ShuriZma
/
RetroArch
mirror of https://github.com/libretro/RetroArch.git
1
0
Fork 0
Code Issues Packages Projects Releases Wiki Activity

Lots of cleanups for XVideo. 

Render video @ 2x scale to combat chroma downsampling. Fix 32-bit
rendering.
This commit is contained in:
Themaister 2011-03-19 18:53:37 +01:00
parent 0441725045
commit bb3ab3b49c
1 changed files with 89 additions and 88 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

177
gfx/xvideo.c
View File

@ -49,7 +49,7 @@ struct xv
Atom quit_atom; Atom quit_atom;
bool focus; bool focus;
int port; XvPortID port;
int depth; int depth;
int visualid; int visualid;
@ -83,29 +83,29 @@ static void xv_set_nonblock_state(void *data, bool state)
SSNES_WARN("Failed to set SYNC_TO_VBLANK attribute.\n"); SSNES_WARN("Failed to set SYNC_TO_VBLANK attribute.\n");
} }
static volatile sig_atomic_t g_quit = false; static volatile sig_atomic_t g_quit = 0;
static void sighandler(int sig) static void sighandler(int sig)
{ {
g_quit = true; g_quit = 1;
} }
static void init_yuv_tables(xv_t *xv) static void init_yuv_tables(xv_t *xv)
{ {
xv->ytable = malloc(0x10000); xv->ytable = malloc(0x8000);
xv->utable = malloc(0x10000); xv->utable = malloc(0x8000);
xv->vtable = malloc(0x10000); xv->vtable = malloc(0x8000);
for (unsigned i = 0; i < 0x10000; i++) for (unsigned i = 0; i < 0x8000; i++)
{ {
// Extract RGB555 color data from i // Extract RGB555 color data from i
uint8_t r = (i >> 10) & 31, g = (i >> 5) & 31, b = (i) & 31; uint8_t r = (i >> 10) & 0x1F, g = (i >> 5) & 0x1F, b = (i) & 0x1F;
r = (r << 3) | (r >> 2); //R5->R8 r = (r << 3) | (r >> 2); //R5->R8
g = (g << 3) | (g >> 2); //G5->G8 g = (g << 3) | (g >> 2); //G5->G8
b = (b << 3) | (b >> 2); //B5->B8 b = (b << 3) | (b >> 2); //B5->B8
int y = (int)( +((double)r * 0.257) + ((double)g * 0.504) + ((double)b * 0.098) + 16.0 ); int y = (int)(+((double)r * 0.257) + ((double)g * 0.504) + ((double)b * 0.098) + 16.0);
int u = (int)( -((double)r * 0.148) - ((double)g * 0.291) + ((double)b * 0.439) + 128.0 ); int u = (int)(-((double)r * 0.148) - ((double)g * 0.291) + ((double)b * 0.439) + 128.0);
int v = (int)( +((double)r * 0.439) - ((double)g * 0.368) - ((double)b * 0.071) + 128.0 ); int v = (int)(+((double)r * 0.439) - ((double)g * 0.368) - ((double)b * 0.071) + 128.0);
xv->ytable[i] = y < 0 ? 0 : y > 255 ? 255 : y; xv->ytable[i] = y < 0 ? 0 : y > 255 ? 255 : y;
xv->utable[i] = u < 0 ? 0 : u > 255 ? 255 : u; xv->utable[i] = u < 0 ? 0 : u > 255 ? 255 : u;
@ -182,6 +182,7 @@ static void xv_init_font(xv_t *xv, const char *font_path, unsigned font_size)
#endif #endif
} }
// We render @ 2x scale to combat chroma downsampling. Also makes fonts more bearable :)
static void render16_yuy2(xv_t *xv, const void *input_, unsigned width, unsigned height, unsigned pitch) static void render16_yuy2(xv_t *xv, const void *input_, unsigned width, unsigned height, unsigned pitch)
{ {
const uint16_t *input = input_; const uint16_t *input = input_;
@ -189,24 +190,24 @@ static void render16_yuy2(xv_t *xv, const void *input_, unsigned width, unsigned
for (unsigned y = 0; y < height; y++) for (unsigned y = 0; y < height; y++)
{ {
for (unsigned x = 0; x < width >> 1; x++) for (unsigned x = 0; x < width; x++)
{ {
uint16_t p0 = *input++; uint16_t p = *input++;
uint16_t p1 = *input++;
uint8_t y0 = xv->ytable[p0]; uint8_t y0 = xv->ytable[p];
uint8_t y1 = xv->ytable[p1]; uint8_t u = xv->utable[p];
uint8_t u = (uint8_t)(((unsigned)xv->utable[p0] + (unsigned)xv->utable[p1]) >> 1); uint8_t v = xv->vtable[p];
uint8_t v = (uint8_t)(((unsigned)xv->vtable[p0] + (unsigned)xv->vtable[p1]) >> 1);
*output++ = y0; unsigned img_width = xv->width << 1;
*output++ = u; output[0] = output[img_width] = y0;
*output++ = y1; output[1] = output[img_width + 1] = u;
*output++ = v; output[2] = output[img_width + 2] = y0;
output[3] = output[img_width + 3] = v;
output += 4;
} }
input += (pitch >> 1) - width; input += (pitch >> 1) - width;
output += (xv->width - width) << 1; output += (xv->width - width) << 2;
} }
} }
@ -217,24 +218,24 @@ static void render16_uyvy(xv_t *xv, const void *input_, unsigned width, unsigned
for (unsigned y = 0; y < height; y++) for (unsigned y = 0; y < height; y++)
{ {
for (unsigned x = 0; x < width >> 1; x++) for (unsigned x = 0; x < width; x++)
{ {
uint16_t p0 = *input++; uint16_t p = *input++;
uint16_t p1 = *input++;
uint8_t y0 = xv->ytable[p0]; uint8_t y0 = xv->ytable[p];
uint8_t y1 = xv->ytable[p1]; uint8_t u = xv->utable[p];
uint8_t u = (uint8_t)(((unsigned)xv->utable[p0] + (unsigned)xv->utable[p1]) >> 1); uint8_t v = xv->vtable[p];
uint8_t v = (uint8_t)(((unsigned)xv->vtable[p0] + (unsigned)xv->vtable[p1]) >> 1);
*output++ = u; unsigned img_width = xv->width << 1;
*output++ = y0; output[0] = output[img_width] = u;
*output++ = v; output[1] = output[img_width + 1] = y0;
*output++ = y1; output[2] = output[img_width + 2] = v;
output[3] = output[img_width + 3] = y0;
output += 4;
} }
input += (pitch >> 1) - width; input += (pitch >> 1) - width;
output += (xv->width - width) << 1; output += (xv->width - width) << 2;
} }
} }
@ -243,28 +244,27 @@ static void render32_yuy2(xv_t *xv, const void *input_, unsigned width, unsigned
const uint32_t *input = input_; const uint32_t *input = input_;
uint8_t *output = (uint8_t*)xv->image->data; uint8_t *output = (uint8_t*)xv->image->data;
for (unsigned y = 0; y < height; y++) for (unsigned y = 0; y < height; y++)
{ {
for (unsigned x = 0; x < width >> 1; x++) for (unsigned x = 0; x < width; x++)
{ {
uint32_t p0 = *input++; uint32_t p = *input++;
uint32_t p1 = *input++; p = ((p >> 17) & 0x7c00) | ((p >> 14) & 0x03e0) | ((p >> 11) & 0x1f); // RGBA -> RGB15
p0 = ((p0 >> 9) & 0x7c00) | ((p0 >> 6) & 0x03e0) | ((p0 >> 3) & 0x1f); // RGBA -> RGB15
p1 = ((p1 >> 9) & 0x7c00) | ((p1 >> 6) & 0x03e0) | ((p1 >> 3) & 0x1f);
uint8_t y0 = xv->ytable[p0]; uint8_t y0 = xv->ytable[p];
uint8_t y1 = xv->ytable[p1]; uint8_t u = xv->utable[p];
uint8_t u = (uint8_t)(((unsigned)xv->utable[p0] + (unsigned)xv->utable[p1]) >> 1); uint8_t v = xv->vtable[p];
uint8_t v = (uint8_t)(((unsigned)xv->vtable[p0] + (unsigned)xv->vtable[p1]) >> 1);
*output++ = y0; unsigned img_width = xv->width << 1;
*output++ = u; output[0] = output[img_width] = y0;
*output++ = y1; output[1] = output[img_width + 1] = u;
*output++ = v; output[2] = output[img_width + 2] = y0;
output[3] = output[img_width + 3] = v;
output += 4;
} }
input += (pitch >> 2) - width; input += (pitch >> 2) - width;
output += (xv->width - width) << 1; output += (xv->width - width) << 2;
} }
} }
@ -275,26 +275,25 @@ static void render32_uyvy(xv_t *xv, const void *input_, unsigned width, unsigned
for (unsigned y = 0; y < height; y++) for (unsigned y = 0; y < height; y++)
{ {
for (unsigned x = 0; x < width >> 1; x++) for (unsigned x = 0; x < width; x++)
{ {
uint32_t p0 = *input++; uint32_t p = *input++;
uint32_t p1 = *input++; p = ((p >> 17) & 0x7c00) | ((p >> 14) & 0x03e0) | ((p >> 11) & 0x1f); // RGBA -> RGB15
p0 = ((p0 >> 9) & 0x7c00) | ((p0 >> 6) & 0x03e0) | ((p0 >> 3) & 0x1f);
p1 = ((p1 >> 9) & 0x7c00) | ((p1 >> 6) & 0x03e0) | ((p1 >> 3) & 0x1f);
uint8_t y0 = xv->ytable[p0]; uint8_t y0 = xv->ytable[p];
uint8_t y1 = xv->ytable[p1]; uint8_t u = xv->utable[p];
uint8_t u = (uint8_t)(((unsigned)xv->utable[p0] + (unsigned)xv->utable[p1]) >> 1); uint8_t v = xv->vtable[p];
uint8_t v = (uint8_t)(((unsigned)xv->vtable[p0] + (unsigned)xv->vtable[p1]) >> 1);
*output++ = u; unsigned img_width = xv->width << 1;
*output++ = y0; output[0] = output[img_width] = u;
*output++ = v; output[1] = output[img_width + 1] = y0;
*output++ = y1; output[2] = output[img_width + 2] = v;
output[3] = output[img_width + 3] = y0;
output += 4;
} }
input += (pitch >> 2) - width; input += (pitch >> 2) - width;
output += xv->width - width; output += (xv->width - width) << 2;
} }
} }
@ -430,8 +429,8 @@ static void* xv_init(video_info_t *video, const input_driver_t **input, void **i
goto error; goto error;
} }
xv->width = 256; xv->width = 512;
xv->height = 256; xv->height = 512;
xv->image = XvShmCreateImage(xv->display, xv->port, xv->fourcc, NULL, xv->width, xv->height, &xv->shminfo); xv->image = XvShmCreateImage(xv->display, xv->port, xv->fourcc, NULL, xv->width, xv->height, &xv->shminfo);
if (!xv->image) if (!xv->image)
@ -488,10 +487,11 @@ error:
static bool check_resize(xv_t *xv, unsigned width, unsigned height) static bool check_resize(xv_t *xv, unsigned width, unsigned height)
{ {
if (xv->width != width || xv->height != height) // We render @ 2x scale to combat chroma downsampling.
if (xv->width != (width << 1) || xv->height != (height << 1))
{ {
xv->width = width; xv->width = width << 1;
xv->height = height; xv->height = height << 1;
XShmDetach(xv->display, &xv->shminfo); XShmDetach(xv->display, &xv->shminfo);
shmdt(xv->shminfo.shmaddr); shmdt(xv->shminfo.shmaddr);
@ -581,28 +581,29 @@ static void xv_render_msg(xv_t *xv, const char *msg, unsigned width, unsigned he
{ {
int base_x = _base_x + head->off_x; int base_x = _base_x + head->off_x;
int base_y = _base_y - head->off_y; int base_y = _base_y - head->off_y;
for (int y = 0; y < head->height && (base_y + y) < height; y++) if (base_y >= 0)
{ {
if (base_y + y < 0) for (int y = 0; y < head->height && (base_y + y) < height; y++)
continue;
const uint8_t *a = head->output + head->pitch * y;
uint8_t *out = (uint8_t*)xv->image->data + (base_y - head->height + y) * (width << 1) + (base_x << 1);
for (int x = 0; x < (head->width << 1) && (base_x + x) < width; x += 2)
{ {
if (base_x + x < 0) if (base_x < 0)
continue; continue;
// Blend luma const uint8_t *a = head->output + head->pitch * y;
uint8_t blend = a[x >> 1]; uint8_t *out = (uint8_t*)xv->image->data + (base_y - head->height + y) * (width << 1) + (base_x << 1);
unsigned blended = blend + (((256 - blend) * (unsigned)out[x + luma_i]) >> 8);
blended = blended > 255 ? 255 : blended;
out[x + luma_i] = blended;
// Blend chroma for (int x = 0; x < (head->width << 1) && (base_x + x) < width; x += 2)
blended = (128 * blend + ((256 - blend) * (unsigned)out[x + chroma_i])) >> 8; {
blended = blended > 255 ? 255 : blended; // Blend luma
out[x + chroma_i] = blended; uint8_t blend = a[x >> 1];
unsigned blended = blend + (((256 - blend) * (unsigned)out[x + luma_i]) >> 8);
blended = blended > 255 ? 255 : blended;
out[x + luma_i] = blended;
// Blend chroma
blended = (128 * blend + ((256 - blend) * (unsigned)out[x + chroma_i])) >> 8;
blended = blended > 255 ? 255 : blended;
out[x + chroma_i] = blended;
}
} }
} }
@ -630,12 +631,12 @@ static bool xv_frame(void *data, const void* frame, unsigned width, unsigned hei
xv->render_func(xv, frame, width, height, pitch); xv->render_func(xv, frame, width, height, pitch);
if (msg) if (msg)
xv_render_msg(xv, msg, width, height); xv_render_msg(xv, msg, xv->width, xv->height);
unsigned x, y, owidth, oheight; unsigned x, y, owidth, oheight;
calc_out_rect(xv->keep_aspect, &x, &y, &owidth, &oheight, target.width, target.height); calc_out_rect(xv->keep_aspect, &x, &y, &owidth, &oheight, target.width, target.height);
XvShmPutImage(xv->display, xv->port, xv->window, xv->gc, xv->image, XvShmPutImage(xv->display, xv->port, xv->window, xv->gc, xv->image,
0, 0, width, height, 0, 0, width << 1, height << 1,
x, y, owidth, oheight, x, y, owidth, oheight,
true); true);
XSync(xv->display, False); XSync(xv->display, False);