void compute_curve_gamma_table_type_parametric(float gamma_table[256], float parameter[7], int count)
{
size_t X;
float interval;
float a, b, c, e, f;
float y = parameter[0];
if (count == 0) {
a = 1;
b = 0;
c = 0;
e = 0;
f = 0;
interval = -INFINITY;
} else if(count == 1) {
a = parameter[1];
b = parameter[2];
c = 0;
e = 0;
f = 0;
interval = -1 * parameter[2] / parameter[1];
} else if(count == 2) {
a = parameter[1];
b = parameter[2];
c = 0;
e = parameter[3];
f = parameter[3];
interval = -1 * parameter[2] / parameter[1];
} else if(count == 3) {
a = parameter[1];
b = parameter[2];
c = parameter[3];
e = -c;
f = 0;
interval = parameter[4];
} else if(count == 4) {
a = parameter[1];
b = parameter[2];
c = parameter[3];
e = parameter[5] - c;
f = parameter[6];
interval = parameter[4];
} else {
assert(0 && "invalid parametric function type.");
a = 1;
b = 0;
c = 0;
e = 0;
f = 0;
interval = -INFINITY;
}
for (X = 0; X < 256; X++) {
if (X >= interval) {
// XXX The equations are not exactly as definied in the spec but are
// algebraic equivilent.
// TODO Should division by 255 be for the whole expression.
gamma_table[X] = clamp_float(pow(a * X / 255. + b, y) + c + e);
} else {
gamma_table[X] = clamp_float(c * X / 255. + f);
}
}
}
int composite_alpha_rgbaz88 \
( BYTE* over_image, \
BYTE* under_image, \
BYTE* blend_image, \
unsigned int image_size )
#endif
{
unsigned int i;
unsigned int full_image_size;
BYTE* over_image_ptr;
BYTE* under_image_ptr;
BYTE* blend_image_ptr;
float* over_image_f_ptr;
float* under_image_f_ptr;
float* blend_image_f_ptr;
BYTE over_r;
BYTE over_g;
BYTE over_b;
BYTE under_r;
BYTE under_g;
BYTE under_b;
float over_r_f;
float over_g_f;
float over_b_f;
float over_a_f;
float over_z_f;
float under_r_f;
float under_g_f;
float under_b_f;
float under_a_f;
float under_z_f;
float blend_r;
float blend_g;
float blend_b;
float blend_a;
float blend_z;
float one_minus_alpha;
blend_image_ptr = (BYTE *)blend_image;
over_image_ptr = (BYTE *)over_image;
under_image_ptr = (BYTE *)under_image;
blend_image_f_ptr = (float *)blend_image;
over_image_f_ptr = (float *)over_image;
under_image_f_ptr = (float *)under_image;
blend_z = 0;
full_image_size = image_size * RGBAZ88; // 11 BYTES
//=====================================
// Depth Sorting
//=====================================
over_image_f_ptr = (float *)&over_image_ptr[ 3 ];
over_z_f = (float) over_image_f_ptr[ 1 ]; // Z
under_image_f_ptr = (float *)&under_image_ptr[ 3 ];
under_z_f = (float)under_image_f_ptr[ 1 ]; // Z
//=====================================
// Shared Memory Parallelism
//=====================================
#if defined ( _OPENMP )
#pragma omp parallel for \
private( i, one_minus_alpha, \
over_r_f, over_g_f, over_b_f, over_a_f, over_z_f, \
under_r, under_g, under_b_f, under_a_f, under_z_f,\
blend_r, blend_g, blend_b, blend_a, blend_z )
#endif
for ( i = 0; i < full_image_size; i += RGBAZ88 ) // SKIP 11 BYTES
{
over_r = (BYTE)over_image_ptr[ i ]; // R
over_g = (BYTE)over_image_ptr[ i + 1 ]; // G
over_b = (BYTE)over_image_ptr[ i + 2 ]; // B
over_image_f_ptr = (float *)&over_image_ptr[ i + 3 ];
over_a_f = (float)over_image_f_ptr[ 0 ]; // A
over_z_f = (float)over_image_f_ptr[ 1 ]; // Z
under_r = (BYTE)under_image_ptr[ i ]; // R
under_g = (BYTE)under_image_ptr[ i + 1 ]; // G
under_b = (BYTE)under_image_ptr[ i + 2 ]; // B
under_image_f_ptr = (float *)&under_image_ptr[ i + 3 ];
under_a_f = (float)under_image_f_ptr[ 0 ]; // A
under_z_f = (float)under_image_f_ptr[ 1 ]; // Z
// Depth sorting if necessary
if ( over_z_f > under_z_f )
{
blend_r = over_r;
blend_g = over_g;
blend_b = over_b;
blend_a = over_a_f;
blend_z = over_z_f;
over_r = under_r;
over_g = under_g;
over_b = under_b;
over_a_f = under_a_f;
over_z_f = under_z_f;
under_r = blend_r;
under_g = blend_g;
under_b = blend_b;
under_a_f = blend_a;
under_z_f = blend_z;
}
// Convert BYTE to Float (Range: 0.0 - 1.0)
over_r_f = (float)( over_r / 255.0f );
over_g_f = (float)( over_g / 255.0f );
over_b_f = (float)( over_b / 255.0f );
under_r_f = (float)( under_r / 255.0f );
under_g_f = (float)( under_g / 255.0f );
under_b_f = (float)( under_b / 255.0f );
// Pre-calculate 1 - Src_A
one_minus_alpha = 1.0f - over_a_f;
// Calculate Final alpha value
blend_a = (float) ( over_a_f + ( under_a_f * one_minus_alpha ));
blend_r = (float)( over_r_f + ( under_r_f * one_minus_alpha ));
blend_g = (float)( over_g_f + ( under_g_f * one_minus_alpha ));
blend_b = (float)( over_b_f + ( under_b_f * one_minus_alpha ));
// =======================================================
blend_r = clamp_float( blend_r, 0.0, 1.0 );
blend_g = clamp_float( blend_g, 0.0, 1.0 );
blend_b = clamp_float( blend_b, 0.0, 1.0 );
blend_a = clamp_float( blend_a, 0.0, 1.0 );
blend_image_ptr[ i ] = (BYTE)( blend_r * 255 ); // R
blend_image_ptr[ i + 1 ] = (BYTE)( blend_g * 255 ); // G
blend_image_ptr[ i + 2 ] = (BYTE)( blend_b * 255 ); // B
blend_image_f_ptr = (float *)&blend_image_ptr[ i + 3 ];
blend_image_f_ptr[ 0 ] = (float)blend_a; // A
blend_image_f_ptr[ 1 ] = (float)over_z_f; // Z
}
return EXIT_SUCCESS;
}
/**
* input_overlay_poll:
* @out : Polled output data.
* @norm_x : Normalized X coordinate.
* @norm_y : Normalized Y coordinate.
*
* Polls input overlay.
*
* @norm_x and @norm_y are the result of
* input_translate_coord_viewport().
**/
static void input_overlay_poll(input_overlay_state_t *out,
int16_t norm_x, int16_t norm_y)
{
size_t i;
float x, y;
input_overlay_t *ol = overlay_ptr;
memset(out, 0, sizeof(*out));
if (!ol->enable)
{
ol->blocked = false;
return;
}
/* norm_x and norm_y is in [-0x7fff, 0x7fff] range,
* like RETRO_DEVICE_POINTER. */
x = (float)(norm_x + 0x7fff) / 0xffff;
y = (float)(norm_y + 0x7fff) / 0xffff;
x -= ol->active->mod_x;
y -= ol->active->mod_y;
x /= ol->active->mod_w;
y /= ol->active->mod_h;
for (i = 0; i < ol->active->size; i++)
{
float x_dist, y_dist;
struct overlay_desc *desc = &ol->active->descs[i];
if (!desc)
continue;
if (!inside_hitbox(desc, x, y))
continue;
desc->updated = true;
x_dist = x - desc->x;
y_dist = y - desc->y;
switch (desc->type)
{
case OVERLAY_TYPE_BUTTONS:
{
uint64_t mask = desc->key_mask;
out->buttons |= mask;
if (mask & (UINT64_C(1) << RARCH_OVERLAY_NEXT))
ol->next_index = desc->next_index;
}
break;
case OVERLAY_TYPE_KEYBOARD:
if (desc->key_mask < RETROK_LAST)
OVERLAY_SET_KEY(out, desc->key_mask);
break;
default:
{
float x_val = x_dist / desc->range_x;
float y_val = y_dist / desc->range_y;
float x_val_sat = x_val / desc->analog_saturate_pct;
float y_val_sat = y_val / desc->analog_saturate_pct;
unsigned int base = (desc->type == OVERLAY_TYPE_ANALOG_RIGHT) ? 2 : 0;
out->analog[base + 0] = clamp_float(x_val_sat, -1.0f, 1.0f) * 32767.0f;
out->analog[base + 1] = clamp_float(y_val_sat, -1.0f, 1.0f) * 32767.0f;
}
break;
}
if (desc->movable)
{
desc->delta_x = clamp_float(x_dist, -desc->range_x, desc->range_x)
* ol->active->mod_w;
desc->delta_y = clamp_float(y_dist, -desc->range_y, desc->range_y)
* ol->active->mod_h;
}
}
if (!out->buttons)
ol->blocked = false;
else if (ol->blocked)
memset(out, 0, sizeof(*out));
}
int composite_alpha_rgba64 \
( BYTE* over_image, \
BYTE* under_image, \
BYTE* blend_image, \
unsigned int image_size )
#endif
{
unsigned int i;
unsigned int full_image_size;
BYTE* blend_image_ptr;
float* over_image_f_ptr;
float* under_image_f_ptr;
float* blend_image_f_ptr;
BYTE under_r;
BYTE under_g;
BYTE under_b;
float under_a;
BYTE over_r;
BYTE over_g;
BYTE over_b;
float over_a;
float over_r_f;
float over_g_f;
float over_b_f;
float over_a_f;
float under_r_f;
float under_g_f;
float under_b_f;
float under_a_f;
float blend_r;
float blend_g;
float blend_b;
float blend_a;
float one_minus_alpha;
blend_image_ptr = (BYTE *)blend_image;
blend_image_f_ptr = (float *)blend_image;
over_image_f_ptr = (float *)over_image;
under_image_f_ptr = (float *)under_image;
//=====================================
// Shared Memory Parallelism
//=====================================
full_image_size = image_size * RGBA64;
#if defined ( _OPENMP )
#pragma omp parallel for \
private( i, one_minus_alpha, \
over_r, over_g, over_b, over_a, \
under_r, under_g, under_b, under_a, \
blend_r, blend_g, blend_b, blend_a )
#endif
for ( i = 0; i < full_image_size; i += RGBA64 ) // SKIP 8 elements
{
// Separate R, G, B and A values of both
// the foreground and background colors
over_r = (BYTE)over_image[ i ];
over_g = (BYTE)over_image[ i + 1 ];
over_b = (BYTE)over_image[ i + 2 ];
over_image_f_ptr = (float *)&over_image[ i + 4 ];
over_a = (float)*over_image_f_ptr;
under_r = (BYTE)under_image[ i ];
under_g = (BYTE)under_image[ i + 1 ];
under_b = (BYTE)under_image[ i + 2 ];
under_image_f_ptr = (float *)&under_image[ i + 4 ];
under_a = (float)*under_image_f_ptr;
// Convert BYTE to Float (Range: 0.0 - 1.0)
over_r_f = (float)( over_r / 255.0f );
over_g_f = (float)( over_g / 255.0f );
over_b_f = (float)( over_b / 255.0f );
over_a_f = (float) over_a ;
under_r_f = (float)( under_r / 255.0f );
under_g_f = (float)( under_g / 255.0f );
under_b_f = (float)( under_b / 255.0f );
under_a_f = (float) under_a ;
// Pre-calculate 1 - Src_A
one_minus_alpha = 1.0f - over_a_f;
// Calculate Final alpha value
blend_a = (float) ( over_a_f + ( under_a_f * one_minus_alpha ));
blend_r = (float)( over_r_f + ( under_r_f * one_minus_alpha ));
blend_g = (float)( over_g_f + ( under_g_f * one_minus_alpha ));
blend_b = (float)( over_b_f + ( under_b_f * one_minus_alpha ));
// =======================================================
// Clamp RGB component values if necessary
blend_r = clamp_float( blend_r, 0.0, 1.0 );
blend_g = clamp_float( blend_g, 0.0, 1.0 );
blend_b = clamp_float( blend_b, 0.0, 1.0 );
blend_a = clamp_float( blend_a, 0.0, 1.0 );
blend_image_ptr[ i ] = (BYTE)( blend_r * 255 ); // R
blend_image_ptr[ i + 1 ] = (BYTE)( blend_g * 255 ); // G
blend_image_ptr[ i + 2 ] = (BYTE)( blend_b * 255 ); // B
blend_image_ptr[ i + 3 ] = (BYTE)0; // X
blend_image_f_ptr = (float *)&blend_image_ptr[ i + 4 ]; // A
*blend_image_f_ptr = (float) blend_a;
}
return EXIT_SUCCESS;
}
int composite_alpha_rgbaz160 \
( float* over_image, \
float* under_image, \
float* blend_image, \
unsigned int image_size )
#endif
{
unsigned int i;
unsigned int full_image_size;
float under_r;
float under_g;
float under_b;
float under_a;
float under_z;
float over_r;
float over_g;
float over_b;
float over_a;
float over_z;
float blend_r;
float blend_g;
float blend_b;
float blend_a;
float blend_z;
float one_minus_alpha;
blend_z = 0.0f;
full_image_size = image_size * RGBAZ; // 5 elements
//=====================================
// Shared Memory Parallelism
//=====================================
#if defined ( _OPENMP )
#pragma omp parallel for \
private( i, one_minus_alpha, \
over_r, over_g, over_b, over_a, over_z, \
under_r, under_g, under_b, under_a, under_z, \
blend_r, blend_g, blend_b, blend_a, blend_z )
#endif
for ( i = 0; i < full_image_size; i += RGBAZ ) // SKIP 5 elements(FLOAT)
{
// Separate R, G, B and A values of both
// the foreground and background colors
over_r = (float)over_image[ i ];
over_g = (float)over_image[ i + 1 ];
over_b = (float)over_image[ i + 2 ];
over_a = (float)over_image[ i + 3 ];
over_z = (float)over_image[ i + 4 ];
under_r = (float)under_image[ i ];
under_g = (float)under_image[ i + 1 ];
under_b = (float)under_image[ i + 2 ];
under_a = (float)under_image[ i + 3 ];
under_z = (float)under_image[ i + 4 ];
// Depth sorting if necessary
if ( over_z > under_z )
{
blend_r = over_r;
blend_g = over_g;
blend_b = over_b;
blend_a = over_a;
blend_z = over_z;
over_r = under_r;
over_g = under_g;
over_b = under_b;
over_a = under_a;
over_z = under_z;
under_r = blend_r;
under_g = blend_g;
under_b = blend_b;
under_a = blend_a;
under_z = blend_z;
}
// =======================================================
// Pre-calculate 1 - Src_A
one_minus_alpha = (float)(1.0f - over_a);
// Calculate Final alpha value
blend_a = (float)( over_a + ( under_a * one_minus_alpha ));
blend_r = (float)( over_r + ( under_r * one_minus_alpha ));
blend_g = (float)( over_g + ( under_g * one_minus_alpha ));
blend_b = (float)( over_b + ( under_b * one_minus_alpha ));
// =======================================================
blend_r = clamp_float( blend_r, 0.0, 1.0 );
blend_g = clamp_float( blend_g, 0.0, 1.0 );
blend_b = clamp_float( blend_b, 0.0, 1.0 );
blend_a = clamp_float( blend_a, 0.0, 1.0 );
blend_image[ i ] = (float)blend_r;
blend_image[ i + 1 ] = (float)blend_g;
blend_image[ i + 2 ] = (float)blend_b;
blend_image[ i + 3 ] = (float)blend_a;
blend_image[ i + 4 ] = (float)over_z;
}
return EXIT_SUCCESS;
}
int composite_alpha_rgba128 \
( float* over_image, \
float* under_image, \
float* blend_image, \
unsigned int image_size )
#endif
{
unsigned int i;
unsigned int full_image_size;
float under_r;
float under_g;
float under_b;
float under_a;
float over_r;
float over_g;
float over_b;
float over_a;
float blend_r;
float blend_g;
float blend_b;
float blend_a;
float one_minus_alpha;
//=====================================
// Shared Memory Parallelism
//=====================================
full_image_size = image_size * RGBA; // 4 elements
#if defined ( _OPENMP )
#pragma omp parallel for \
private( i, one_minus_alpha, \
over_r, over_g, over_b, over_a, \
under_r, under_g, under_b, under_a, \
blend_r, blend_g, blend_b, blend_a )
#endif
for ( i = 0; i < full_image_size; i += RGBA ) // SKIP 4 elements(FLOAT)
{
// Separate R, G, B and A values of both
// the foreground and background colors
over_r = (float)over_image[ i ];
over_g = (float)over_image[ i + 1 ];
over_b = (float)over_image[ i + 2 ];
over_a = (float)over_image[ i + 3 ];
under_r = (float)under_image[ i ];
under_g = (float)under_image[ i + 1 ];
under_b = (float)under_image[ i + 2 ];
under_a = (float)under_image[ i + 3 ];
// =============================================
// Eliminate branching for compiler optimization
// =============================================
// Pre-calculate 1 - Src_A
one_minus_alpha = (float)(1.0f - over_a);
// Calculate Final alpha value
blend_a = (float)( over_a + ( under_a * one_minus_alpha ));
blend_r = (float)( over_r + ( under_r * one_minus_alpha ));
blend_g = (float)( over_g + ( under_g * one_minus_alpha ));
blend_b = (float)( over_b + ( under_b * one_minus_alpha ));
// =======================================================
blend_r = clamp_float( blend_r, 0.0, 1.0 );
blend_g = clamp_float( blend_g, 0.0, 1.0 );
blend_b = clamp_float( blend_b, 0.0, 1.0 );
blend_a = clamp_float( blend_a, 0.0, 1.0 );
blend_image[ i ] = (float)blend_r;
blend_image[ i + 1 ] = (float)blend_g;
blend_image[ i + 2 ] = (float)blend_b;
blend_image[ i + 3 ] = (float)blend_a;
}
return EXIT_SUCCESS;
}
void vector3_clamp(vector3* vector, float a, float b){
vector->x = clamp_float(vector->x, a, b);
vector->y = clamp_float(vector->y, a, b);
vector->z = clamp_float(vector->z, a, b);
}
static void tick_camera(struct Input *input, struct Camera *camera, float dt)
{
//
// move
//
vec2 camera_up = vec2_direction(camera->rotation);
vec2 camera_right = vec2_direction(camera->rotation - PI_OVER_TWO);
vec2 move_acceleration = vec2_zero();
if (key_down(GLFW_KEY_W, input))
move_acceleration = vec2_add(move_acceleration, camera_up);
if (key_down(GLFW_KEY_A, input))
move_acceleration = vec2_add(move_acceleration, vec2_negate(camera_right));
if (key_down(GLFW_KEY_S, input))
move_acceleration = vec2_add(move_acceleration, vec2_negate(camera_up));
if (key_down(GLFW_KEY_D, input))
move_acceleration = vec2_add(move_acceleration, camera_right);
if (vec2_length2(move_acceleration) > FLOAT_EPSILON)
{
const float acceleration_magnitude = camera->zoom * 10.0f;
move_acceleration = vec2_normalize(move_acceleration);
move_acceleration = vec2_mul(move_acceleration, acceleration_magnitude);
// v = v0 + (a*t)
camera->move_velocity = vec2_add(camera->move_velocity, vec2_mul(move_acceleration, dt));
// Clamp move velocity.
const float max_move_speed = 5.0f * sqrtf(camera->zoom);
if (vec2_length2(camera->move_velocity) > (max_move_speed * max_move_speed))
camera->move_velocity = vec2_mul(vec2_normalize(camera->move_velocity), max_move_speed);
}
// Number of seconds required for friction to stop movement completely.
// (because velocity is sampled each frame, actual stop time longer than this)
float stop_time = 0.1f;
float inv_stop_time = 1.0f / stop_time;
if (move_acceleration.x == 0.0f)
camera->move_velocity.x -= camera->move_velocity.x * inv_stop_time * dt;
if (move_acceleration.y == 0.0f)
camera->move_velocity.y -= camera->move_velocity.y * inv_stop_time * dt;
// r = r0 + (v0*t) + (a*t^2)/2
camera->position = vec2_add(vec2_add(camera->position, vec2_mul(camera->move_velocity, dt)), vec2_div(vec2_mul(move_acceleration, dt * dt), 2.0f));
//
// zoom
//
float zoom_acceleration = 0.0f;
if (key_down(GLFW_KEY_SPACE, input))
zoom_acceleration += 1.0f;
if (key_down(GLFW_KEY_LEFT_CONTROL, input))
zoom_acceleration -= 1.0f;
zoom_acceleration *= camera->zoom * 50.0f;
if (zoom_acceleration == 0.0f)
camera->zoom_velocity -= camera->zoom_velocity * (1.0f / 0.1f) * dt;
// v = v0 + (a*t)
camera->zoom_velocity += zoom_acceleration * dt;
// r = r0 + (v0*t) + (a*t^2)/2
camera->zoom += (camera->zoom_velocity * dt) + (zoom_acceleration * dt * dt) / 2.0f;
// Clamp zoom velocity.
const float max_zoom_speed = 3.0f * camera->zoom;
camera->zoom_velocity = clamp_float(camera->zoom_velocity, -max_zoom_speed, max_zoom_speed);
float unclamped_zoom = camera->zoom;
camera->zoom = clamp_float(camera->zoom, 1.0f, 1000.0f);
if (camera->zoom != unclamped_zoom)
camera->zoom_velocity = 0.0f;
}