Patch by Steinar Gunderson.
#include <framework/mlt.h>
#include <string.h>
#include <framework/mlt.h>
#include <string.h>
#include <assert.h>
#include "glsl_manager.h"
#include <movit/white_balance_effect.h>
#include <assert.h>
#include "glsl_manager.h"
#include <movit/white_balance_effect.h>
+static double srgb8_to_linear(int c)
+{
+ double x = c / 255.0f;
+ if (x < 0.04045f) {
+ return (1.0/12.92f) * x;
+ } else {
+ return pow((x + 0.055) * (1.0/1.055f), 2.4);
+ }
+}
+
static int get_image( mlt_frame frame, uint8_t **image, mlt_image_format *format, int *width, int *height, int writable )
{
mlt_filter filter = (mlt_filter) mlt_frame_pop_service( frame );
static int get_image( mlt_frame frame, uint8_t **image, mlt_image_format *format, int *width, int *height, int writable )
{
mlt_filter filter = (mlt_filter) mlt_frame_pop_service( frame );
mlt_position length = mlt_filter_get_length2( filter, frame );
int color_int = mlt_properties_anim_get_int( properties, "neutral_color", position, length );
RGBTriplet color(
mlt_position length = mlt_filter_get_length2( filter, frame );
int color_int = mlt_properties_anim_get_int( properties, "neutral_color", position, length );
RGBTriplet color(
- float((color_int >> 24) & 0xff) / 255.0f,
- float((color_int >> 16) & 0xff) / 255.0f,
- float((color_int >> 8) & 0xff) / 255.0f
+ srgb8_to_linear((color_int >> 24) & 0xff),
+ srgb8_to_linear((color_int >> 16) & 0xff),
+ srgb8_to_linear((color_int >> 8) & 0xff)
);
bool ok = effect->set_vec3( "neutral_color", (float*) &color );
ok |= effect->set_float( "output_color_temperature",
);
bool ok = effect->set_vec3( "neutral_color", (float*) &color );
ok |= effect->set_float( "output_color_temperature",