* /author Zachary Drew, Copyright 2005
*
* Currently only uses Gamma data for comparisonon (bug or feature?)
- * Vector optimization coming soon.
+ * SSE optimized where available.
*
* Vector orientation: The vector data that is generated for the current frame specifies
* the motion from the previous frame to the current frame. To know how a macroblock
#include <sys/time.h>
#include <unistd.h>
+#ifdef USE_SSE
#include "sad_sse.h"
+#endif
#define NDEBUG
#include <assert.h>
#define DIAMOND_SEARCH 0x0
#define FULL_SEARCH 0x1
-#define SHIFT 8
+#define SHIFT 8
#define MIN(a,b) ((a) > (b) ? (b) : (a))
#define ABS(a) ((a) >= 0 ? (a) : (-(a)))
int mb_w, mb_h;
int xstride, ystride;
uint8_t *cache_image; // Copy of current frame
- uint8_t *former_image; // Copy of former frame
+ uint8_t *former_image; // Copy of former frame
int search_method;
int skip_prediction;
int shot_change;
};
// This is used to constrains pixel operations between two blocks to be within the image boundry
-inline static int constrain( int *x, int *y, int *w, int *h,
+inline static int constrain( int *x, int *y, int *w, int *h,
const int dx, const int dy,
const int left, const int right,
const int top, const int bottom)
else if( *y + *h > bottom || y2 + *h > bottom )
h_remains = bottom - ((*y > y2) ? *y : y2);
- if( w_remains == *w && h_remains == *h ) return penalty;
+ if( w_remains == *w && h_remains == *h ) return penalty;
if( w_remains <= 0 || h_remains <= 0) return 0; // Block is clipped out of existance
penalty = (*w * *h * penalty)
/ ( w_remains * h_remains); // Recipricol of the fraction of the block that remains
/** /brief Reference Sum of Absolute Differences comparison function
*
*/
-inline static int sad_reference( uint8_t *block1, uint8_t *block2, const int xstride, const int ystride, const int w, const int h )
+static int sad_reference( uint8_t *block1, uint8_t *block2, const int xstride, const int ystride, const int w, const int h )
{
int i, j, score = 0;
for ( j = 0; j < h; j++ ){
int score;
- // Default comparison may be overridden by the slower, more capable reference comparison
+ // Default comparison may be overridden by the slower, more capable reference comparison
int (*cmp)(uint8_t *, uint8_t *, int, int, int, int) = c->compare_optimized;
// vector displacement limited has been exceeded
// Some gotchas
if( penalty == 0 ) // Clipped out of existance: Return worst score
return MAX_MSAD;
- else if( penalty != 1<<SHIFT ) // Nonstandard macroblock dimensions: Disable SIMD optimizizations.
+ else if( penalty != 1<<SHIFT ) // Nonstandard macroblock dimensions: Disable SIMD optimizizations.
cmp = c->compare_reference;
// Calculate the memory locations of the macroblocks
- block1 += x * c->xstride + y * c->ystride;
- block2 += (x+dx) * c->xstride + (y+dy) * c->ystride;
+ block1 += x * c->xstride + y * c->ystride;
+ block2 += (x+dx) * c->xstride + (y+dy) * c->ystride;
- #ifdef DEBUG_ASM
+ #ifdef DEBUG_ASM
if( penalty == 1<<SHIFT ){
score = c->compare_reference( block1, block2, c->xstride, c->ystride, mb_w, mb_h );
int score2 = c->compare_optimized( block1, block2, c->xstride, c->ystride, mb_w, mb_h );
// Keep track of best and former best candidates
motion_vector best, former;
+ best.dx = 0;
+ best.dy = 0;
+ former.dx = 0;
+ former.dy = 0;
// The direction of the refinement needs to be known
motion_vector current;
// Loop through the search pattern
while( 1 ) {
-
+
current.dx = result->dx;
current.dy = result->dy;
candidates[1].dy = result->dy + former.dy;
i = 2;
}
-
+
former.dx = best.dx; former.dy = best.dy; // Keep track of new former best
}
-
- check_candidates ( ref, candidate_base, x, y, candidates, i, 1, result, c );
+
+ check_candidates ( ref, candidate_base, x, y, candidates, i, 1, result, c );
// Which candidate was the best?
best.dx = result->dx - current.dx;
}
}
-/* /brief Full (brute) search
+/* /brief Full (brute) search
* Operates on a single macroblock
-*/
+*/
__attribute__((used))
static void full_search(
uint8_t *ref, //<! Image data from previous frame
int ncompare (const void * a, const void * b)
{
- return ( *(int*)a - *(int*)b );
+ return ( *(const int*)a - *(const int*)b );
}
// motion vector denoising
{
xvalues[n ] = CURRENT(i+1,j)->dx; // Right
yvalues[n++] = CURRENT(i+1,j)->dy;
-
-
+
+
if( j > c->top_mb ) {
xvalues[n ] = CURRENT(i+1,j-1)->dx; // Upper Right
yvalues[n++] = CURRENT(i+1,j-1)->dy;
// For every macroblock, perform motion vector estimation
for( i = c->left_mb; i <= c->right_mb; i++ ){
- for( j = c->top_mb; j <= c->bottom_mb; j++ ){
+ for( j = c->top_mb; j <= c->bottom_mb; j++ ){
here = CURRENT(i,j);
here->valid = 1;
candidates[n ].dx = FORMER(i,j-1)->dx;
candidates[n++].dy = FORMER(i,j-1)->dy;
}
-
+
// Left of colocated
if( i > c->prev_left_mb ){// && COL_LEFT->valid ){
candidates[n ].dx = FORMER(i-1,j)->dx;
candidates[n++].dy = FORMER(i-1,j)->dy;
}
-
+
// Right of colocated
if( i < c->prev_right_mb ){// && COL_RIGHT->valid ){
candidates[n ].dx = FORMER(i+1,j)->dx;
candidates[n++].dy = FORMER(i+1,j)->dy;
}
-
+
// Bottom of colocated
if( j < c->prev_bottom_mb ){// && COL_BOTTOM->valid ){
candidates[n ].dx = FORMER(i,j+1)->dx;
// Top-Right, macroblocks not in the right row
if ( i < c->right_mb ){// && TOP_RIGHT->valid ) {
candidates[n ].dx = CURRENT(i+1,j-1)->dx;
- candidates[n++].dy = CURRENT(i+1,j-1)->dy;
+ candidates[n++].dy = CURRENT(i+1,j-1)->dy;
}
}
/* Median predictor vector (median of left, top, and top right adjacent vectors) */
if ( i > c->left_mb && j > c->top_mb && i < c->right_mb
)//&& LEFT->valid && TOP->valid && TOP_RIGHT->valid )
- {
+ {
candidates[n ].dx = median_predictor( CURRENT(i-1,j)->dx, CURRENT(i,j-1)->dx, CURRENT(i+1,j-1)->dx);
candidates[n++].dy = median_predictor( CURRENT(i-1,j)->dy, CURRENT(i,j-1)->dy, CURRENT(i+1,j-1)->dy);
}
int x = i * c->mb_w;
int y = j * c->mb_h;
- check_candidates ( to, from, x, y, candidates, n, 0, here, c );
+ check_candidates ( to, from, x, y, candidates, n, 0, here, c );
#ifndef FULLSEARCH
- diamond_search( to, from, x, y, here, c);
+ diamond_search( to, from, x, y, here, c);
#else
- full_search( to, from, x, y, here, c);
+ full_search( to, from, x, y, here, c);
#endif
assert( x + c->mb_w + here->dx > 0 ); // All macroblocks must have area > 0
} /* End column loop */
} /* End row loop */
+#ifdef USE_SSE
asm volatile ( "emms" );
+#endif
#ifdef COUNT_COMPARES
fprintf(stderr, "%d comparisons per block were made", compares/count);
int i, j, count = 0;
for ( i = c->left_mb; i <= c->right_mb; i++ ){
- for ( j = c->top_mb; j <= c->bottom_mb; j++ ){
+ for ( j = c->top_mb; j <= c->bottom_mb; j++ ){
count++;
c->comparison_average += CURRENT(i,j)->msad;
static void init_optimizations( struct motion_est_context_s *c )
{
switch(c->mb_w){
+#ifdef USE_SSE
case 4: if(c->mb_h == 4) c->compare_optimized = sad_sse_422_luma_4x4;
else c->compare_optimized = sad_sse_422_luma_4w;
break;
break;
case 64: c->compare_optimized = sad_sse_422_luma_64w;
break;
+#endif
default: c->compare_optimized = sad_reference;
break;
}
image[n+2] = 79;
image[n+3] = 237;
}
-
+
}
static void show_residual( uint8_t *result, struct motion_est_context_s *c )
// set_red(result,c);
- for( j = c->top_mb; j <= c->bottom_mb; j++ ){
+ for( j = c->top_mb; j <= c->bottom_mb; j++ ){
for( i = c->left_mb; i <= c->right_mb; i++ ){
dx = CURRENT(i,j)->dx;
x = i * w;
y = j * h;
- // Denoise function caused some blocks to be completely clipped, ignore them
+ // Denoise function caused some blocks to be completely clipped, ignore them
if (constrain( &x, &y, &w, &h, dx, dy, 0, c->width, 0, c->height) == 0 )
continue;
int tx,ty;
for( i = c->left_mb; i <= c->right_mb; i++ ){
- for( j = c->top_mb; j <= c->bottom_mb; j++ ){
+ for( j = c->top_mb; j <= c->bottom_mb; j++ ){
dx = CURRENT(i,j)->dx;
dy = CURRENT(i,j)->dy;
x = i * w;
y = j * h;
- // Denoise function caused some blocks to be completely clipped, ignore them
+ // Denoise function caused some blocks to be completely clipped, ignore them
if (constrain( &x, &y, &w, &h, dx, dy, 0, c->width, 0, c->height) == 0 )
continue;
c->mv_size = c->mv_buffer_width * c->mv_buffer_height * sizeof(struct motion_vector_s);
// Allocate the motion vector buffers
- c->former_vectors = mlt_pool_alloc( c->mv_size );
- c->current_vectors = mlt_pool_alloc( c->mv_size );
- c->denoise_vectors = mlt_pool_alloc( c->mv_size );
+ c->former_vectors = mlt_pool_alloc( c->mv_size );
+ c->current_vectors = mlt_pool_alloc( c->mv_size );
+ c->denoise_vectors = mlt_pool_alloc( c->mv_size );
// Register motion buffers for destruction
mlt_properties_set_data( properties, "current_motion_vectors", (void *)c->current_vectors, 0, mlt_pool_release, NULL );
case mlt_image_yuv422:
c->xstride = 2;
c->ystride = c->xstride * *width;
- break;
+ break;
default:
// I don't know
fprintf(stderr, "\"I am unfamiliar with your new fangled pixel format!\" -filter_motion_est\n");
c->bounds.h = *height;
}
- // If video is advancing, run motion vector algorithm and etc...
+ // If video is advancing, run motion vector algorithm and etc...
if( c->former_frame_position + 1 == c->current_frame_position )
{
/** Constructor for the filter.
*/
-mlt_filter filter_motion_est_init( char *arg )
+mlt_filter filter_motion_est_init( mlt_profile profile, mlt_service_type type, const char *id, char *arg )
{
mlt_filter this = mlt_filter_new( );
if ( this != NULL )
projects / melted / blobdiff