Skip to content

V
vlc-gpu
Commit 7bca0d7b authored by Rocky Bernstein's avatar Rocky Bernstein
Browse files
Options

Attempt to make easier to understand and easier to complete.

parent 78e3a1d6
Hide whitespace changes
Inline Side-by-side
Showing with 85 additions and 81 deletions
modules/codec/ogt/render.c
View file @ 7bca0d7b
......@@ -2,7 +2,7 @@
* render.c : Philips OGT (SVCD Subtitle) renderer
*****************************************************************************
* Copyright (C) 2003 VideoLAN
* $Id: render.c,v 1.7 2004/01/01 15:56:56 rocky Exp $
* $Id: render.c,v 1.8 2004/01/02 04:44:34 rocky Exp $
*
* Author: Rocky Bernstein
* based on code from:
......@@ -123,34 +123,31 @@ static void RenderI420( vout_thread_t *p_vout, picture_t *p_pic,
struct subpicture_sys_t *p_sys = p_spu->p_sys;
dbg_print( (DECODE_DBG_CALL|DECODE_DBG_RENDER),
"spu width: %d, height %d, pitch (%d, %d, %d)",
p_spu->i_width, p_spu->i_height,
"spu width x height: (%dx%d), x-y (%d,%d), yuv pitch (%d,%d,%d)",
p_spu->i_width, p_spu->i_height, p_spu->i_x, p_spu->i_y,
p_pic->Y_PITCH, p_pic->U_PITCH, p_pic->V_PITCH );
p_pixel_base = p_pic->p[Y_PLANE].p_pixels + p_spu->i_x + p_spu->i_width
+ p_pic->p[Y_PLANE].i_pitch * ( p_spu->i_y + p_spu->i_height );
p_pixel_base = p_pic->p[Y_PLANE].p_pixels + p_spu->i_x
+ p_pic->p[Y_PLANE].i_pitch * p_spu->i_y;
i_x_start = p_spu->i_width - p_sys->i_x_end;
i_y_start = p_pic->p[Y_PLANE].i_pitch
* (p_spu->i_height - p_sys->i_y_end );
i_x_start = p_sys->i_x_start;
i_y_start = p_pic->p[Y_PLANE].i_pitch * p_sys->i_y_start;
i_x_end = p_spu->i_width - p_sys->i_x_start;
i_y_end = p_pic->p[Y_PLANE].i_pitch
* (p_spu->i_height - p_sys->i_y_start );
i_x_end = p_sys->i_x_end;
i_y_end = p_pic->p[Y_PLANE].i_pitch * p_sys->i_y_end;
p_source = (ogt_yuvt_t *)p_sys->p_data;
/* Draw until we reach the bottom of the subtitle */
for( i_y = p_spu->i_height * p_pic->p[Y_PLANE].i_pitch ;
i_y ;
i_y -= p_pic->p[Y_PLANE].i_pitch )
for( i_y = 0;
i_y < p_spu->i_height * p_pic->p[Y_PLANE].i_pitch ;
i_y += p_pic->p[Y_PLANE].i_pitch )
{
/* printf("+++begin line: %d,\n", i++); */
/* Draw until we reach the end of the line */
for( i_x = p_spu->i_width ; i_x ; i_x--, p_source++ )
for( i_x = 0; i_x < p_spu->i_width; i_x++, p_source++ )
{
if( b_crop
&& ( i_x < i_x_start || i_x > i_x_end
|| i_y < i_y_start || i_y > i_y_end ) )
......@@ -172,10 +169,10 @@ static void RenderI420( vout_thread_t *p_vout, picture_t *p_pic,
/* Completely opaque. Completely overwrite underlying
pixel with subtitle pixel. */
uint8_t *p_pixel = p_pixel_base - i_x - i_y;
uint16_t i_colprecomp =
(uint16_t) ( p_source->plane[Y_PLANE] * MAX_ALPHA );
*p_pixel = i_colprecomp >> ALPHA_SCALEDOWN;
/* This is the location that's going to get changed.*/
uint8_t *p_pixel = p_pixel_base + i_x + i_y;
*p_pixel = p_source->plane[Y_PLANE];
break;
}
......@@ -184,20 +181,36 @@ static void RenderI420( vout_thread_t *p_vout, picture_t *p_pic,
{
/* Blend in underlying pixel subtitle pixel. */
/* To be able to scale correctly for full opaqueness, we
add 1 to the alpha. This means alpha value 0 won't
be completely transparent and is not correct, but
that's handled in a special case above anyway. */
uint8_t *p_pixel = p_pixel_base - i_x - i_y;
uint16_t i_colprecomp =
(uint16_t) ( p_source->plane[Y_PLANE]
* (uint16_t) (p_source->s.t+1) );
uint16_t i_destalpha = MAX_ALPHA - p_source->s.t;
/* This is the location that's going to get changed.*/
uint8_t *p_pixel = p_pixel_base + i_x + i_y;
/* This is the weighted part of the subtitle. The
color plane is 8 bits and transparancy is 4 bits so
when multiplied we get up to 12 bits.
*/
uint16_t i_sub_color =
(uint16_t) ( p_source->plane[Y_PLANE] *
(uint16_t) (p_source->s.t) );
/* This is the weighted part of the underlying pixel.
For the same reasons above, the result is up to 12
bits. However since the transparancies are
inverses, the sum of i_sub_color and i_pixel_color
will not exceed 12 bits.
*/
uint16_t i_pixel_color =
(uint16_t) ( *p_pixel *
(uint16_t) (MAX_ALPHA - p_source->s.t) ) ;
*p_pixel = ( i_colprecomp +
(uint16_t) (*p_pixel) * i_destalpha )
>> ALPHA_SCALEDOWN;
/* Scale the 12-bit result back down to 8 bits. A
precise scaling after adding the two components,
would divide by one less than a power of 2. However
to simplify and speed things we use a power of
2. This means the boundaries (either all
transparent and all opaque) aren't handled properly.
But we deal with them in special cases above. */
*p_pixel = ( i_sub_color + i_pixel_color ) >> 4;
break;
}
......@@ -206,38 +219,31 @@ static void RenderI420( vout_thread_t *p_vout, picture_t *p_pic,
}
#if FIXED
for ( i_plane = U_PLANE; i_plane <= U_PLANE ; i_plane++ )
#if 0
for ( i_plane = U_PLANE; i_plane <= V_PLANE ; i_plane++ )
{
p_pixel_base = p_pic->p[i_plane].p_pixels
+ ( p_spu->i_x + p_spu->i_width
+ p_pic->p[i_plane].i_pitch * (p_spu->i_y + p_spu->i_height) )/2 ;
p_pixel_base = p_pic->p[i_plane].p_pixels + p_spu->i_x / 2
+ p_pic->p[i_plane].i_pitch * p_spu->i_y / 2;
i_x_start = (p_spu->i_width - p_sys->i_x_end) >> 1;
i_y_start = p_pic->p[i_plane].i_pitch
* (p_spu->i_height - p_sys->i_y_end );
i_x_end = (p_spu->i_width - p_sys->i_x_start ) >> 1;
i_y_end = p_pic->p[i_plane].i_pitch
* (p_spu->i_height - p_sys->i_y_start );
i_x_start = p_sys->i_x_start / 2;
i_y_start = p_pic->p[Y_PLANE].i_pitch * p_sys->i_y_start;
i_x_end = p_sys->i_x_end / 2;
i_y_end = p_pic->p[Y_PLANE].i_pitch * p_sys->i_y_end;
p_source = (ogt_yuvt_t *)p_sys->p_data;
/* Draw until we reach the bottom of the subtitle */
for( i_y = p_spu->i_height * p_pic->p[i_plane].i_pitch ;
i_y > 0;
i_y -= p_pic->p[i_plane].i_pitch )
for( i_y = 0;
i_y < p_spu->i_height * p_pic->p[i_plane].i_pitch ;
i_y += p_pic->p[i_plane].i_pitch )
{
vlc_bool_t high_nibble = VLC_TRUE;
/* printf("+++begin line: %d,\n", i++); */
/* Draw until we reach the end of the line */
for( i_x = p_spu->i_width / 2 ; i_x > 0 ; i_x--, p_source ++ )
for( i_x = 0; i_x < p_spu->i_width; i_x += 2, p_source += 2 )
{
high_nibble = !high_nibble;
if( b_crop
&& ( i_x < i_x_start || i_x > i_x_end
|| i_y < i_y_start || i_y > i_y_end ) )
......@@ -253,26 +259,23 @@ static void RenderI420( vout_thread_t *p_vout, picture_t *p_pic,
case 0x00:
/* Completely transparent. Don't change pixel. */
break;
default:
case MAX_ALPHA:
{
/* Completely opaque. Completely overwrite underlying
pixel with subtitle pixel. */
uint8_t *p_pixel = p_pixel_base - (i_x - i_y) / 2;
uint16_t i_colprecomp =
(uint16_t) (p_source->plane[i_plane] >> 4);
if (high_nibble)
*p_pixel = (i_colprecomp << 4) | ((*p_pixel) & 0x0f);
else
*p_pixel = (i_colprecomp) | ((*p_pixel) & 0xf0);
/* This is the location that's going to get changed.*/
uint8_t *p_pixel = p_pixel_base + i_y/2 + i_x/2;
uint8_t alpha = p_source->plane[i_plane];
*p_pixel = ( ( *p_pixel * ( 0xFF - alpha ) ) >> 8 ) +
( 0x80 * alpha >> 8 );
break;
}
default:
#if 0
{
/* Blend in underlying pixel subtitle pixel. */
......@@ -281,25 +284,26 @@ static void RenderI420( vout_thread_t *p_vout, picture_t *p_pic,
alpha value 0 won't be completely transparent and
is not correct, but that's handled in a special
case above anyway. */
uint8_t *p_pixel = p_pixel_base - (i_x - i_y) / 2;
uint16_t i_colprecomp = (uint16_t)
(p_source->plane[i_plane]
* (uint16_t) (p_source->s.t+1)) >> (4+4);
uint16_t i_destalpha = (MAX_ALPHA - p_source->s.t) >> 4;
if (high_nibble)
*p_pixel =
(( (i_colprecomp) +
(((uint16_t) *p_pixel * i_destalpha) & 0x0f) ) <<4)
| ((*p_pixel) & 0x0f);
uint8_t *p_pixel = p_pixel_base + i_y/2 + i_x/2;
uint8_t low_val = *p_pixel & 0x0f;
uint8_t high_val = *p_pixel & 0xf0 >> 4;
uint16_t i_colprecomp = (p_source->plane[i_plane]
* (uint16_t) (p_source->s.t+1)) >> 4;
uint8_t i_destalpha = MAX_ALPHA - p_source->s.t;
uint16_t new_low = ( i_colprecomp + low_val*i_destalpha )
/ 2;
uint16_t new_high;
else
*p_pixel =
( (i_colprecomp) +
(((uint16_t) *p_pixel * i_destalpha) & 0x0f) )
| ((*p_pixel) & 0xf0);
p_source++;
i_colprecomp = (p_source->plane[i_plane]
* (uint16_t) (p_source->s.t+1)) >> 4;
i_destalpha = MAX_ALPHA - p_source->s.t;
new_high = ( i_colprecomp + high_val*i_destalpha ) / 2;
*p_pixel = (new_high << 4) | new_low;
break;
}
#endif
}
}
}
......
Markdown is supported
0% or
You are about to add 0 people to the discussion. Proceed with caution.
Finish editing this message first!
Please register or to comment