
pi_top_skip[i_plane] += p_vout->p_sys->pp_vout[ i_vout ].i_height
|
|
* p_pic->p[i_plane].i_lines
|
|
/ p_vout->output.i_height
|
|
* p_pic->p[i_plane].i_pitch;
|
|
}
|
|
|
|
}
|
|
#ifdef OVERLAP
|
|
if (p_vout->p_sys->b_has_changed) p_vout->p_sys->b_has_changed = VLC_FALSE;
|
|
#endif
|
|
}
|
|
|
|
|
|
/*****************************************************************************
|
|
* RenderPackedRGB: displays previously rendered output
|
|
*****************************************************************************
|
|
* This function send the currently rendered image to Wall image, waits
|
|
* until it is displayed and switch the two rendering buffers, preparing next
|
|
* frame.
|
|
*****************************************************************************/
|
|
static void RenderPackedRGB( vout_thread_t *p_vout, picture_t *p_pic )
|
|
{
|
|
picture_t *p_outpic = NULL;
|
|
int i_col, i_row, i_vout, i_plane;
|
|
int pi_left_skip[VOUT_MAX_PLANES], pi_top_skip[VOUT_MAX_PLANES];
|
|
#ifdef OVERLAP
|
|
int LeftOffset, TopOffset;
|
|
int Denom;
|
|
int a_2;
|
|
int a_1;
|
|
int a_0;
|
|
int i_index, i_index2;
|
|
#endif
|
|
|
|
i_vout = 0;
|
|
|
|
for( i_plane = 0 ; i_plane < p_pic->i_planes ; i_plane++ )
|
|
{
|
|
pi_top_skip[i_plane] = 0;
|
|
}
|
|
|
|
for( i_row = 0; i_row < p_vout->p_sys->i_row; i_row++ )
|
|
{
|
|
for( i_plane = 0 ; i_plane < p_pic->i_planes ; i_plane++ )
|
|
{
|
|
pi_left_skip[i_plane] = 0;
|
|
}
|
|
|
|
for( i_col = 0; i_col < p_vout->p_sys->i_col; i_col++ )
|
|
{
|
|
if( !p_vout->p_sys->pp_vout[ i_vout ].b_active )
|
|
{
|
|
for( i_plane = 0 ; i_plane < p_pic->i_planes ; i_plane++ )
|
|
{
|
|
pi_left_skip[i_plane] +=
|
|
p_vout->p_sys->pp_vout[ i_vout ].i_width * p_pic->p->i_pixel_pitch;
|
|
}
|
|
i_vout++;
|
|
continue;
|
|
}
|
|
|
|
while( ( p_outpic =
|
|
vout_CreatePicture( p_vout->p_sys->pp_vout[ i_vout ].p_vout,
|
|
0, 0, 0 )
|
|
) == NULL )
|
|
{
|
|
if( p_vout->b_die || p_vout->b_error )
|
|
{
|
|
vout_DestroyPicture(
|
|
p_vout->p_sys->pp_vout[ i_vout ].p_vout, p_outpic );
|
|
return;
|
|
}
|
|
|
|
msleep( VOUT_OUTMEM_SLEEP );
|
|
}
|
|
|
|
vout_DatePicture( p_vout->p_sys->pp_vout[ i_vout ].p_vout,
|
|
p_outpic, p_pic->date );
|
|
vout_LinkPicture( p_vout->p_sys->pp_vout[ i_vout ].p_vout,
|
|
p_outpic );
|
|
|
|
for( i_plane = 0 ; i_plane < p_pic->i_planes ; i_plane++ )
|
|
{
|
|
uint8_t *p_in, *p_in_end, *p_out;
|
|
int i_in_pitch = p_pic->p[i_plane].i_pitch;
|
|
int i_out_pitch = p_outpic->p[i_plane].i_pitch;
|
|
int i_copy_pitch = p_outpic->p[i_plane].i_visible_pitch;
|
|
|
|
#ifdef OVERLAP
|
|
if (i_col) pi_left_skip[i_plane] -= (2 * p_vout->p_sys->i_halfLength) * p_pic->p->i_pixel_pitch;
|
|
if ((i_row) && (!i_col)) pi_top_skip[i_plane] -= (2 * p_vout->p_sys->i_halfHeight * p_pic->p[i_plane].i_pitch);
|
|
if ((!p_vout->p_sys->pp_vout[p_vout->p_sys->i_col].b_active))
|
|
pi_top_skip[i_plane] -= (2 * p_vout->p_sys->i_halfHeight * i_row * p_pic->p[i_plane].i_pitch);
|
|
// i_n : previous inactive pp_vout
|
|
int i_n=0;
|
|
while ((!p_vout->p_sys->pp_vout[i_row * p_vout->p_sys->i_col + i_col - 1 - i_n].b_active) && (i_col - i_n > 1)) i_n++;
|
|
if ((i_col > 1) && i_n)
|
|
pi_left_skip[i_plane] -= i_n*(2 * p_vout->p_sys->i_halfLength ) * p_pic->p->i_pixel_pitch;
|
|
|
|
p_in = p_pic->p[i_plane].p_pixels
|
|
/* Wall proprities */
|
|
+ pi_top_skip[i_plane] + pi_left_skip[i_plane];
|
|
|
|
int i_lines = p_outpic->p[i_plane].i_visible_lines;
|
|
// 1088 lines bug in a mpeg2 stream of 1080 lines
|
|
if ((p_vout->p_sys->i_row - 1 == i_row) &&
|
|
(p_pic->p[i_plane].i_lines == 1088))
|
|
i_lines -= 8;
|
|
|
|
p_in_end = p_in + i_lines * p_pic->p[i_plane].i_pitch;
|
|
#else
|
|
p_in = p_pic->p[i_plane].p_pixels
|
|
+ pi_top_skip[i_plane] + pi_left_skip[i_plane];
|
|
|
|
p_in_end = p_in + p_outpic->p[i_plane].i_visible_lines
|
|
* p_pic->p[i_plane].i_pitch;
|
|
#endif //OVERLAP
|
|
|
|
p_out = p_outpic->p[i_plane].p_pixels;
|
|
|
|
|
|
#ifdef OVERLAP
|
|
if ((p_vout->p_sys->i_row > 2) && (!i_row))
|
|
p_out += (p_outpic->p[i_plane].i_pitch * (2 * p_vout->p_sys->i_halfHeight) * p_pic->p->i_pixel_pitch);
|
|
|
|
int length;
|
|
length = 2 * p_vout->p_sys->i_halfLength * p_pic->p->i_pixel_pitch;
|
|
|
|
if (p_vout->p_sys->b_has_changed)
|
|
{
|
|
int i_plane_;
|
|
int i_col_mod;
|
|
Denom = F2(length / p_pic->p->i_pixel_pitch);
|
|
a_2 = p_vout->p_sys->a_2 * (ACCURACY / 100);
|
|
a_1 = p_vout->p_sys->a_1 * 2 * p_vout->p_sys->i_halfLength * (ACCURACY / 100);
|
|
a_0 = p_vout->p_sys->a_0 * Denom * (ACCURACY / 100);
|
|
for(i_col_mod = 0; i_col_mod < 2; i_col_mod++)
|
|
for (i_index = 0; i_index < length / p_pic->p->i_pixel_pitch; i_index++)
|
|
for (i_plane_ = 0; i_plane_ < p_pic->p->i_pixel_pitch; i_plane_++)
|
|
p_vout->p_sys->lambda[i_col_mod][i_plane_][i_index] = CLIP_0A(!i_col_mod ? ACCURACY - (F4(a_2, a_1, i_index) + a_0) / Denom : ACCURACY - (F4(a_2, a_1,(length / p_pic->p->i_pixel_pitch) - i_index) + a_0) / Denom);
|
|
}
|
|
#endif
|
|
while( p_in < p_in_end )
|
|
{
|
|
#ifndef OVERLAP
|
|
p_vout->p_libvlc->pf_memcpy( p_out, p_in, i_copy_pitch);
|
|
#else
|
|
if (p_vout->p_sys->i_col > 2)
|
|
{
|
|
// vertical blend
|
|
length /= 2;
|
|
if (i_col == 0)
|
|
p_vout->p_libvlc->pf_memcpy( p_out + length, p_in, i_copy_pitch - length);
|
|
else if (i_col + 1 == p_vout->p_sys->i_col)
|
|
p_vout->p_libvlc->pf_memcpy( p_out, p_in - length, i_copy_pitch - length);
|
|
else
|
|
p_vout->p_libvlc->pf_memcpy( p_out, p_in - length, i_copy_pitch);
|
|
|
|
if ((i_col == 0))
|
|
// black bar
|
|
{
|
|
LeftOffset = 0;
|
|
p_out += LeftOffset;
|
|
p_in += LeftOffset;
|
|
for (i_index = 0; i_index < length; i_index++)
|
|
*(p_out + i_index) = 0;
|
|
p_out -= LeftOffset;
|
|
p_in -= LeftOffset;
|
|
}
|
|
else if ((i_col + 1 == p_vout->p_sys->i_col ))
|
|
// black bar
|
|
{
|
|
LeftOffset = i_copy_pitch - length;
|
|
p_out += LeftOffset;
|
|
p_in += LeftOffset;
|
|
for (i_index = 0; i_index < length; i_index++)
|
|
*(p_out + i_index) = 0;
|
|
p_out -= LeftOffset;
|
|
p_in -= LeftOffset;
|
|
}
|
|
length *= 2;
|
|
}
|
|
else
|
|
p_vout->p_libvlc->pf_memcpy( p_out, p_in, i_copy_pitch);
|
|
|
|
// vertical blend
|
|
// first blended zone
|
|
if (i_col)
|
|
{
|
|
LeftOffset = 0;
|
|
p_out += LeftOffset;
|
|
for (i_index = 0; i_index < length; i_index++)
|
|
#ifndef GAMMA
|
|
*(p_out + i_index) = (p_vout->p_sys->lambda[1][i_index % p_pic->p->i_pixel_pitch][i_index / p_pic->p->i_pixel_pitch] *
|
|
(*(p_out + i_index))) / ACCURACY;
|
|
#else
|
|
*(p_out + i_index) = p_vout->p_sys->LUT[i_index % p_pic->p->i_pixel_pitch][p_vout->p_sys->lambda[1][i_index % p_pic->p->i_pixel_pitch][i_index / p_pic->p->i_pixel_pitch]][*(p_out + i_index)];
|
|
#endif
|
|
p_out -= LeftOffset;
|
|
}
|
|
// second blended zone
|
|
if (i_col + 1 < p_vout->p_sys->i_col)
|
|
{
|
|
LeftOffset = i_copy_pitch - length;
|
|
p_out += LeftOffset;
|
|
for (i_index = 0; i_index < length; i_index++)
|
|
#ifndef GAMMA
|
|
*(p_out + i_index) = (p_vout->p_sys->lambda[0][i_index % p_pic->p->i_pixel_pitch][i_index / p_pic->p->i_pixel_pitch] *
|
|
(*(p_out + i_index))) / ACCURACY;
|
|
#else
|
|
*(p_out + i_index) = p_vout->p_sys->LUT[i_index % p_pic->p->i_pixel_pitch][p_vout->p_sys->lambda[0][i_index % p_pic->p->i_pixel_pitch][i_index / p_pic->p->i_pixel_pitch]][*(p_out + i_index)];
|
|
#endif
|
|
p_out -= LeftOffset;
|
|
}
|
|
// end blended zone
|
|
#endif //OVERLAP
|
|
p_in += i_in_pitch;
|
|
p_out += i_out_pitch;
|
|
}
|
|
#ifdef OVERLAP
|
|
// horizontal blend
|
|
if (!p_vout->p_sys->b_attenuate)
|
|
{
|
|
if ((i_row == 0) && (p_vout->p_sys->i_row > 2))
|
|
// black bar
|
|
{
|
|
TopOffset = i_lines + (2 * p_vout->p_sys->i_halfHeight);
|
|
p_out -= TopOffset * p_outpic->p[i_plane].i_pitch;
|
|
for (i_index = 0; i_index < length; i_index++)
|
|
for (i_index2 = 0; i_index2 < i_copy_pitch; i_index2++)
|
|
*(p_out + (i_index * p_outpic->p[i_plane].i_pitch) + i_index2) = 0;
|
|
p_out += TopOffset * p_outpic->p[i_plane].i_pitch;
|
|
}
|
|
else if ((i_row + 1 == p_vout->p_sys->i_row) && (p_vout->p_sys->i_row > 2))
|
|
// black bar
|
|
{
|
|
TopOffset = length - (2 * p_vout->p_sys->i_halfHeight);
|
|
p_out -= TopOffset * p_outpic->p[i_plane].i_pitch;
|
|
for (i_index = 0; i_index < length; i_index++)
|
|
for (i_index2 = 0; i_index2 < i_copy_pitch; i_index2++)
|
|
*(p_out + (i_index * p_outpic->p[i_plane].i_pitch) + i_index2) = 0;
|
|
p_out += TopOffset * p_outpic->p[i_plane].i_pitch;
|
|
}
|
|
}
|
|
else
|
|
{
|
|
if (p_vout->p_sys->i_row >= 2)
|
|
{
|
|
length = 2 * p_vout->p_sys->i_halfHeight;
|
|
if (p_vout->p_sys->b_has_changed)
|
|
{
|
|
int i_plane_;
|
|
int i_row_mod;
|
|
Denom = F2(length);
|
|
a_2 = p_vout->p_sys->a_2 * (ACCURACY / 100);
|
|
a_1 = p_vout->p_sys->a_1 * length * (ACCURACY / 100);
|
|
a_0 = p_vout->p_sys->a_0 * Denom * (ACCURACY / 100);
|
|
for(i_row_mod = 0; i_row_mod < 2; i_row_mod++)
|
|
for (i_index = 0; i_index < length; i_index++)
|
|
for (i_plane_ = 0; i_plane_ < p_pic->p->i_pixel_pitch; i_plane_++)
|
|
p_vout->p_sys->lambda2[i_row_mod][i_plane_][i_index] = CLIP_0A(!i_row_mod ? ACCURACY - (F4(a_2, a_1, i_index) + a_0) / Denom : ACCURACY - (F4(a_2, a_1,(length) - i_index) + a_0) / Denom);
|
|
}
|
|
// first blended zone
|
|
|
|
if (i_row)
|
|
{
|
|
TopOffset = i_lines;
|
|
p_out -= TopOffset * p_outpic->p[i_plane].i_pitch;
|
|
for (i_index = 0; i_index < length; i_index++)
|
|
for (i_index2 = 0; i_index2 < i_copy_pitch; i_index2++)
|
|
#ifndef GAMMA
|
|
*(p_out + (i_index * p_outpic->p[i_plane].i_pitch) + i_index2) = (p_vout->p_sys->lambda2[1][i_index2 % p_pic->p->i_pixel_pitch][i_index] *
|
|
(*(p_out + (i_index * p_outpic->p[i_plane].i_pitch) + i_index2))) / ACCURACY;
|
|
#else
|
|
*(p_out + (i_index * p_outpic->p[i_plane].i_pitch) + i_index2) = p_vout->p_sys->LUT[i_index2 % p_pic->p->i_pixel_pitch][p_vout->p_sys->lambda2[1][i_index2 % p_pic->p->i_pixel_pitch][i_index]][*(p_out + (i_index * p_outpic->p[i_plane].i_pitch) + i_index2)];
|
|
#endif
|
|
p_out += TopOffset * p_outpic->p[i_plane].i_pitch;
|
|
}
|
|
else if (p_vout->p_sys->i_row > 2)
|
|
// black bar
|
|
{
|
|
TopOffset = i_lines + (2 * p_vout->p_sys->i_halfHeight);
|
|
p_out -= TopOffset * p_outpic->p[i_plane].i_pitch;
|
|
for (i_index = 0; i_index < length; i_index++)
|
|
for (i_index2 = 0; i_index2 < i_copy_pitch; i_index2++)
|
|
*(p_out + (i_index * p_outpic->p[i_plane].i_pitch) + i_index2) = 0;
|
|
p_out += TopOffset * p_outpic->p[i_plane].i_pitch;
|
|
}
|
|
|
|
// second blended zone
|
|
|
|
if (i_row + 1 < p_vout->p_sys->i_row)
|
|
{
|
|
TopOffset = length;
|
|
p_out -= TopOffset * p_outpic->p[i_plane].i_pitch;
|
|
for (i_index = 0; i_index < length; i_index++)
|
|
for (i_index2 = 0; i_index2 < i_copy_pitch; i_index2++)
|
|
#ifndef GAMMA
|
|
*(p_out + (i_index * p_outpic->p[i_plane].i_pitch) + i_index2) = (p_vout->p_sys->lambda2[0][i_index2 % p_pic->p->i_pixel_pitch][i_index] *
|
|
(*(p_out + (i_index * p_outpic->p[i_plane].i_pitch) + i_index2))) / ACCURACY;
|
|
#else
|
|
*(p_out + (i_index * p_outpic->p[i_plane].i_pitch) + i_index2) = p_vout->p_sys->LUT[i_index2 % p_pic->p->i_pixel_pitch][p_vout->p_sys->lambda2[0][i_index2 % p_pic->p->i_pixel_pitch][i_index]][*(p_out + (i_index * p_outpic->p[i_plane].i_pitch) + i_index2)];
|
|
|
|
#endif
|
|
p_out += TopOffset * p_outpic->p[i_plane].i_pitch;
|
|
}
|
|
else if (p_vout->p_sys->i_row > 2)
|
|
// black bar
|
|
{
|
|
TopOffset = length - (2 * p_vout->p_sys->i_halfHeight);
|
|
p_out -= TopOffset * p_outpic->p[i_plane].i_pitch;
|
|
for (i_index = 0; i_index < length; i_index++)
|
|
for (i_index2 = 0; i_index2 < i_copy_pitch; i_index2++)
|
|
*(p_out + (i_index * p_outpic->p[i_plane].i_pitch) + i_index2) = 0;
|
|
p_out += TopOffset * p_outpic->p[i_plane].i_pitch;
|
|
}
|
|
// end blended zone
|
|
}
|
|
}
|
|
#endif
|
|
// bug for wall filter : fix by CC
|
|
// pi_left_skip[i_plane] += i_out_pitch;
|
|
pi_left_skip[i_plane] += i_copy_pitch;
|
|
}
|
|
|
|
vout_UnlinkPicture( p_vout->p_sys->pp_vout[ i_vout ].p_vout,
|
|
p_outpic );
|
|
vout_DisplayPicture( p_vout->p_sys->pp_vout[ i_vout ].p_vout,
|
|
p_outpic );
|
|
i_vout++;
|
|
}
|
|
|
|
for( i_plane = 0 ; i_plane < p_pic->i_planes ; i_plane++ )
|
|
{
|
|
pi_top_skip[i_plane] += p_vout->p_sys->pp_vout[ i_vout ].i_height
|
|
* p_pic->p[i_plane].i_lines
|
|
/ p_vout->output.i_height
|
|
* p_pic->p[i_plane].i_pitch;
|
|
}
|
|
}
|
|
#ifdef OVERLAP
|
|
if (p_vout->p_sys->b_has_changed) p_vout->p_sys->b_has_changed = VLC_FALSE;
|
|
#endif
|
|
}
|
|
|
|
|
|
#ifdef PACKED_YUV
|
|
// WARNING : NO DEBUGGED
|
|
/*****************************************************************************
|
|
* RenderPackedYUV: displays previously rendered output
|
|
*****************************************************************************
|
|
* This function send the currently rendered image to Wall image, waits
|
|
* until it is displayed and switch the two rendering buffers, preparing next
|
|
* frame.
|
|
*****************************************************************************/
|
|
static void RenderPackedYUV( vout_thread_t *p_vout, picture_t *p_pic )
|
|
{
|
|
picture_t *p_outpic = NULL;
|
|
int i_col, i_row, i_vout, i_plane;
|
|
int pi_left_skip[VOUT_MAX_PLANES], pi_top_skip[VOUT_MAX_PLANES];
|
|
#ifdef OVERLAP
|
|
int LeftOffset, TopOffset;
|
|
int constantYUV[3] = {0,128,128};
|
|
int Denom;
|
|
int a_2;
|
|
int a_1;
|
|
int a_0;
|
|
int i_index, i_index2;
|
|
#endif
|
|
|
|
|
|
i_vout = 0;
|
|
|
|
for( i_plane = 0 ; i_plane < p_pic->i_planes ; i_plane++ )
|
|
{
|
|
pi_top_skip[i_plane] = 0;
|
|
}
|
|
|
|
for( i_row = 0; i_row < p_vout->p_sys->i_row; i_row++ )
|
|
{
|
|
for( i_plane = 0 ; i_plane < p_pic->i_planes ; i_plane++ )
|
|
{
|
|
pi_left_skip[i_plane] = 0;
|
|
}
|
|
|
|
for( i_col = 0; i_col < p_vout->p_sys->i_col; i_col++ )
|
|
{
|
|
if( !p_vout->p_sys->pp_vout[ i_vout ].b_active )
|
|
{
|
|
for( i_plane = 0 ; i_plane < p_pic->i_planes ; i_plane++ )
|
|
{
|
|
pi_left_skip[i_plane] +=
|
|
p_vout->p_sys->pp_vout[ i_vout ].i_width
|
|
* p_pic->p[i_plane].i_pitch / p_vout->output.i_width;
|
|
}
|
|
i_vout++;
|
|
continue;
|
|
}
|
|
|
|
while( ( p_outpic =
|
|
vout_CreatePicture( p_vout->p_sys->pp_vout[ i_vout ].p_vout,
|
|
0, 0, 0 )
|
|
) == NULL )
|
|
{
|
|
if( p_vout->b_die || p_vout->b_error )
|
|
{
|
|
vout_DestroyPicture(
|
|
p_vout->p_sys->pp_vout[ i_vout ].p_vout, p_outpic );
|
|
return;
|
|
}
|
|
|
|
msleep( VOUT_OUTMEM_SLEEP );
|
|
}
|
|
|
|
vout_DatePicture( p_vout->p_sys->pp_vout[ i_vout ].p_vout,
|
|
p_outpic, p_pic->date );
|
|
vout_LinkPicture( p_vout->p_sys->pp_vout[ i_vout ].p_vout,
|
|
p_outpic );
|
|
|
|
for( i_plane = 0 ; i_plane < p_pic->i_planes ; i_plane++ )
|
|
{
|
|
uint8_t *p_in, *p_in_end, *p_out;
|
|
int i_in_pitch = p_pic->p[i_plane].i_pitch;
|
|
int i_out_pitch = p_outpic->p[i_plane].i_pitch;
|
|
int i_copy_pitch = p_outpic->p[i_plane].i_visible_pitch;
|
|
|
|
#ifdef OVERLAP
|
|
if (i_col) pi_left_skip[i_plane] -= (2 * p_vout->p_sys->i_halfLength ) / (p_vout->p_sys->pp_vout[i_vout].i_width / i_copy_pitch);
|
|
if ((i_row) && (!i_col)) pi_top_skip[i_plane] -= (2 * p_vout->p_sys->i_halfHeight * p_pic->p[i_plane].i_pitch) / (p_vout->p_sys->pp_vout[i_vout].i_width / i_copy_pitch);
|
|
if ((p_vout->p_sys->i_row > 2) && (i_row == 1) && (!i_col)) pi_top_skip[i_plane] -= (2 * p_vout->p_sys->i_halfHeight * p_pic->p[i_plane].i_pitch) / (p_vout->p_sys->pp_vout[i_vout].i_width / i_copy_pitch);
|
|
if ((!p_vout->p_sys->pp_vout[p_vout->p_sys->i_col].b_active))
|
|
pi_top_skip[i_plane] -= (2 * p_vout->p_sys->i_halfHeight * i_row * p_pic->p[i_plane].i_pitch) / (p_vout->p_sys->pp_vout[i_vout].i_width / i_copy_pitch);
|
|
// i_n : previous inactive pp_vout
|
|
int i_n=0;
|
|
while ((!p_vout->p_sys->pp_vout[i_row * p_vout->p_sys->i_col + i_col - 1 - i_n].b_active) && (i_col - i_n > 1)) i_n++;
|
|
if ((i_col > 1) && i_n)
|
|
pi_left_skip[i_plane] -= i_n*(2 * p_vout->p_sys->i_halfLength ) / (p_vout->p_sys->pp_vout[i_vout].i_width / i_copy_pitch);
|
|
|
|
p_in = p_pic->p[i_plane].p_pixels
|
|
/* Wall proprities */
|
|
+ pi_top_skip[i_plane] + pi_left_skip[i_plane];
|
|
|
|
int i_lines = p_outpic->p[i_plane].i_visible_lines;
|
|
// 1088 lines bug in a mpeg2 stream of 1080 lines
|
|
if ((p_vout->p_sys->i_row - 1 == i_row) &&
|
|
(p_pic->p[i_plane].i_lines == 1088))
|
|
i_lines -= 8;
|
|
|
|
p_in_end = p_in + i_lines * p_pic->p[i_plane].i_pitch;
|
|
#else
|
|
p_in = p_pic->p[i_plane].p_pixels
|
|
+ pi_top_skip[i_plane] + pi_left_skip[i_plane];
|
|
|
|
p_in_end = p_in + p_outpic->p[i_plane].i_visible_lines
|
|
* p_pic->p[i_plane].i_pitch;
|
|
#endif
|
|
p_out = p_outpic->p[i_plane].p_pixels;
|
|
#ifdef OVERLAP
|
|
int length;
|
|
length = 2 * p_vout->p_sys->i_halfLength * p_pic->p->i_pixel_pitch;
|
|
LeftOffset = (i_col ? 0 : i_copy_pitch - length);
|
|
if (p_vout->p_sys->b_has_changed)
|
|
{
|
|
#ifdef GAMMA
|
|
int i_plane_;
|
|
for (i_index = 0; i_index < length / p_pic->p->i_pixel_pitch; i_index++)
|
|
for (i_plane_ = 0; i_plane_ < p_pic->p->i_pixel_pitch; i_plane_++)
|
|
for (i_index2 = 0; i_index2 < 256; i_index2++)
|
|
p_vout->p_sys->LUT[i_plane_][i_index2][i_index] = F(i_index2, (length / p_pic->p->i_pixel_pitch, i_index, p_vout->p_sys->f_gamma[i_plane_]));
|
|
#endif
|
|
switch (p_vout->output.i_chroma)
|
|
{
|
|
case VLC_FOURCC('Y','U','Y','2'): // packed by 2
|
|
case VLC_FOURCC('Y','U','N','V'): // packed by 2
|
|
Denom = F2(length / p_pic->p->i_pixel_pitch);
|
|
a_2 = p_vout->p_sys->a_2 * (ACCURACY / 100);
|
|
a_1 = p_vout->p_sys->a_1 * 2 * p_vout->p_sys->i_halfLength * (ACCURACY / 100);
|
|
a_0 = p_vout->p_sys->a_0 * Denom * (ACCURACY / 100);
|
|
for (i_index = 0; i_index < length / p_pic->p->i_pixel_pitch; i_index+=p_pic->p->i_pixel_pitch)
|
|
// for each macropixel
|
|
{
|
|
// first image pixel
|
|
p_vout->p_sys->lambda[i_col][0][i_index] = CLIP_0A(!i_col ? ACCURACY - (F4(a_2, a_1, i_index) + a_0) / Denom : ACCURACY - (F4(a_2, a_1,(length / p_pic->p->i_pixel_pitch) - i_index) + a_0) / Denom);
|
|
p_vout->p_sys->cstYUV[i_col][0][i_index] = ((ACCURACY - p_vout->p_sys->lambda[i_col][0][i_index]) * constantYUV[0]) / ACCURACY;
|
|
p_vout->p_sys->lambda[i_col][1][i_index] = CLIP_0A(!i_col ? ACCURACY - (F4(a_2, a_1, i_index) + a_0) / Denom : ACCURACY - (F4(a_2, a_1,(length / p_pic->p->i_pixel_pitch) - i_index) + a_0) / Denom);
|
|
p_vout->p_sys->cstYUV[i_col][1][i_index] = ((ACCURACY - p_vout->p_sys->lambda[i_col][1][i_index]) * constantYUV[1]) / ACCURACY;
|
|
// second image pixel
|
|
p_vout->p_sys->lambda[i_col][0][i_index + 1] = CLIP_0A(!i_col ? ACCURACY - (F4(a_2, a_1, i_index + 1) + a_0) / Denom : ACCURACY - (F4(a_2, a_1,(length / p_pic->p->i_pixel_pitch) - (i_index + 1)) + a_0) / Denom);
|
|
p_vout->p_sys->cstYUV[i_col][0][i_index + 1] = ((ACCURACY - p_vout->p_sys->lambda[i_col][0][i_index]) * constantYUV[0]) / ACCURACY;
|
|
p_vout->p_sys->lambda[i_col][1][i_index + 1] = p_vout->p_sys->lambda[i_col][1][i_index];
|
|
p_vout->p_sys->cstYUV[i_col][1][i_index + 1] = p_vout->p_sys->cstYUV[i_col][1][i_index];
|
|
}
|
|
break;
|
|
case VLC_FOURCC('U','Y','V','Y'): // packed by 2
|
|
case VLC_FOURCC('U','Y','N','V'): // packed by 2
|
|
case VLC_FOURCC('Y','4','2','2'): // packed by 2
|
|
Denom = F2(length / p_pic->p->i_pixel_pitch);
|
|
a_2 = p_vout->p_sys->a_2 * (ACCURACY / 100);
|
|
a_1 = p_vout->p_sys->a_1 * 2 * p_vout->p_sys->i_halfLength * (ACCURACY / 100);
|
|
a_0 = p_vout->p_sys->a_0 * Denom * (ACCURACY / 100);
|
|
for (i_index = 0; i_index < length / p_pic->p->i_pixel_pitch; i_index+=p_pic->p->i_pixel_pitch)
|
|
// for each macropixel
|
|
{
|
|
// first image pixel
|
|
p_vout->p_sys->lambda[i_col][0][i_index] = CLIP_0A(!i_col ? ACCURACY - (F4(a_2, a_1, i_index) + a_0) / Denom : ACCURACY - (F4(a_2, a_1,(length / p_pic->p->i_pixel_pitch) - i_index) + a_0) / Denom);
|
|
p_vout->p_sys->cstYUV[i_col][0][i_index] = ((ACCURACY - p_vout->p_sys->lambda[i_col][0][i_index]) * constantYUV[1]) / ACCURACY;
|
|
p_vout->p_sys->lambda[i_col][1][i_index] = CLIP_0A(!i_col ? ACCURACY - (F4(a_2, a_1, i_index) + a_0) / Denom : ACCURACY - (F4(a_2, a_1,(length / p_pic->p->i_pixel_pitch) - i_index) + a_0) / Denom);
|
|
p_vout->p_sys->cstYUV[i_col][1][i_index] = ((ACCURACY - p_vout->p_sys->lambda[i_col][1][i_index]) * constantYUV[0]) / ACCURACY;
|
|
// second image pixel
|
|
p_vout->p_sys->lambda[i_col][0][i_index + 1] = CLIP_0A(!i_col ? ACCURACY - (F4(a_2, a_1, i_index + 1) + a_0) / Denom : ACCURACY - (F4(a_2, a_1,(length / p_pic->p->i_pixel_pitch) - (i_index + 1)) + a_0) / Denom);
|
|
p_vout->p_sys->cstYUV[i_col][0][i_index + 1] = ((ACCURACY - p_vout->p_sys->lambda[i_col][0][i_index]) * constantYUV[1]) / ACCURACY;
|
|
p_vout->p_sys->lambda[i_col][1][i_index + 1] = p_vout->p_sys->lambda[i_col][1][i_index];
|
|
p_vout->p_sys->cstYUV[i_col][1][i_index + 1] = p_vout->p_sys->cstYUV[i_col][1][i_index];
|
|
}
|
|
break;
|
|
default :
|
|
break;
|
|
}
|
|
}
|
|
#endif
|
|
while( p_in < p_in_end )
|
|
{
|
|
#ifndef OVERLAP
|
|
p_vout->p_libvlc->pf_memcpy( p_out, p_in, i_copy_pitch);
|
|
#else
|
|
p_vout->p_libvlc->pf_memcpy( p_out + i_col * length, p_in + i_col * length, i_copy_pitch - length);
|
|
p_out += LeftOffset;
|
|
p_in += LeftOffset;
|
|
#ifndef GAMMA
|
|
for (i_index = 0; i_index < length; i_index++)
|
|
*(p_out + i_index) = (p_vout->p_sys->lambda[i_col][i_index % p_pic->p->i_pixel_pitch][i_index / p_pic->p->i_pixel_pitch] *
|
|
(*(p_in + i_index))) / ACCURACY +
|
|
p_vout->p_sys->cstYUV[i_col][i_index % p_pic->p->i_pixel_pitch][i_index / p_pic->p->i_pixel_pitch];
|
|
#else
|
|
for (i_index = 0; i_index < length; i_index++)
|
|
*(p_out + i_index) = p_vout->p_sys->LUT[i_index % p_pic->p->i_pixel_pitch][(p_vout->p_sys->lambda[i_col][i_index % p_pic->p->i_pixel_pitch][i_index / p_pic->p->i_pixel_pitch] *
|
|
(*(p_in + i_index))) / ACCURACY +
|
|
p_vout->p_sys->cstYUV[i_col][i_index % p_pic->p->i_pixel_pitch][i_index / p_pic->p->i_pixel_pitch]][i_index / p_pic->p->i_pixel_pitch];
|
|
#endif
|
|
p_out -= LeftOffset;
|
|
p_in -= LeftOffset;
|
|
#endif
|
|
p_in += i_in_pitch;
|
|
p_out += i_out_pitch;
|
|
}
|
|
#ifdef OVERLAP
|
|
if (p_vout->p_sys->i_row == 2)
|
|
{
|
|
length = 2 * p_vout->p_sys->i_halfHeight * p_pic->p->i_pixel_pitch;
|
|
TopOffset = (i_row ? i_lines : length / p_pic->p->i_pixel_pitch);
|
|
if (p_vout->p_sys->b_has_changed)
|
|
{
|
|
#ifdef GAMMA
|
|
int i_plane_;
|
|
for (i_index = 0; i_index < length / p_pic->p->i_pixel_pitch; i_index++)
|
|
for (i_plane_ = 0; i_plane_ < p_pic->p->i_pixel_pitch; i_plane_++)
|
|
for (i_index2 = 0; i_index2 < 256; i_index2++)
|
|
p_vout->p_sys->LUT2[i_plane_][i_index2][i_index] = F(i_index2, (length / p_pic->p->i_pixel_pitch, i_index, p_vout->p_sys->f_gamma[i_plane_]));
|
|
#endif
|
|
switch (p_vout->output.i_chroma)
|
|
{
|
|
case VLC_FOURCC('Y','U','Y','2'): // packed by 2
|
|
case VLC_FOURCC('Y','U','N','V'): // packed by 2
|
|
Denom = F2(length / p_pic->p->i_pixel_pitch);
|
|
a_2 = p_vout->p_sys->a_2 * (ACCURACY / 100);
|
|
a_1 = p_vout->p_sys->a_1 * 2 * p_vout->p_sys->i_halfHeight * (ACCURACY / 100);
|
|
a_0 = p_vout->p_sys->a_0 * Denom * (ACCURACY / 100);
|
|
for (i_index = 0; i_index < length / p_pic->p->i_pixel_pitch; i_index+=p_pic->p->i_pixel_pitch)
|
|
// for each macropixel
|
|
{
|
|
// first image pixel
|
|
p_vout->p_sys->lambda2[i_row][0][i_index] = CLIP_0A(!i_row ? ACCURACY - (F4(a_2, a_1, i_index) + a_0) / Denom : ACCURACY - (F4(a_2, a_1,(length / p_pic->p->i_pixel_pitch) - i_index) + a_0) / Denom);
|
|
p_vout->p_sys->cstYUV2[i_row][0][i_index] = ((ACCURACY - p_vout->p_sys->lambda2[i_row][0][i_index]) * constantYUV[0]) / ACCURACY;
|
|
p_vout->p_sys->lambda2[i_row][1][i_index] = CLIP_0A(!i_row ? ACCURACY - (F4(a_2, a_1, i_index) + a_0) / Denom : ACCURACY - (F4(a_2, a_1,(length / p_pic->p->i_pixel_pitch) - i_index) + a_0) / Denom);
|
|
p_vout->p_sys->cstYUV2[i_row][1][i_index] = ((ACCURACY - p_vout->p_sys->lambda2[i_row][1][i_index]) * constantYUV[1]) / ACCURACY;
|
|
// second image pixel
|
|
p_vout->p_sys->lambda2[i_row][0][i_index + 1] = CLIP_0A(!i_row ? ACCURACY - (F4(a_2, a_1, i_index + 1) + a_0) / Denom : ACCURACY - (F4(a_2, a_1,(length / p_pic->p->i_pixel_pitch) - (i_index + 1)) + a_0) / Denom);
|
|
p_vout->p_sys->cstYUV2[i_row][0][i_index + 1] = ((ACCURACY - p_vout->p_sys->lambda2[i_row][0][i_index]) * constantYUV[0]) / ACCURACY;
|
|
p_vout->p_sys->lambda2[i_row][1][i_index + 1] = p_vout->p_sys->lambda2[i_row][1][i_index];
|
|
p_vout->p_sys->cstYUV2[i_row][1][i_index + 1] = p_vout->p_sys->cstYUV2[i_row][1][i_index];
|
|
}
|
|
break;
|
|
case VLC_FOURCC('U','Y','V','Y'): // packed by 2
|
|
case VLC_FOURCC('U','Y','N','V'): // packed by 2
|
|
case VLC_FOURCC('Y','4','2','2'): // packed by 2
|
|
Denom = F2(length / p_pic->p->i_pixel_pitch);
|
|
a_2 = p_vout->p_sys->a_2 * (ACCURACY / 100);
|
|
a_1 = p_vout->p_sys->a_1 * 2 * p_vout->p_sys->i_halfHeight * (ACCURACY / 100);
|
|
a_0 = p_vout->p_sys->a_0 * Denom * (ACCURACY / 100);
|
|
for (i_index = 0; i_index < length / p_pic->p->i_pixel_pitch; i_index+=p_pic->p->i_pixel_pitch)
|
|
// for each macropixel
|
|
{
|
|
// first image pixel
|
|
p_vout->p_sys->lambda2[i_row][0][i_index] = CLIP_0A(!i_row ? ACCURACY - (F4(a_2, a_1, i_index) + a_0) / Denom : ACCURACY - (F4(a_2, a_1,(length / p_pic->p->i_pixel_pitch) - i_index) + a_0) / Denom);
|
|
p_vout->p_sys->cstYUV2[i_row][0][i_index] = ((ACCURACY - p_vout->p_sys->lambda2[i_col][0][i_index]) * constantYUV[1]) / ACCURACY;
|
|
p_vout->p_sys->lambda2[i_row][1][i_index] = CLIP_0A(!i_row ? ACCURACY - (F4(a_2, a_1, i_index) + a_0) / Denom : ACCURACY - (F4(a_2, a_1,(length / p_pic->p->i_pixel_pitch) - i_index) + a_0) / Denom);
|
|
p_vout->p_sys->cstYUV2[i_row][1][i_index] = ((ACCURACY - p_vout->p_sys->lambda2[i_row][1][i_index]) * constantYUV[0]) / ACCURACY;
|
|
// second image pixel
|
|
p_vout->p_sys->lambda2[i_row][0][i_index + 1] = CLIP_0A(!i_row ? ACCURACY - (F4(a_2, a_1, i_index + 1) + a_0) / Denom : ACCURACY - (F4(a_2, a_1,(length / p_pic->p->i_pixel_pitch) - (i_index + 1)) + a_0) / Denom);
|
|
p_vout->p_sys->cstYUV2[i_row][0][i_index + 1] = ((ACCURACY - p_vout->p_sys->lambda2[i_row][0][i_index]) * constantYUV[1]) / ACCURACY;
|
|
p_vout->p_sys->lambda2[i_row][1][i_index + 1] = p_vout->p_sys->lambda2[i_row][1][i_index];
|
|
p_vout->p_sys->cstYUV2[i_row][1][i_index + 1] = p_vout->p_sys->cstYUV2[i_row][1][i_index];
|
|
}
|
|
break;
|
|
default :
|
|
break;
|
|
}
|
|
}
|
|
p_out -= TopOffset * p_outpic->p[i_plane].i_pitch;
|
|
#ifndef GAMMA
|
|
for (i_index = 0; i_index < length / p_pic->p->i_pixel_pitch; i_index++)
|
|
for (i_index2 = 0; i_index2 < i_copy_pitch; i_index2++)
|
|
*(p_out + (i_index * p_outpic->p[i_plane].i_pitch) + i_index2) = (p_vout->p_sys->lambda2[i_row][i_index2 % p_pic->p->i_pixel_pitch][i_index] *
|
|
(*(p_out + (i_index * p_outpic->p[i_plane].i_pitch) + i_index2))) / ACCURACY +
|
|
p_vout->p_sys->cstYUV2[i_row][i_index2 % p_pic->p->i_pixel_pitch][i_index];
|
|
#else
|
|
for (i_index = 0; i_index < length / p_pic->p->i_pixel_pitch; i_index++)
|
|
for (i_index2 = 0; i_index2 < i_copy_pitch; i_index2++)
|
|
*(p_out + (i_index * p_outpic->p[i_plane].i_pitch) + i_index2) = p_vout->p_sys->LUT[i_index % p_pic->p->i_pixel_pitch][(p_vout->p_sys->lambda2[i_row][i_index2 % p_pic->p->i_pixel_pitch][i_index] *
|
|
(*(p_out + (i_index * p_outpic->p[i_plane].i_pitch) + i_index2))) / ACCURACY +
|
|
p_vout->p_sys->cstYUV2[i_row][i_index2 % p_pic->p->i_pixel_pitch][i_index]][i_index / p_pic->p->i_pixel_pitch];
|
|
|
|
#endif
|
|
p_out += TopOffset * p_outpic->p[i_plane].i_pitch;
|
|
}
|
|
#endif
|
|
// bug for wall filter : fix by CC
|
|
// pi_left_skip[i_plane] += i_out_pitch;
|
|
pi_left_skip[i_plane] += i_copy_pitch;
|
|
}
|
|
|
|
vout_UnlinkPicture( p_vout->p_sys->pp_vout[ i_vout ].p_vout,
|
|
p_outpic );
|
|
vout_DisplayPicture( p_vout->p_sys->pp_vout[ i_vout ].p_vout,
|
|
p_outpic );
|
|
i_vout++;
|
|
}
|
|
|
|
for( i_plane = 0 ; i_plane < p_pic->i_planes ; i_plane++ )
|
|
{
|
|
pi_top_skip[i_plane] += p_vout->p_sys->pp_vout[ i_vout ].i_height
|
|
* p_pic->p[i_plane].i_lines
|
|
/ p_vout->output.i_height
|
|
* p_pic->p[i_plane].i_pitch;
|
|
}
|
|
}
|
|
#ifdef OVERLAP
|
|
if (p_vout->p_sys->b_has_changed) p_vout->p_sys->b_has_changed = VLC_FALSE;
|
|
#endif
|
|
}
|
|
#endif
|
|
|
|
|
|
/*****************************************************************************
|
|
* RemoveAllVout: destroy all the child video output threads
|
|
*****************************************************************************/
|
|
static void RemoveAllVout( vout_thread_t *p_vout )
|
|
{
|
|
while( p_vout->p_sys->i_vout )
|
|
{
|
|
--p_vout->p_sys->i_vout;
|
|
if( p_vout->p_sys->pp_vout[ p_vout->p_sys->i_vout ].b_active )
|
|
{
|
|
DEL_CALLBACKS(
|
|
p_vout->p_sys->pp_vout[ p_vout->p_sys->i_vout ].p_vout,
|
|
SendEvents );
|
|
vlc_object_detach(
|
|
p_vout->p_sys->pp_vout[ p_vout->p_sys->i_vout ].p_vout );
|
|
vout_Destroy(
|
|
p_vout->p_sys->pp_vout[ p_vout->p_sys->i_vout ].p_vout );
|
|
}
|
|
}
|
|
}
|
|
|
|
/*****************************************************************************
|
|
* SendEvents: forward mouse and keyboard events to the parent p_vout
|
|
*****************************************************************************/
|
|
static int SendEvents( vlc_object_t *p_this, char const *psz_var,
|
|
vlc_value_t oldval, vlc_value_t newval, void *_p_vout )
|
|
{
|
|
vout_thread_t *p_vout = (vout_thread_t *)_p_vout;
|
|
int i_vout;
|
|
vlc_value_t sentval = newval;
|
|
|
|
/* Find the video output index */
|
|
for( i_vout = 0; i_vout < p_vout->p_sys->i_vout; i_vout++ )
|
|
{
|
|
if( p_this == (vlc_object_t *)p_vout->p_sys->pp_vout[ i_vout ].p_vout )
|
|
{
|
|
break;
|
|
}
|
|
}
|
|
|
|
if( i_vout == p_vout->p_sys->i_vout )
|
|
{
|
|
return VLC_EGENERIC;
|
|
}
|
|
|
|
/* Translate the mouse coordinates */
|
|
if( !strcmp( psz_var, "mouse-x" ) )
|
|
{
|
|
#ifdef OVERLAP
|
|
int i_overlap = ((p_vout->p_sys->i_col > 2) ? 0 : 2 * p_vout->p_sys->i_halfLength);
|
|
sentval.i_int += (p_vout->output.i_width - i_overlap)
|
|
#else
|
|
sentval.i_int += p_vout->output.i_width
|
|
#endif
|
|
* (i_vout % p_vout->p_sys->i_col)
|
|
/ p_vout->p_sys->i_col;
|
|
}
|
|
else if( !strcmp( psz_var, "mouse-y" ) )
|
|
{
|
|
#ifdef OVERLAP
|
|
int i_overlap = ((p_vout->p_sys->i_row > 2) ? 0 : 2 * p_vout->p_sys->i_halfHeight);
|
|
sentval.i_int += (p_vout->output.i_height - i_overlap)
|
|
#else
|
|
sentval.i_int += p_vout->output.i_height
|
|
#endif
|
|
//bug fix in Wall plug-in
|
|
// * (i_vout / p_vout->p_sys->i_row)
|
|
* (i_vout / p_vout->p_sys->i_col)
|
|
/ p_vout->p_sys->i_row;
|
|
}
|
|
|
|
var_Set( p_vout, psz_var, sentval );
|
|
|
|
return VLC_SUCCESS;
|
|
}
|
|
|
|
/*****************************************************************************
|
|
* SendEventsToChild: forward events to the child/children vout
|
|
*****************************************************************************/
|
|
static int SendEventsToChild( vlc_object_t *p_this, char const *psz_var,
|
|
vlc_value_t oldval, vlc_value_t newval, void *p_data )
|
|
{
|
|
|
|
vout_thread_t *p_vout = (vout_thread_t *)p_this;
|
|
int i_row, i_col, i_vout = 0;
|
|
|
|
for( i_row = 0; i_row < p_vout->p_sys->i_row; i_row++ )
|
|
{
|
|
for( i_col = 0; i_col < p_vout->p_sys->i_col; i_col++ )
|
|
{
|
|
var_Set( p_vout->p_sys->pp_vout[ i_vout ].p_vout, psz_var, newval);
|
|
if( !strcmp( psz_var, "fullscreen" ) ) break;
|
|
i_vout++;
|
|
}
|
|
}
|
|
|
|
return VLC_SUCCESS;
|
|
}
|