int lumFilterSize, const int16_t *chrFilter,
const int16_t **chrUSrc, const int16_t **chrVSrc,
int chrFilterSize, const int16_t **alpSrc,
- uint16_t *dest, uint16_t *uDest, uint16_t *vDest,
- uint16_t *aDest, int dstW, int chrDstW,
+ uint16_t *dest[4], int dstW, int chrDstW,
int big_endian, int output_bits)
{
//FIXME Optimize (just quickly written not optimized..)
int i;
int shift = 11 + 16 - output_bits;
+ uint16_t *yDest = dest[0], *uDest = dest[1], *vDest = dest[2],
+ *aDest = CONFIG_SWSCALE_ALPHA ? dest[3] : NULL;
#define output_pixel(pos, val) \
if (big_endian) { \
for (j = 0; j < lumFilterSize; j++)
val += lumSrc[j][i] * lumFilter[j];
- output_pixel(&dest[i], val);
+ output_pixel(&yDest[i], val);
}
if (uDest) {
const int16_t *chrFilter, const int16_t **chrUSrc, \
const int16_t **chrVSrc, \
int chrFilterSize, const int16_t **alpSrc, \
- uint8_t *_dest, uint8_t *_uDest, uint8_t *_vDest, \
- uint8_t *_aDest, int dstW, int chrDstW) \
+ uint8_t *_dest[4], int dstW, int chrDstW) \
{ \
- uint16_t *dest = (uint16_t *) _dest, *uDest = (uint16_t *) _uDest, \
- *vDest = (uint16_t *) _vDest, *aDest = (uint16_t *) _aDest; \
yuv2yuvX16_c_template(lumFilter, lumSrc, lumFilterSize, \
chrFilter, chrUSrc, chrVSrc, chrFilterSize, \
- alpSrc, \
- dest, uDest, vDest, aDest, \
+ alpSrc, (uint16_t **) _dest, \
dstW, chrDstW, is_be, bits); \
}
yuv2NBPS( 9, BE, 1);
const int16_t *chrFilter, const int16_t **chrUSrc,
const int16_t **chrVSrc,
int chrFilterSize, const int16_t **alpSrc,
- uint8_t *dest, uint8_t *uDest, uint8_t *vDest,
- uint8_t *aDest, int dstW, int chrDstW,
+ uint8_t *dest[4], int dstW, int chrDstW,
const uint8_t *lumDither, const uint8_t *chrDither)
{
- //FIXME Optimize (just quickly written not optimized..)
+ uint8_t *yDest = dest[0], *uDest = dest[1], *vDest = dest[2],
+ *aDest = CONFIG_SWSCALE_ALPHA ? dest[3] : NULL;
int i;
+
+ //FIXME Optimize (just quickly written not optimized..)
for (i=0; i<dstW; i++) {
int val = lumDither[i&7] << 12;
int j;
for (j=0; j<lumFilterSize; j++)
val += lumSrc[j][i] * lumFilter[j];
- dest[i]= av_clip_uint8(val>>19);
+ yDest[i]= av_clip_uint8(val>>19);
}
if (uDest)
}
static void yuv2yuv1_c(SwsContext *c, const int16_t *lumSrc,
- const int16_t *chrUSrc, const int16_t *chrVSrc,
- const int16_t *alpSrc,
- uint8_t *dest, uint8_t *uDest, uint8_t *vDest,
- uint8_t *aDest, int dstW, int chrDstW, const uint8_t *lumDither, const uint8_t *chrDither)
+ const int16_t *chrUSrc, const int16_t *chrVSrc,
+ const int16_t *alpSrc,
+ uint8_t *dest[4], int dstW, int chrDstW,
+ const uint8_t *lumDither, const uint8_t *chrDither)
{
+ uint8_t *yDest = dest[0], *uDest = dest[1], *vDest = dest[2],
+ *aDest = CONFIG_SWSCALE_ALPHA ? dest[3] : NULL;
int i;
for (i=0; i<dstW; i++) {
int val= (lumSrc[i]+lumDither[i&7])>>7;
- dest[i]= av_clip_uint8(val);
+ yDest[i]= av_clip_uint8(val);
}
if (uDest)
const int16_t **lumSrc, int lumFilterSize,
const int16_t *chrFilter, const int16_t **chrUSrc,
const int16_t **chrVSrc, int chrFilterSize,
- const int16_t **alpSrc, uint8_t *dest, uint8_t *uDest,
- uint8_t *vDest, uint8_t *aDest,
+ const int16_t **alpSrc, uint8_t *dest[4],
int dstW, int chrDstW,
const uint8_t *lumDither, const uint8_t *chrDither)
{
+ uint8_t *yDest = dest[0], *uDest = dest[1];
enum PixelFormat dstFormat = c->dstFormat;
//FIXME Optimize (just quickly written not optimized..)
for (j=0; j<lumFilterSize; j++)
val += lumSrc[j][i] * lumFilter[j];
- dest[i]= av_clip_uint8(val>>19);
+ yDest[i]= av_clip_uint8(val>>19);
}
if (!uDest)
}
static av_always_inline void
-yuv2gray16_2_c_template(SwsContext *c, const uint16_t *buf0,
- const uint16_t *buf1, const uint16_t *ubuf0,
- const uint16_t *ubuf1, const uint16_t *vbuf0,
- const uint16_t *vbuf1, const uint16_t *abuf0,
- const uint16_t *abuf1, uint8_t *dest, int dstW,
+yuv2gray16_2_c_template(SwsContext *c, const int16_t *buf[2],
+ const int16_t *ubuf[2], const int16_t *vbuf[2],
+ const int16_t *abuf[2], uint8_t *dest, int dstW,
int yalpha, int uvalpha, int y,
enum PixelFormat target)
{
- int yalpha1 = 4095 - yalpha; \
+ int yalpha1 = 4095 - yalpha;
int i;
+ const int16_t *buf0 = buf[0], *buf1 = buf[1];
for (i = 0; i < (dstW >> 1); i++) {
const int i2 = 2 * i;
}
static av_always_inline void
-yuv2gray16_1_c_template(SwsContext *c, const uint16_t *buf0,
- const uint16_t *ubuf0, const uint16_t *ubuf1,
- const uint16_t *vbuf0, const uint16_t *vbuf1,
- const uint16_t *abuf0, uint8_t *dest, int dstW,
- int uvalpha, enum PixelFormat dstFormat,
- int flags, int y, enum PixelFormat target)
+yuv2gray16_1_c_template(SwsContext *c, const int16_t *buf0,
+ const int16_t *ubuf[2], const int16_t *vbuf[2],
+ const int16_t *abuf0, uint8_t *dest, int dstW,
+ int uvalpha, int y, enum PixelFormat target)
{
int i;
alpSrc, dest, dstW, y, fmt); \
} \
\
-static void name ## ext ## _2_c(SwsContext *c, const uint16_t *buf0, \
- const uint16_t *buf1, const uint16_t *ubuf0, \
- const uint16_t *ubuf1, const uint16_t *vbuf0, \
- const uint16_t *vbuf1, const uint16_t *abuf0, \
- const uint16_t *abuf1, uint8_t *dest, int dstW, \
+static void name ## ext ## _2_c(SwsContext *c, const int16_t *buf[2], \
+ const int16_t *ubuf[2], const int16_t *vbuf[2], \
+ const int16_t *abuf[2], uint8_t *dest, int dstW, \
int yalpha, int uvalpha, int y) \
{ \
- name ## base ## _2_c_template(c, buf0, buf1, ubuf0, ubuf1, \
- vbuf0, vbuf1, abuf0, abuf1, \
+ name ## base ## _2_c_template(c, buf, ubuf, vbuf, abuf, \
dest, dstW, yalpha, uvalpha, y, fmt); \
} \
\
-static void name ## ext ## _1_c(SwsContext *c, const uint16_t *buf0, \
- const uint16_t *ubuf0, const uint16_t *ubuf1, \
- const uint16_t *vbuf0, const uint16_t *vbuf1, \
- const uint16_t *abuf0, uint8_t *dest, int dstW, \
- int uvalpha, enum PixelFormat dstFormat, \
- int flags, int y) \
+static void name ## ext ## _1_c(SwsContext *c, const int16_t *buf0, \
+ const int16_t *ubuf[2], const int16_t *vbuf[2], \
+ const int16_t *abuf0, uint8_t *dest, int dstW, \
+ int uvalpha, int y) \
{ \
- name ## base ## _1_c_template(c, buf0, ubuf0, ubuf1, vbuf0, \
- vbuf1, abuf0, dest, dstW, uvalpha, \
- dstFormat, flags, y, fmt); \
+ name ## base ## _1_c_template(c, buf0, ubuf, vbuf, abuf0, dest, \
+ dstW, uvalpha, y, fmt); \
}
YUV2PACKEDWRAPPER(yuv2gray16,, LE, PIX_FMT_GRAY16LE);
}
static av_always_inline void
-yuv2mono_2_c_template(SwsContext *c, const uint16_t *buf0,
- const uint16_t *buf1, const uint16_t *ubuf0,
- const uint16_t *ubuf1, const uint16_t *vbuf0,
- const uint16_t *vbuf1, const uint16_t *abuf0,
- const uint16_t *abuf1, uint8_t *dest, int dstW,
+yuv2mono_2_c_template(SwsContext *c, const int16_t *buf[2],
+ const int16_t *ubuf[2], const int16_t *vbuf[2],
+ const int16_t *abuf[2], uint8_t *dest, int dstW,
int yalpha, int uvalpha, int y,
enum PixelFormat target)
{
+ const int16_t *buf0 = buf[0], *buf1 = buf[1];
const uint8_t * const d128 = dither_8x8_220[y & 7];
uint8_t *g = c->table_gU[128] + c->table_gV[128];
int yalpha1 = 4095 - yalpha;
}
static av_always_inline void
-yuv2mono_1_c_template(SwsContext *c, const uint16_t *buf0,
- const uint16_t *ubuf0, const uint16_t *ubuf1,
- const uint16_t *vbuf0, const uint16_t *vbuf1,
- const uint16_t *abuf0, uint8_t *dest, int dstW,
- int uvalpha, enum PixelFormat dstFormat,
- int flags, int y, enum PixelFormat target)
+yuv2mono_1_c_template(SwsContext *c, const int16_t *buf0,
+ const int16_t *ubuf[2], const int16_t *vbuf[2],
+ const int16_t *abuf0, uint8_t *dest, int dstW,
+ int uvalpha, int y, enum PixelFormat target)
{
const uint8_t * const d128 = dither_8x8_220[y & 7];
uint8_t *g = c->table_gU[128] + c->table_gV[128];
}
static av_always_inline void
-yuv2422_2_c_template(SwsContext *c, const uint16_t *buf0,
- const uint16_t *buf1, const uint16_t *ubuf0,
- const uint16_t *ubuf1, const uint16_t *vbuf0,
- const uint16_t *vbuf1, const uint16_t *abuf0,
- const uint16_t *abuf1, uint8_t *dest, int dstW,
+yuv2422_2_c_template(SwsContext *c, const int16_t *buf[2],
+ const int16_t *ubuf[2], const int16_t *vbuf[2],
+ const int16_t *abuf[2], uint8_t *dest, int dstW,
int yalpha, int uvalpha, int y,
enum PixelFormat target)
{
+ const int16_t *buf0 = buf[0], *buf1 = buf[1],
+ *ubuf0 = ubuf[0], *ubuf1 = ubuf[1],
+ *vbuf0 = vbuf[0], *vbuf1 = vbuf[1];
int yalpha1 = 4095 - yalpha;
int uvalpha1 = 4095 - uvalpha;
int i;
}
static av_always_inline void
-yuv2422_1_c_template(SwsContext *c, const uint16_t *buf0,
- const uint16_t *ubuf0, const uint16_t *ubuf1,
- const uint16_t *vbuf0, const uint16_t *vbuf1,
- const uint16_t *abuf0, uint8_t *dest, int dstW,
- int uvalpha, enum PixelFormat dstFormat,
- int flags, int y, enum PixelFormat target)
+yuv2422_1_c_template(SwsContext *c, const int16_t *buf0,
+ const int16_t *ubuf[2], const int16_t *vbuf[2],
+ const int16_t *abuf0, uint8_t *dest, int dstW,
+ int uvalpha, int y, enum PixelFormat target)
{
+ const int16_t *ubuf0 = ubuf[0], *ubuf1 = ubuf[1],
+ *vbuf0 = vbuf[0], *vbuf1 = vbuf[1];
int i;
if (uvalpha < 2048) {
}
static av_always_inline void
-yuv2rgb48_2_c_template(SwsContext *c, const uint16_t *buf0,
- const uint16_t *buf1, const uint16_t *ubuf0,
- const uint16_t *ubuf1, const uint16_t *vbuf0,
- const uint16_t *vbuf1, const uint16_t *abuf0,
- const uint16_t *abuf1, uint8_t *dest, int dstW,
+yuv2rgb48_2_c_template(SwsContext *c, const int16_t *buf[2],
+ const int16_t *ubuf[2], const int16_t *vbuf[2],
+ const int16_t *abuf[2], uint8_t *dest, int dstW,
int yalpha, int uvalpha, int y,
enum PixelFormat target)
{
+ const int16_t *buf0 = buf[0], *buf1 = buf[1],
+ *ubuf0 = ubuf[0], *ubuf1 = ubuf[1],
+ *vbuf0 = vbuf[0], *vbuf1 = vbuf[1];
int yalpha1 = 4095 - yalpha;
int uvalpha1 = 4095 - uvalpha;
int i;
}
static av_always_inline void
-yuv2rgb48_1_c_template(SwsContext *c, const uint16_t *buf0,
- const uint16_t *ubuf0, const uint16_t *ubuf1,
- const uint16_t *vbuf0, const uint16_t *vbuf1,
- const uint16_t *abuf0, uint8_t *dest, int dstW,
- int uvalpha, enum PixelFormat dstFormat,
- int flags, int y, enum PixelFormat target)
+yuv2rgb48_1_c_template(SwsContext *c, const int16_t *buf0,
+ const int16_t *ubuf[2], const int16_t *vbuf[2],
+ const int16_t *abuf0, uint8_t *dest, int dstW,
+ int uvalpha, int y, enum PixelFormat target)
{
+ const int16_t *ubuf0 = ubuf[0], *ubuf1 = ubuf[1],
+ *vbuf0 = vbuf[0], *vbuf1 = vbuf[1];
int i;
if (uvalpha < 2048) {
YUV2PACKEDWRAPPER(yuv2, rgb48, bgr48be, PIX_FMT_BGR48BE);
//YUV2PACKEDWRAPPER(yuv2, rgb48, bgr48le, PIX_FMT_BGR48LE);
-#define YSCALE_YUV_2_RGBX_C(type,alpha) \
- for (i=0; i<(dstW>>1); i++) {\
- int j;\
- int Y1 = 1<<18;\
- int Y2 = 1<<18;\
- int U = 1<<18;\
- int V = 1<<18;\
- int av_unused A1, A2;\
- type av_unused *r, *b, *g;\
- const int i2= 2*i;\
- \
- for (j=0; j<lumFilterSize; j++) {\
- Y1 += lumSrc[j][i2] * lumFilter[j];\
- Y2 += lumSrc[j][i2+1] * lumFilter[j];\
- }\
- for (j=0; j<chrFilterSize; j++) {\
- U += chrUSrc[j][i] * chrFilter[j];\
- V += chrVSrc[j][i] * chrFilter[j];\
- }\
- Y1>>=19;\
- Y2>>=19;\
- U >>=19;\
- V >>=19;\
- if ((Y1|Y2|U|V)&0x100) {\
- Y1 = av_clip_uint8(Y1); \
- Y2 = av_clip_uint8(Y2); \
- U = av_clip_uint8(U); \
- V = av_clip_uint8(V); \
- }\
- if (alpha) {\
- A1 = 1<<18;\
- A2 = 1<<18;\
- for (j=0; j<lumFilterSize; j++) {\
- A1 += alpSrc[j][i2 ] * lumFilter[j];\
- A2 += alpSrc[j][i2+1] * lumFilter[j];\
- }\
- A1>>=19;\
- A2>>=19;\
- if ((A1|A2)&0x100) {\
- A1 = av_clip_uint8(A1); \
- A2 = av_clip_uint8(A2); \
- }\
- }\
- /* FIXME fix tables so that clipping is not needed and then use _NOCLIP*/\
- r = (type *)c->table_rV[V]; \
- g = (type *)(c->table_gU[U] + c->table_gV[V]); \
- b = (type *)c->table_bU[U];
-
-#define YSCALE_YUV_2_RGBX_FULL_C(rnd,alpha) \
- for (i=0; i<dstW; i++) {\
- int j;\
- int Y = 1<<9;\
- int U = (1<<9)-(128<<19);\
- int V = (1<<9)-(128<<19);\
- int av_unused A;\
- int R,G,B;\
- \
- for (j=0; j<lumFilterSize; j++) {\
- Y += lumSrc[j][i ] * lumFilter[j];\
- }\
- for (j=0; j<chrFilterSize; j++) {\
- U += chrUSrc[j][i] * chrFilter[j];\
- V += chrVSrc[j][i] * chrFilter[j];\
- }\
- Y >>=10;\
- U >>=10;\
- V >>=10;\
- if (alpha) {\
- A = rnd>>3;\
- for (j=0; j<lumFilterSize; j++)\
- A += alpSrc[j][i ] * lumFilter[j];\
- A >>=19;\
- if (A&0x100)\
- A = av_clip_uint8(A);\
- }\
- Y-= c->yuv2rgb_y_offset;\
- Y*= c->yuv2rgb_y_coeff;\
- Y+= rnd;\
- R= Y + V*c->yuv2rgb_v2r_coeff;\
- G= Y + V*c->yuv2rgb_v2g_coeff + U*c->yuv2rgb_u2g_coeff;\
- B= Y + U*c->yuv2rgb_u2b_coeff;\
- if ((R|G|B)&(0xC0000000)) {\
- R = av_clip_uintp2(R, 30); \
- G = av_clip_uintp2(G, 30); \
- B = av_clip_uintp2(B, 30); \
+static av_always_inline void
+yuv2rgb_write(uint8_t *_dest, int i, int Y1, int Y2,
+ int U, int V, int A1, int A2,
+ const void *_r, const void *_g, const void *_b, int y,
+ enum PixelFormat target, int hasAlpha)
+{
+ if (target == PIX_FMT_ARGB || target == PIX_FMT_RGBA ||
+ target == PIX_FMT_ABGR || target == PIX_FMT_BGRA) {
+ uint32_t *dest = (uint32_t *) _dest;
+ const uint32_t *r = (const uint32_t *) _r;
+ const uint32_t *g = (const uint32_t *) _g;
+ const uint32_t *b = (const uint32_t *) _b;
+
+#if CONFIG_SMALL
+ int sh = hasAlpha ? ((fmt == PIX_FMT_RGB32_1 || fmt == PIX_FMT_BGR32_1) ? 0 : 24) : 0;
+
+ dest[i * 2 + 0] = r[Y1] + g[Y1] + b[Y1] + (hasAlpha ? A1 << sh : 0);
+ dest[i * 2 + 1] = r[Y2] + g[Y2] + b[Y2] + (hasAlpha ? A2 << sh : 0);
+#else
+ if (hasAlpha) {
+ int sh = (target == PIX_FMT_RGB32_1 || target == PIX_FMT_BGR32_1) ? 0 : 24;
+
+ dest[i * 2 + 0] = r[Y1] + g[Y1] + b[Y1] + (A1 << sh);
+ dest[i * 2 + 1] = r[Y2] + g[Y2] + b[Y2] + (A2 << sh);
+ } else {
+ dest[i * 2 + 0] = r[Y1] + g[Y1] + b[Y1];
+ dest[i * 2 + 1] = r[Y2] + g[Y2] + b[Y2];
+ }
+#endif
+ } else if (target == PIX_FMT_RGB24 || target == PIX_FMT_BGR24) {
+ uint8_t *dest = (uint8_t *) _dest;
+ const uint8_t *r = (const uint8_t *) _r;
+ const uint8_t *g = (const uint8_t *) _g;
+ const uint8_t *b = (const uint8_t *) _b;
+
+#define r_b ((target == PIX_FMT_RGB24) ? r : b)
+#define b_r ((target == PIX_FMT_RGB24) ? b : r)
+ dest[i * 6 + 0] = r_b[Y1];
+ dest[i * 6 + 1] = g[Y1];
+ dest[i * 6 + 2] = b_r[Y1];
+ dest[i * 6 + 3] = r_b[Y2];
+ dest[i * 6 + 4] = g[Y2];
+ dest[i * 6 + 5] = b_r[Y2];
+#undef r_b
+#undef b_r
+ } else if (target == PIX_FMT_RGB565 || target == PIX_FMT_BGR565 ||
+ target == PIX_FMT_RGB555 || target == PIX_FMT_BGR555 ||
+ target == PIX_FMT_RGB444 || target == PIX_FMT_BGR444) {
+ uint16_t *dest = (uint16_t *) _dest;
+ const uint16_t *r = (const uint16_t *) _r;
+ const uint16_t *g = (const uint16_t *) _g;
+ const uint16_t *b = (const uint16_t *) _b;
+ int dr1, dg1, db1, dr2, dg2, db2;
+
+ if (target == PIX_FMT_RGB565 || target == PIX_FMT_BGR565) {
+ dr1 = dither_2x2_8[ y & 1 ][0];
+ dg1 = dither_2x2_4[ y & 1 ][0];
+ db1 = dither_2x2_8[(y & 1) ^ 1][0];
+ dr2 = dither_2x2_8[ y & 1 ][1];
+ dg2 = dither_2x2_4[ y & 1 ][1];
+ db2 = dither_2x2_8[(y & 1) ^ 1][1];
+ } else if (target == PIX_FMT_RGB555 || target == PIX_FMT_BGR555) {
+ dr1 = dither_2x2_8[ y & 1 ][0];
+ dg1 = dither_2x2_8[ y & 1 ][1];
+ db1 = dither_2x2_8[(y & 1) ^ 1][0];
+ dr2 = dither_2x2_8[ y & 1 ][1];
+ dg2 = dither_2x2_8[ y & 1 ][0];
+ db2 = dither_2x2_8[(y & 1) ^ 1][1];
+ } else {
+ dr1 = dither_4x4_16[ y & 3 ][0];
+ dg1 = dither_4x4_16[ y & 3 ][1];
+ db1 = dither_4x4_16[(y & 3) ^ 3][0];
+ dr2 = dither_4x4_16[ y & 3 ][1];
+ dg2 = dither_4x4_16[ y & 3 ][0];
+ db2 = dither_4x4_16[(y & 3) ^ 3][1];
+ }
+
+ dest[i * 2 + 0] = r[Y1 + dr1] + g[Y1 + dg1] + b[Y1 + db1];
+ dest[i * 2 + 1] = r[Y2 + dr2] + g[Y2 + dg2] + b[Y2 + db2];
+ } else /* 8/4-bit */ {
+ uint8_t *dest = (uint8_t *) _dest;
+ const uint8_t *r = (const uint8_t *) _r;
+ const uint8_t *g = (const uint8_t *) _g;
+ const uint8_t *b = (const uint8_t *) _b;
+ int dr1, dg1, db1, dr2, dg2, db2;
+
+ if (target == PIX_FMT_RGB8 || target == PIX_FMT_BGR8) {
+ const uint8_t * const d64 = dither_8x8_73[y & 7];
+ const uint8_t * const d32 = dither_8x8_32[y & 7];
+ dr1 = dg1 = d32[(i * 2 + 0) & 7];
+ db1 = d64[(i * 2 + 0) & 7];
+ dr2 = dg2 = d32[(i * 2 + 1) & 7];
+ db2 = d64[(i * 2 + 1) & 7];
+ } else {
+ const uint8_t * const d64 = dither_8x8_73 [y & 7];
+ const uint8_t * const d128 = dither_8x8_220[y & 7];
+ dr1 = db1 = d128[(i * 2 + 0) & 7];
+ dg1 = d64[(i * 2 + 0) & 7];
+ dr2 = db2 = d128[(i * 2 + 1) & 7];
+ dg2 = d64[(i * 2 + 1) & 7];
}
-#define YSCALE_YUV_2_RGB2_C(type,alpha) \
- for (i=0; i<(dstW>>1); i++) { \
- const int i2= 2*i; \
- int Y1= (buf0[i2 ]*yalpha1+buf1[i2 ]*yalpha)>>19; \
- int Y2= (buf0[i2+1]*yalpha1+buf1[i2+1]*yalpha)>>19; \
- int U= (ubuf0[i]*uvalpha1+ubuf1[i]*uvalpha)>>19; \
- int V= (vbuf0[i]*uvalpha1+vbuf1[i]*uvalpha)>>19; \
- type av_unused *r, *b, *g; \
- int av_unused A1, A2; \
- if (alpha) {\
- A1= (abuf0[i2 ]*yalpha1+abuf1[i2 ]*yalpha)>>19; \
- A2= (abuf0[i2+1]*yalpha1+abuf1[i2+1]*yalpha)>>19; \
- }\
- r = (type *)c->table_rV[V];\
- g = (type *)(c->table_gU[U] + c->table_gV[V]);\
- b = (type *)c->table_bU[U];
-
-#define YSCALE_YUV_2_RGB1_C(type,alpha) \
- for (i=0; i<(dstW>>1); i++) {\
- const int i2= 2*i;\
- int Y1= buf0[i2 ]>>7;\
- int Y2= buf0[i2+1]>>7;\
- int U= (ubuf1[i])>>7;\
- int V= (vbuf1[i])>>7;\
- type av_unused *r, *b, *g;\
- int av_unused A1, A2;\
- if (alpha) {\
- A1= abuf0[i2 ]>>7;\
- A2= abuf0[i2+1]>>7;\
- }\
- r = (type *)c->table_rV[V];\
- g = (type *)(c->table_gU[U] + c->table_gV[V]);\
- b = (type *)c->table_bU[U];
-
-#define YSCALE_YUV_2_RGB1B_C(type,alpha) \
- for (i=0; i<(dstW>>1); i++) {\
- const int i2= 2*i;\
- int Y1= buf0[i2 ]>>7;\
- int Y2= buf0[i2+1]>>7;\
- int U= (ubuf0[i] + ubuf1[i])>>8;\
- int V= (vbuf0[i] + vbuf1[i])>>8;\
- type av_unused *r, *b, *g;\
- int av_unused A1, A2;\
- if (alpha) {\
- A1= abuf0[i2 ]>>7;\
- A2= abuf0[i2+1]>>7;\
- }\
- r = (type *)c->table_rV[V];\
- g = (type *)(c->table_gU[U] + c->table_gV[V]);\
- b = (type *)c->table_bU[U];
-
-#define YSCALE_YUV_2_ANYRGB_C(func)\
- switch(c->dstFormat) {\
- case PIX_FMT_RGBA:\
- case PIX_FMT_BGRA:\
- if (CONFIG_SMALL) {\
- int needAlpha = CONFIG_SWSCALE_ALPHA && c->alpPixBuf;\
- func(uint32_t,needAlpha)\
- ((uint32_t*)dest)[i2+0]= r[Y1] + g[Y1] + b[Y1] + (needAlpha ? (A1<<24) : 0);\
- ((uint32_t*)dest)[i2+1]= r[Y2] + g[Y2] + b[Y2] + (needAlpha ? (A2<<24) : 0);\
- }\
- } else {\
- if (CONFIG_SWSCALE_ALPHA && c->alpPixBuf) {\
- func(uint32_t,1)\
- ((uint32_t*)dest)[i2+0]= r[Y1] + g[Y1] + b[Y1] + (A1<<24);\
- ((uint32_t*)dest)[i2+1]= r[Y2] + g[Y2] + b[Y2] + (A2<<24);\
- }\
- } else {\
- func(uint32_t,0)\
- ((uint32_t*)dest)[i2+0]= r[Y1] + g[Y1] + b[Y1];\
- ((uint32_t*)dest)[i2+1]= r[Y2] + g[Y2] + b[Y2];\
- }\
- }\
- }\
- break;\
- case PIX_FMT_ARGB:\
- case PIX_FMT_ABGR:\
- if (CONFIG_SMALL) {\
- int needAlpha = CONFIG_SWSCALE_ALPHA && c->alpPixBuf;\
- func(uint32_t,needAlpha)\
- ((uint32_t*)dest)[i2+0]= r[Y1] + g[Y1] + b[Y1] + (needAlpha ? A1 : 0);\
- ((uint32_t*)dest)[i2+1]= r[Y2] + g[Y2] + b[Y2] + (needAlpha ? A2 : 0);\
- }\
- } else {\
- if (CONFIG_SWSCALE_ALPHA && c->alpPixBuf) {\
- func(uint32_t,1)\
- ((uint32_t*)dest)[i2+0]= r[Y1] + g[Y1] + b[Y1] + A1;\
- ((uint32_t*)dest)[i2+1]= r[Y2] + g[Y2] + b[Y2] + A2;\
- }\
- } else {\
- func(uint32_t,0)\
- ((uint32_t*)dest)[i2+0]= r[Y1] + g[Y1] + b[Y1];\
- ((uint32_t*)dest)[i2+1]= r[Y2] + g[Y2] + b[Y2];\
- }\
- }\
- } \
- break;\
- case PIX_FMT_RGB24:\
- func(uint8_t,0)\
- ((uint8_t*)dest)[0]= r[Y1];\
- ((uint8_t*)dest)[1]= g[Y1];\
- ((uint8_t*)dest)[2]= b[Y1];\
- ((uint8_t*)dest)[3]= r[Y2];\
- ((uint8_t*)dest)[4]= g[Y2];\
- ((uint8_t*)dest)[5]= b[Y2];\
- dest+=6;\
- }\
- break;\
- case PIX_FMT_BGR24:\
- func(uint8_t,0)\
- ((uint8_t*)dest)[0]= b[Y1];\
- ((uint8_t*)dest)[1]= g[Y1];\
- ((uint8_t*)dest)[2]= r[Y1];\
- ((uint8_t*)dest)[3]= b[Y2];\
- ((uint8_t*)dest)[4]= g[Y2];\
- ((uint8_t*)dest)[5]= r[Y2];\
- dest+=6;\
- }\
- break;\
- case PIX_FMT_RGB565:\
- case PIX_FMT_BGR565:\
- {\
- const int dr1= dither_2x2_8[y&1 ][0];\
- const int dg1= dither_2x2_4[y&1 ][0];\
- const int db1= dither_2x2_8[(y&1)^1][0];\
- const int dr2= dither_2x2_8[y&1 ][1];\
- const int dg2= dither_2x2_4[y&1 ][1];\
- const int db2= dither_2x2_8[(y&1)^1][1];\
- func(uint16_t,0)\
- ((uint16_t*)dest)[i2+0]= r[Y1+dr1] + g[Y1+dg1] + b[Y1+db1];\
- ((uint16_t*)dest)[i2+1]= r[Y2+dr2] + g[Y2+dg2] + b[Y2+db2];\
- }\
- }\
- break;\
- case PIX_FMT_RGB555:\
- case PIX_FMT_BGR555:\
- {\
- const int dr1= dither_2x2_8[y&1 ][0];\
- const int dg1= dither_2x2_8[y&1 ][1];\
- const int db1= dither_2x2_8[(y&1)^1][0];\
- const int dr2= dither_2x2_8[y&1 ][1];\
- const int dg2= dither_2x2_8[y&1 ][0];\
- const int db2= dither_2x2_8[(y&1)^1][1];\
- func(uint16_t,0)\
- ((uint16_t*)dest)[i2+0]= r[Y1+dr1] + g[Y1+dg1] + b[Y1+db1];\
- ((uint16_t*)dest)[i2+1]= r[Y2+dr2] + g[Y2+dg2] + b[Y2+db2];\
- }\
- }\
- break;\
- case PIX_FMT_RGB444:\
- case PIX_FMT_BGR444:\
- {\
- const int dr1= dither_4x4_16[y&3 ][0];\
- const int dg1= dither_4x4_16[y&3 ][1];\
- const int db1= dither_4x4_16[(y&3)^3][0];\
- const int dr2= dither_4x4_16[y&3 ][1];\
- const int dg2= dither_4x4_16[y&3 ][0];\
- const int db2= dither_4x4_16[(y&3)^3][1];\
- func(uint16_t,0)\
- ((uint16_t*)dest)[i2+0]= r[Y1+dr1] + g[Y1+dg1] + b[Y1+db1];\
- ((uint16_t*)dest)[i2+1]= r[Y2+dr2] + g[Y2+dg2] + b[Y2+db2];\
- }\
- }\
- break;\
- case PIX_FMT_RGB8:\
- case PIX_FMT_BGR8:\
- {\
- const uint8_t * const d64= dither_8x8_73[y&7];\
- const uint8_t * const d32= dither_8x8_32[y&7];\
- func(uint8_t,0)\
- ((uint8_t*)dest)[i2+0]= r[Y1+d32[(i2+0)&7]] + g[Y1+d32[(i2+0)&7]] + b[Y1+d64[(i2+0)&7]];\
- ((uint8_t*)dest)[i2+1]= r[Y2+d32[(i2+1)&7]] + g[Y2+d32[(i2+1)&7]] + b[Y2+d64[(i2+1)&7]];\
- }\
- }\
- break;\
- case PIX_FMT_RGB4:\
- case PIX_FMT_BGR4:\
- {\
- const uint8_t * const d64= dither_8x8_73 [y&7];\
- const uint8_t * const d128=dither_8x8_220[y&7];\
- func(uint8_t,0)\
- ((uint8_t*)dest)[i]= r[Y1+d128[(i2+0)&7]] + g[Y1+d64[(i2+0)&7]] + b[Y1+d128[(i2+0)&7]]\
- + ((r[Y2+d128[(i2+1)&7]] + g[Y2+d64[(i2+1)&7]] + b[Y2+d128[(i2+1)&7]])<<4);\
- }\
- }\
- break;\
- case PIX_FMT_RGB4_BYTE:\
- case PIX_FMT_BGR4_BYTE:\
- {\
- const uint8_t * const d64= dither_8x8_73 [y&7];\
- const uint8_t * const d128=dither_8x8_220[y&7];\
- func(uint8_t,0)\
- ((uint8_t*)dest)[i2+0]= r[Y1+d128[(i2+0)&7]] + g[Y1+d64[(i2+0)&7]] + b[Y1+d128[(i2+0)&7]];\
- ((uint8_t*)dest)[i2+1]= r[Y2+d128[(i2+1)&7]] + g[Y2+d64[(i2+1)&7]] + b[Y2+d128[(i2+1)&7]];\
- }\
- }\
- break;\
- }
-
-static void yuv2packedX_c(SwsContext *c, const int16_t *lumFilter,
- const int16_t **lumSrc, int lumFilterSize,
- const int16_t *chrFilter, const int16_t **chrUSrc,
- const int16_t **chrVSrc, int chrFilterSize,
- const int16_t **alpSrc, uint8_t *dest, int dstW, int y)
-{
- int i;
- YSCALE_YUV_2_ANYRGB_C(YSCALE_YUV_2_RGBX_C)
+ if (target == PIX_FMT_RGB4 || target == PIX_FMT_BGR4) {
+ dest[i] = r[Y1 + dr1] + g[Y1 + dg1] + b[Y1 + db1] +
+ ((r[Y2 + dr2] + g[Y2 + dg2] + b[Y2 + db2]) << 4);
+ } else {
+ dest[i * 2 + 0] = r[Y1 + dr1] + g[Y1 + dg1] + b[Y1 + db1];
+ dest[i * 2 + 1] = r[Y2 + dr2] + g[Y2 + dg2] + b[Y2 + db2];
+ }
+ }
}
-static void yuv2rgbX_c_full(SwsContext *c, const int16_t *lumFilter,
- const int16_t **lumSrc, int lumFilterSize,
- const int16_t *chrFilter, const int16_t **chrUSrc,
- const int16_t **chrVSrc, int chrFilterSize,
- const int16_t **alpSrc, uint8_t *dest, int dstW, int y)
+static av_always_inline void
+yuv2rgb_X_c_template(SwsContext *c, const int16_t *lumFilter,
+ const int16_t **lumSrc, int lumFilterSize,
+ const int16_t *chrFilter, const int16_t **chrUSrc,
+ const int16_t **chrVSrc, int chrFilterSize,
+ const int16_t **alpSrc, uint8_t *dest, int dstW,
+ int y, enum PixelFormat target, int hasAlpha)
{
int i;
- int step= c->dstFormatBpp/8;
- int aidx= 3;
-
- switch(c->dstFormat) {
- case PIX_FMT_ARGB:
- dest++;
- aidx= 0;
- case PIX_FMT_RGB24:
- aidx--;
- case PIX_FMT_RGBA:
- if (CONFIG_SMALL) {
- int needAlpha = CONFIG_SWSCALE_ALPHA && c->alpPixBuf;
- YSCALE_YUV_2_RGBX_FULL_C(1<<21, needAlpha)
- dest[aidx]= needAlpha ? A : 255;
- dest[0]= R>>22;
- dest[1]= G>>22;
- dest[2]= B>>22;
- dest+= step;
- }
- } else {
- if (CONFIG_SWSCALE_ALPHA && c->alpPixBuf) {
- YSCALE_YUV_2_RGBX_FULL_C(1<<21, 1)
- dest[aidx]= A;
- dest[0]= R>>22;
- dest[1]= G>>22;
- dest[2]= B>>22;
- dest+= step;
- }
- } else {
- YSCALE_YUV_2_RGBX_FULL_C(1<<21, 0)
- dest[aidx]= 255;
- dest[0]= R>>22;
- dest[1]= G>>22;
- dest[2]= B>>22;
- dest+= step;
- }
- }
+
+ for (i = 0; i < (dstW >> 1); i++) {
+ int j;
+ int Y1 = 1 << 18;
+ int Y2 = 1 << 18;
+ int U = 1 << 18;
+ int V = 1 << 18;
+ int av_unused A1, A2;
+ const void *r, *g, *b;
+
+ for (j = 0; j < lumFilterSize; j++) {
+ Y1 += lumSrc[j][i * 2] * lumFilter[j];
+ Y2 += lumSrc[j][i * 2 + 1] * lumFilter[j];
+ }
+ for (j = 0; j < chrFilterSize; j++) {
+ U += chrUSrc[j][i] * chrFilter[j];
+ V += chrVSrc[j][i] * chrFilter[j];
+ }
+ Y1 >>= 19;
+ Y2 >>= 19;
+ U >>= 19;
+ V >>= 19;
+ if ((Y1 | Y2 | U | V) & 0x100) {
+ Y1 = av_clip_uint8(Y1);
+ Y2 = av_clip_uint8(Y2);
+ U = av_clip_uint8(U);
+ V = av_clip_uint8(V);
}
- break;
- case PIX_FMT_ABGR:
- dest++;
- aidx= 0;
- case PIX_FMT_BGR24:
- aidx--;
- case PIX_FMT_BGRA:
- if (CONFIG_SMALL) {
- int needAlpha = CONFIG_SWSCALE_ALPHA && c->alpPixBuf;
- YSCALE_YUV_2_RGBX_FULL_C(1<<21, needAlpha)
- dest[aidx]= needAlpha ? A : 255;
- dest[0]= B>>22;
- dest[1]= G>>22;
- dest[2]= R>>22;
- dest+= step;
+ if (hasAlpha) {\
+ A1 = 1 << 18;
+ A2 = 1 << 18;
+ for (j = 0; j < lumFilterSize; j++) {
+ A1 += alpSrc[j][i * 2 ] * lumFilter[j];
+ A2 += alpSrc[j][i * 2 + 1] * lumFilter[j];
}
- } else {
- if (CONFIG_SWSCALE_ALPHA && c->alpPixBuf) {
- YSCALE_YUV_2_RGBX_FULL_C(1<<21, 1)
- dest[aidx]= A;
- dest[0]= B>>22;
- dest[1]= G>>22;
- dest[2]= R>>22;
- dest+= step;
- }
- } else {
- YSCALE_YUV_2_RGBX_FULL_C(1<<21, 0)
- dest[aidx]= 255;
- dest[0]= B>>22;
- dest[1]= G>>22;
- dest[2]= R>>22;
- dest+= step;
- }
+ A1 >>= 19;
+ A2 >>= 19;
+ if ((A1 | A2) & 0x100) {
+ A1 = av_clip_uint8(A1);
+ A2 = av_clip_uint8(A2);
}
}
- break;
- default:
- assert(0);
+
+ /* FIXME fix tables so that clipping is not needed and then use _NOCLIP*/
+ r = c->table_rV[V];
+ g = (c->table_gU[U] + c->table_gV[V]);
+ b = c->table_bU[U];
+
+ yuv2rgb_write(dest, i, Y1, Y2, U, V, hasAlpha ? A1 : 0, hasAlpha ? A2 : 0,
+ r, g, b, y, target, hasAlpha);
}
}
-/**
- * vertical bilinear scale YV12 to RGB
- */
-static void yuv2packed2_c(SwsContext *c, const uint16_t *buf0,
- const uint16_t *buf1, const uint16_t *ubuf0,
- const uint16_t *ubuf1, const uint16_t *vbuf0,
- const uint16_t *vbuf1, const uint16_t *abuf0,
- const uint16_t *abuf1, uint8_t *dest, int dstW,
- int yalpha, int uvalpha, int y)
-{
- int yalpha1=4095- yalpha;
- int uvalpha1=4095-uvalpha;
+static av_always_inline void
+yuv2rgb_2_c_template(SwsContext *c, const int16_t *buf[2],
+ const int16_t *ubuf[2], const int16_t *vbuf[2],
+ const int16_t *abuf[2], uint8_t *dest, int dstW,
+ int yalpha, int uvalpha, int y,
+ enum PixelFormat target, int hasAlpha)
+{
+ const int16_t *buf0 = buf[0], *buf1 = buf[1],
+ *ubuf0 = ubuf[0], *ubuf1 = ubuf[1],
+ *vbuf0 = vbuf[0], *vbuf1 = vbuf[1],
+ *abuf0 = abuf[0], *abuf1 = abuf[1];
+ int yalpha1 = 4095 - yalpha;
+ int uvalpha1 = 4095 - uvalpha;
int i;
- YSCALE_YUV_2_ANYRGB_C(YSCALE_YUV_2_RGB2_C)
+ for (i = 0; i < (dstW >> 1); i++) {
+ int Y1 = (buf0[i * 2] * yalpha1 + buf1[i * 2] * yalpha) >> 19;
+ int Y2 = (buf0[i * 2 + 1] * yalpha1 + buf1[i * 2 + 1] * yalpha) >> 19;
+ int U = (ubuf0[i] * uvalpha1 + ubuf1[i] * uvalpha) >> 19;
+ int V = (vbuf0[i] * uvalpha1 + vbuf1[i] * uvalpha) >> 19;
+ int A1, A2;
+ const void *r = c->table_rV[V],
+ *g = (c->table_gU[U] + c->table_gV[V]),
+ *b = c->table_bU[U];
+
+ if (hasAlpha) {
+ A1 = (abuf0[i * 2 ] * yalpha1 + abuf1[i * 2 ] * yalpha) >> 19;
+ A2 = (abuf0[i * 2 + 1] * yalpha1 + abuf1[i * 2 + 1] * yalpha) >> 19;
+ }
+
+ yuv2rgb_write(dest, i, Y1, Y2, U, V, hasAlpha ? A1 : 0, hasAlpha ? A2 : 0,
+ r, g, b, y, target, hasAlpha);
+ }
}
-/**
- * YV12 to RGB without scaling or interpolating
- */
-static void yuv2packed1_c(SwsContext *c, const uint16_t *buf0,
- const uint16_t *ubuf0, const uint16_t *ubuf1,
- const uint16_t *vbuf0, const uint16_t *vbuf1,
- const uint16_t *abuf0, uint8_t *dest, int dstW,
- int uvalpha, enum PixelFormat dstFormat,
- int flags, int y)
-{
+static av_always_inline void
+yuv2rgb_1_c_template(SwsContext *c, const int16_t *buf0,
+ const int16_t *ubuf[2], const int16_t *vbuf[2],
+ const int16_t *abuf0, uint8_t *dest, int dstW,
+ int uvalpha, int y, enum PixelFormat target,
+ int hasAlpha)
+{
+ const int16_t *ubuf0 = ubuf[0], *ubuf1 = ubuf[1],
+ *vbuf0 = vbuf[0], *vbuf1 = vbuf[1];
int i;
if (uvalpha < 2048) {
- YSCALE_YUV_2_ANYRGB_C(YSCALE_YUV_2_RGB1_C)
+ for (i = 0; i < (dstW >> 1); i++) {
+ int Y1 = buf0[i * 2] >> 7;
+ int Y2 = buf0[i * 2 + 1] >> 7;
+ int U = ubuf1[i] >> 7;
+ int V = vbuf1[i] >> 7;
+ int A1, A2;
+ const void *r = c->table_rV[V],
+ *g = (c->table_gU[U] + c->table_gV[V]),
+ *b = c->table_bU[U];
+
+ if (hasAlpha) {
+ A1 = abuf0[i * 2 ] >> 7;
+ A2 = abuf0[i * 2 + 1] >> 7;
+ }
+
+ yuv2rgb_write(dest, i, Y1, Y2, U, V, hasAlpha ? A1 : 0, hasAlpha ? A2 : 0,
+ r, g, b, y, target, hasAlpha);
+ }
} else {
- YSCALE_YUV_2_ANYRGB_C(YSCALE_YUV_2_RGB1B_C)
+ for (i = 0; i < (dstW >> 1); i++) {
+ int Y1 = buf0[i * 2] >> 7;
+ int Y2 = buf0[i * 2 + 1] >> 7;
+ int U = (ubuf0[i] + ubuf1[i]) >> 8;
+ int V = (vbuf0[i] + vbuf1[i]) >> 8;
+ int A1, A2;
+ const void *r = c->table_rV[V],
+ *g = (c->table_gU[U] + c->table_gV[V]),
+ *b = c->table_bU[U];
+
+ if (hasAlpha) {
+ A1 = abuf0[i * 2 ] >> 7;
+ A2 = abuf0[i * 2 + 1] >> 7;
+ }
+
+ yuv2rgb_write(dest, i, Y1, Y2, U, V, hasAlpha ? A1 : 0, hasAlpha ? A2 : 0,
+ r, g, b, y, target, hasAlpha);
+ }
}
}
+#define YUV2RGBWRAPPERX(name, base, ext, fmt, hasAlpha) \
+static void name ## ext ## _X_c(SwsContext *c, const int16_t *lumFilter, \
+ const int16_t **lumSrc, int lumFilterSize, \
+ const int16_t *chrFilter, const int16_t **chrUSrc, \
+ const int16_t **chrVSrc, int chrFilterSize, \
+ const int16_t **alpSrc, uint8_t *dest, int dstW, \
+ int y) \
+{ \
+ name ## base ## _X_c_template(c, lumFilter, lumSrc, lumFilterSize, \
+ chrFilter, chrUSrc, chrVSrc, chrFilterSize, \
+ alpSrc, dest, dstW, y, fmt, hasAlpha); \
+}
+#define YUV2RGBWRAPPER(name, base, ext, fmt, hasAlpha) \
+YUV2RGBWRAPPERX(name, base, ext, fmt, hasAlpha) \
+static void name ## ext ## _2_c(SwsContext *c, const int16_t *buf[2], \
+ const int16_t *ubuf[2], const int16_t *vbuf[2], \
+ const int16_t *abuf[2], uint8_t *dest, int dstW, \
+ int yalpha, int uvalpha, int y) \
+{ \
+ name ## base ## _2_c_template(c, buf, ubuf, vbuf, abuf, \
+ dest, dstW, yalpha, uvalpha, y, fmt, hasAlpha); \
+} \
+ \
+static void name ## ext ## _1_c(SwsContext *c, const int16_t *buf0, \
+ const int16_t *ubuf[2], const int16_t *vbuf[2], \
+ const int16_t *abuf0, uint8_t *dest, int dstW, \
+ int uvalpha, int y) \
+{ \
+ name ## base ## _1_c_template(c, buf0, ubuf, vbuf, abuf0, dest, \
+ dstW, uvalpha, y, fmt, hasAlpha); \
+}
+
+#if CONFIG_SMALL
+YUV2RGBWRAPPER(yuv2rgb,, 32_1, PIX_FMT_RGB32_1, CONFIG_SWSCALE_ALPHA && c->alpPixBuf);
+YUV2RGBWRAPPER(yuv2rgb,, 32, PIX_FMT_RGB32, CONFIG_SWSCALE_ALPHA && c->alpPixBuf);
+#else
+#if CONFIG_SWSCALE_ALPHA
+YUV2RGBWRAPPER(yuv2rgb,, a32_1, PIX_FMT_RGB32_1, 1);
+YUV2RGBWRAPPER(yuv2rgb,, a32, PIX_FMT_RGB32, 1);
+#endif
+YUV2RGBWRAPPER(yuv2rgb,, x32_1, PIX_FMT_RGB32_1, 0);
+YUV2RGBWRAPPER(yuv2rgb,, x32, PIX_FMT_RGB32, 0);
+#endif
+YUV2RGBWRAPPER(yuv2, rgb, rgb24, PIX_FMT_RGB24, 0);
+YUV2RGBWRAPPER(yuv2, rgb, bgr24, PIX_FMT_BGR24, 0);
+YUV2RGBWRAPPER(yuv2rgb,, 16, PIX_FMT_RGB565, 0);
+YUV2RGBWRAPPER(yuv2rgb,, 15, PIX_FMT_RGB555, 0);
+YUV2RGBWRAPPER(yuv2rgb,, 12, PIX_FMT_RGB444, 0);
+YUV2RGBWRAPPER(yuv2rgb,, 8, PIX_FMT_RGB8, 0);
+YUV2RGBWRAPPER(yuv2rgb,, 4, PIX_FMT_RGB4, 0);
+YUV2RGBWRAPPER(yuv2rgb,, 4b, PIX_FMT_RGB4_BYTE, 0);
+
+static av_always_inline void
+yuv2rgb_full_X_c_template(SwsContext *c, const int16_t *lumFilter,
+ const int16_t **lumSrc, int lumFilterSize,
+ const int16_t *chrFilter, const int16_t **chrUSrc,
+ const int16_t **chrVSrc, int chrFilterSize,
+ const int16_t **alpSrc, uint8_t *dest,
+ int dstW, int y, enum PixelFormat target, int hasAlpha)
+{
+ int i;
+ int step = (target == PIX_FMT_RGB24 || target == PIX_FMT_BGR24) ? 3 : 4;
+
+ for (i = 0; i < dstW; i++) {
+ int j;
+ int Y = 1<<9;
+ int U = (1<<9)-(128 << 19);
+ int V = (1<<9)-(128 << 19);
+ int av_unused A;
+ int R, G, B;
+
+ for (j = 0; j < lumFilterSize; j++) {
+ Y += lumSrc[j][i] * lumFilter[j];
+ }
+ for (j = 0; j < chrFilterSize; j++) {
+ U += chrUSrc[j][i] * chrFilter[j];
+ V += chrVSrc[j][i] * chrFilter[j];
+ }
+ Y >>= 10;
+ U >>= 10;
+ V >>= 10;
+ if (hasAlpha) {
+ A = 1 << 18;
+ for (j = 0; j < lumFilterSize; j++) {
+ A += alpSrc[j][i] * lumFilter[j];
+ }
+ A >>= 19;
+ if (A & 0x100)
+ A = av_clip_uint8(A);
+ }
+ Y -= c->yuv2rgb_y_offset;
+ Y *= c->yuv2rgb_y_coeff;
+ Y += 1 << 21;
+ R = Y + V*c->yuv2rgb_v2r_coeff;
+ G = Y + V*c->yuv2rgb_v2g_coeff + U*c->yuv2rgb_u2g_coeff;
+ B = Y + U*c->yuv2rgb_u2b_coeff;
+ if ((R | G | B) & 0xC0000000) {
+ R = av_clip_uintp2(R, 30);
+ G = av_clip_uintp2(G, 30);
+ B = av_clip_uintp2(B, 30);
+ }
+
+ switch(target) {
+ case PIX_FMT_ARGB:
+ dest[0] = hasAlpha ? A : 255;
+ dest[1] = R >> 22;
+ dest[2] = G >> 22;
+ dest[3] = B >> 22;
+ break;
+ case PIX_FMT_RGB24:
+ dest[0] = R >> 22;
+ dest[1] = G >> 22;
+ dest[2] = B >> 22;
+ break;
+ case PIX_FMT_RGBA:
+ dest[0] = R >> 22;
+ dest[1] = G >> 22;
+ dest[2] = B >> 22;
+ dest[3] = hasAlpha ? A : 255;
+ break;
+ case PIX_FMT_ABGR:
+ dest[0] = hasAlpha ? A : 255;
+ dest[1] = B >> 22;
+ dest[2] = G >> 22;
+ dest[3] = R >> 22;
+ break;
+ case PIX_FMT_BGR24:
+ dest[0] = B >> 22;
+ dest[1] = G >> 22;
+ dest[2] = R >> 22;
+ break;
+ case PIX_FMT_BGRA:
+ dest[0] = B >> 22;
+ dest[1] = G >> 22;
+ dest[2] = R >> 22;
+ dest[3] = hasAlpha ? A : 255;
+ break;
+ }
+ dest += step;
+ }
+}
+
+#if CONFIG_SMALL
+YUV2RGBWRAPPERX(yuv2, rgb_full, bgra32_full, PIX_FMT_BGRA, CONFIG_SWSCALE_ALPHA && c->alpPixBuf);
+YUV2RGBWRAPPERX(yuv2, rgb_full, abgr32_full, PIX_FMT_ABGR, CONFIG_SWSCALE_ALPHA && c->alpPixBuf);
+YUV2RGBWRAPPERX(yuv2, rgb_full, rgba32_full, PIX_FMT_RGBA, CONFIG_SWSCALE_ALPHA && c->alpPixBuf);
+YUV2RGBWRAPPERX(yuv2, rgb_full, argb32_full, PIX_FMT_ARGB, CONFIG_SWSCALE_ALPHA && c->alpPixBuf);
+#else
+#if CONFIG_SWSCALE_ALPHA
+YUV2RGBWRAPPERX(yuv2, rgb_full, bgra32_full, PIX_FMT_BGRA, 1);
+YUV2RGBWRAPPERX(yuv2, rgb_full, abgr32_full, PIX_FMT_ABGR, 1);
+YUV2RGBWRAPPERX(yuv2, rgb_full, rgba32_full, PIX_FMT_RGBA, 1);
+YUV2RGBWRAPPERX(yuv2, rgb_full, argb32_full, PIX_FMT_ARGB, 1);
+#endif
+YUV2RGBWRAPPERX(yuv2, rgb_full, bgrx32_full, PIX_FMT_BGRA, 0);
+YUV2RGBWRAPPERX(yuv2, rgb_full, xbgr32_full, PIX_FMT_ABGR, 0);
+YUV2RGBWRAPPERX(yuv2, rgb_full, rgbx32_full, PIX_FMT_RGBA, 0);
+YUV2RGBWRAPPERX(yuv2, rgb_full, xrgb32_full, PIX_FMT_ARGB, 0);
+#endif
+YUV2RGBWRAPPERX(yuv2, rgb_full, bgr24_full, PIX_FMT_BGR24, 0);
+YUV2RGBWRAPPERX(yuv2, rgb_full, rgb24_full, PIX_FMT_RGB24, 0);
+
static av_always_inline void fillPlane(uint8_t* plane, int stride,
int width, int height,
int y, uint8_t val)
*yuv2yuvX = yuv2yuvX_c;
}
if(c->flags & SWS_FULL_CHR_H_INT) {
- *yuv2packedX = yuv2rgbX_c_full;
+ switch (dstFormat) {
+ case PIX_FMT_RGBA:
+#if CONFIG_SMALL
+ *yuv2packedX = yuv2rgba32_full_X_c;
+#else
+#if CONFIG_SWSCALE_ALPHA
+ if (c->alpPixBuf) {
+ *yuv2packedX = yuv2rgba32_full_X_c;
+ } else
+#endif /* CONFIG_SWSCALE_ALPHA */
+ {
+ *yuv2packedX = yuv2rgbx32_full_X_c;
+ }
+#endif /* !CONFIG_SMALL */
+ break;
+ case PIX_FMT_ARGB:
+#if CONFIG_SMALL
+ *yuv2packedX = yuv2argb32_full_X_c;
+#else
+#if CONFIG_SWSCALE_ALPHA
+ if (c->alpPixBuf) {
+ *yuv2packedX = yuv2argb32_full_X_c;
+ } else
+#endif /* CONFIG_SWSCALE_ALPHA */
+ {
+ *yuv2packedX = yuv2xrgb32_full_X_c;
+ }
+#endif /* !CONFIG_SMALL */
+ break;
+ case PIX_FMT_BGRA:
+#if CONFIG_SMALL
+ *yuv2packedX = yuv2bgra32_full_X_c;
+#else
+#if CONFIG_SWSCALE_ALPHA
+ if (c->alpPixBuf) {
+ *yuv2packedX = yuv2bgra32_full_X_c;
+ } else
+#endif /* CONFIG_SWSCALE_ALPHA */
+ {
+ *yuv2packedX = yuv2bgrx32_full_X_c;
+ }
+#endif /* !CONFIG_SMALL */
+ break;
+ case PIX_FMT_ABGR:
+#if CONFIG_SMALL
+ *yuv2packedX = yuv2abgr32_full_X_c;
+#else
+#if CONFIG_SWSCALE_ALPHA
+ if (c->alpPixBuf) {
+ *yuv2packedX = yuv2abgr32_full_X_c;
+ } else
+#endif /* CONFIG_SWSCALE_ALPHA */
+ {
+ *yuv2packedX = yuv2xbgr32_full_X_c;
+ }
+#endif /* !CONFIG_SMALL */
+ break;
+ case PIX_FMT_RGB24:
+ *yuv2packedX = yuv2rgb24_full_X_c;
+ break;
+ case PIX_FMT_BGR24:
+ *yuv2packedX = yuv2bgr24_full_X_c;
+ break;
+ }
+ if(!*yuv2packedX)
+ goto YUV_PACKED;
} else {
+ YUV_PACKED:
switch (dstFormat) {
case PIX_FMT_GRAY16BE:
*yuv2packed1 = yuv2gray16BE_1_c;
*yuv2packed2 = yuv2bgr48be_2_c;
*yuv2packedX = yuv2bgr48be_X_c;
break;
- default:
- *yuv2packed1 = yuv2packed1_c;
- *yuv2packed2 = yuv2packed2_c;
- *yuv2packedX = yuv2packedX_c;
+ case PIX_FMT_RGB32:
+ case PIX_FMT_BGR32:
+#if CONFIG_SMALL
+ *yuv2packed1 = yuv2rgb32_1_c;
+ *yuv2packed2 = yuv2rgb32_2_c;
+ *yuv2packedX = yuv2rgb32_X_c;
+#else
+#if CONFIG_SWSCALE_ALPHA
+ if (c->alpPixBuf) {
+ *yuv2packed1 = yuv2rgba32_1_c;
+ *yuv2packed2 = yuv2rgba32_2_c;
+ *yuv2packedX = yuv2rgba32_X_c;
+ } else
+#endif /* CONFIG_SWSCALE_ALPHA */
+ {
+ *yuv2packed1 = yuv2rgbx32_1_c;
+ *yuv2packed2 = yuv2rgbx32_2_c;
+ *yuv2packedX = yuv2rgbx32_X_c;
+ }
+#endif /* !CONFIG_SMALL */
+ break;
+ case PIX_FMT_RGB32_1:
+ case PIX_FMT_BGR32_1:
+#if CONFIG_SMALL
+ *yuv2packed1 = yuv2rgb32_1_1_c;
+ *yuv2packed2 = yuv2rgb32_1_2_c;
+ *yuv2packedX = yuv2rgb32_1_X_c;
+#else
+#if CONFIG_SWSCALE_ALPHA
+ if (c->alpPixBuf) {
+ *yuv2packed1 = yuv2rgba32_1_1_c;
+ *yuv2packed2 = yuv2rgba32_1_2_c;
+ *yuv2packedX = yuv2rgba32_1_X_c;
+ } else
+#endif /* CONFIG_SWSCALE_ALPHA */
+ {
+ *yuv2packed1 = yuv2rgbx32_1_1_c;
+ *yuv2packed2 = yuv2rgbx32_1_2_c;
+ *yuv2packedX = yuv2rgbx32_1_X_c;
+ }
+#endif /* !CONFIG_SMALL */
+ break;
+ case PIX_FMT_RGB24:
+ *yuv2packed1 = yuv2rgb24_1_c;
+ *yuv2packed2 = yuv2rgb24_2_c;
+ *yuv2packedX = yuv2rgb24_X_c;
+ break;
+ case PIX_FMT_BGR24:
+ *yuv2packed1 = yuv2bgr24_1_c;
+ *yuv2packed2 = yuv2bgr24_2_c;
+ *yuv2packedX = yuv2bgr24_X_c;
+ break;
+ case PIX_FMT_RGB565LE:
+ case PIX_FMT_RGB565BE:
+ case PIX_FMT_BGR565LE:
+ case PIX_FMT_BGR565BE:
+ *yuv2packed1 = yuv2rgb16_1_c;
+ *yuv2packed2 = yuv2rgb16_2_c;
+ *yuv2packedX = yuv2rgb16_X_c;
+ break;
+ case PIX_FMT_RGB555LE:
+ case PIX_FMT_RGB555BE:
+ case PIX_FMT_BGR555LE:
+ case PIX_FMT_BGR555BE:
+ *yuv2packed1 = yuv2rgb15_1_c;
+ *yuv2packed2 = yuv2rgb15_2_c;
+ *yuv2packedX = yuv2rgb15_X_c;
+ break;
+ case PIX_FMT_RGB444LE:
+ case PIX_FMT_RGB444BE:
+ case PIX_FMT_BGR444LE:
+ case PIX_FMT_BGR444BE:
+ *yuv2packed1 = yuv2rgb12_1_c;
+ *yuv2packed2 = yuv2rgb12_2_c;
+ *yuv2packedX = yuv2rgb12_X_c;
+ break;
+ case PIX_FMT_RGB8:
+ case PIX_FMT_BGR8:
+ *yuv2packed1 = yuv2rgb8_1_c;
+ *yuv2packed2 = yuv2rgb8_2_c;
+ *yuv2packedX = yuv2rgb8_X_c;
+ break;
+ case PIX_FMT_RGB4:
+ case PIX_FMT_BGR4:
+ *yuv2packed1 = yuv2rgb4_1_c;
+ *yuv2packed2 = yuv2rgb4_2_c;
+ *yuv2packedX = yuv2rgb4_X_c;
+ break;
+ case PIX_FMT_RGB4_BYTE:
+ case PIX_FMT_BGR4_BYTE:
+ *yuv2packed1 = yuv2rgb4b_1_c;
+ *yuv2packed2 = yuv2rgb4b_2_c;
+ *yuv2packedX = yuv2rgb4b_X_c;
break;
}
}
lastDstY= dstY;
for (;dstY < dstH; dstY++) {
- unsigned char *dest =dst[0]+dstStride[0]*dstY;
const int chrDstY= dstY>>c->chrDstVSubSample;
- unsigned char *uDest=dst[1]+dstStride[1]*chrDstY;
- unsigned char *vDest=dst[2]+dstStride[2]*chrDstY;
- unsigned char *aDest=(CONFIG_SWSCALE_ALPHA && alpPixBuf) ? dst[3]+dstStride[3]*dstY : NULL;
+ uint8_t *dest[4] = {
+ dst[0] + dstStride[0] * dstY,
+ dst[1] + dstStride[1] * chrDstY,
+ dst[2] + dstStride[2] * chrDstY,
+ (CONFIG_SWSCALE_ALPHA && alpPixBuf) ? dst[3] + dstStride[3] * dstY : NULL,
+ };
const uint8_t *lumDither= should_dither ? dithers[7][dstY &7] : flat64;
const uint8_t *chrDither= should_dither ? dithers[7][chrDstY&7] : flat64;
if (isPlanarYUV(dstFormat) || dstFormat==PIX_FMT_GRAY8) { //YV12 like
const int chrSkipMask= (1<<c->chrDstVSubSample)-1;
- if ((dstY&chrSkipMask) || isGray(dstFormat)) uDest=vDest= NULL; //FIXME split functions in lumi / chromi
+ if ((dstY&chrSkipMask) || isGray(dstFormat))
+ dest[1] = dest[2] = NULL; //FIXME split functions in lumi / chromi
if (c->yuv2yuv1 && vLumFilterSize == 1 && vChrFilterSize == 1) { // unscaled YV12
- const int16_t *lumBuf = lumSrcPtr[0];
- const int16_t *chrUBuf= chrUSrcPtr[0];
- const int16_t *chrVBuf= chrVSrcPtr[0];
const int16_t *alpBuf= (CONFIG_SWSCALE_ALPHA && alpPixBuf) ? alpSrcPtr[0] : NULL;
- yuv2yuv1(c, lumBuf, chrUBuf, chrVBuf, alpBuf, dest,
- uDest, vDest, aDest, dstW, chrDstW, lumDither, chrDither);
+ yuv2yuv1(c, lumSrcPtr[0], chrUSrcPtr[0], chrVSrcPtr[0], alpBuf,
+ dest, dstW, chrDstW, lumDither, chrDither);
} else { //General YV12
- yuv2yuvX(c,
- vLumFilter+dstY*vLumFilterSize , lumSrcPtr, vLumFilterSize,
- vChrFilter+chrDstY*vChrFilterSize, chrUSrcPtr,
- chrVSrcPtr, vChrFilterSize,
- alpSrcPtr, dest, uDest, vDest, aDest, dstW, chrDstW, lumDither, chrDither);
+ yuv2yuvX(c, vLumFilter + dstY * vLumFilterSize,
+ lumSrcPtr, vLumFilterSize,
+ vChrFilter + chrDstY * vChrFilterSize,
+ chrUSrcPtr, chrVSrcPtr, vChrFilterSize,
+ alpSrcPtr, dest, dstW, chrDstW, lumDither, chrDither);
}
} else {
assert(lumSrcPtr + vLumFilterSize - 1 < lumPixBuf + vLumBufSize*2);
assert(chrUSrcPtr + vChrFilterSize - 1 < chrUPixBuf + vChrBufSize*2);
if (c->yuv2packed1 && vLumFilterSize == 1 && vChrFilterSize == 2) { //unscaled RGB
- int chrAlpha= vChrFilter[2*dstY+1];
- yuv2packed1(c, *lumSrcPtr, *chrUSrcPtr, *(chrUSrcPtr+1),
- *chrVSrcPtr, *(chrVSrcPtr+1),
- alpPixBuf ? *alpSrcPtr : NULL,
- dest, dstW, chrAlpha, dstFormat, flags, dstY);
+ int chrAlpha = vChrFilter[2 * dstY + 1];
+ yuv2packed1(c, *lumSrcPtr, chrUSrcPtr, chrVSrcPtr,
+ alpPixBuf ? *alpSrcPtr : NULL,
+ dest[0], dstW, chrAlpha, dstY);
} else if (c->yuv2packed2 && vLumFilterSize == 2 && vChrFilterSize == 2) { //bilinear upscale RGB
- int lumAlpha= vLumFilter[2*dstY+1];
- int chrAlpha= vChrFilter[2*dstY+1];
- lumMmxFilter[2]=
- lumMmxFilter[3]= vLumFilter[2*dstY ]*0x10001;
- chrMmxFilter[2]=
- chrMmxFilter[3]= vChrFilter[2*chrDstY]*0x10001;
- yuv2packed2(c, *lumSrcPtr, *(lumSrcPtr+1), *chrUSrcPtr, *(chrUSrcPtr+1),
- *chrVSrcPtr, *(chrVSrcPtr+1),
- alpPixBuf ? *alpSrcPtr : NULL, alpPixBuf ? *(alpSrcPtr+1) : NULL,
- dest, dstW, lumAlpha, chrAlpha, dstY);
+ int lumAlpha = vLumFilter[2 * dstY + 1];
+ int chrAlpha = vChrFilter[2 * dstY + 1];
+ lumMmxFilter[2] =
+ lumMmxFilter[3] = vLumFilter[2 * dstY ] * 0x10001;
+ chrMmxFilter[2] =
+ chrMmxFilter[3] = vChrFilter[2 * chrDstY] * 0x10001;
+ yuv2packed2(c, lumSrcPtr, chrUSrcPtr, chrVSrcPtr,
+ alpPixBuf ? alpSrcPtr : NULL,
+ dest[0], dstW, lumAlpha, chrAlpha, dstY);
} else { //general RGB
- yuv2packedX(c,
- vLumFilter+dstY*vLumFilterSize, lumSrcPtr, vLumFilterSize,
- vChrFilter+dstY*vChrFilterSize, chrUSrcPtr, chrVSrcPtr, vChrFilterSize,
- alpSrcPtr, dest, dstW, dstY);
+ yuv2packedX(c, vLumFilter + dstY * vLumFilterSize,
+ lumSrcPtr, vLumFilterSize,
+ vChrFilter + dstY * vChrFilterSize,
+ chrUSrcPtr, chrVSrcPtr, vChrFilterSize,
+ alpSrcPtr, dest[0], dstW, dstY);
}
}
}
int srcStride[], int srcSliceY, int srcSliceH,
uint8_t* dst[], int dstStride[]);
+/**
+ * Write one line of horizontally scaled Y/U/V/A to planar output
+ * without any additional vertical scaling (or point-scaling).
+ *
+ * @param c SWS scaling context
+ * @param lumSrc scaled luma (Y) source data, 15bit for 8bit output
+ * @param chrUSrc scaled chroma (U) source data, 15bit for 8bit output
+ * @param chrVSrc scaled chroma (V) source data, 15bit for 8bit output
+ * @param alpSrc scaled alpha (A) source data, 15bit for 8bit output
+ * @param dest pointer to the 4 output planes (Y/U/V/A)
+ * @param dstW width of dest[0], dest[3], lumSrc and alpSrc in pixels
+ * @param chrDstW width of dest[1], dest[2], chrUSrc and chrVSrc
+ */
typedef void (*yuv2planar1_fn) (struct SwsContext *c,
const int16_t *lumSrc, const int16_t *chrUSrc,
const int16_t *chrVSrc, const int16_t *alpSrc,
- uint8_t *dest,
- uint8_t *uDest, uint8_t *vDest, uint8_t *aDest,
- int dstW, int chrDstW, const uint8_t *lumDither, const uint8_t *chrDither);
-typedef void (*yuv2planarX_fn) (struct SwsContext *c,
- const int16_t *lumFilter, const int16_t **lumSrc, int lumFilterSize,
+ uint8_t *dest[4], int dstW, int chrDstW,
+ const uint8_t *lumDither, const uint8_t *chrDither);
+/**
+ * Write one line of horizontally scaled Y/U/V/A to planar output
+ * with multi-point vertical scaling between input pixels.
+ *
+ * @param c SWS scaling context
+ * @param lumFilter vertical luma/alpha scaling coefficients, 12bit [0,4096]
+ * @param lumSrc scaled luma (Y) source data, 15bit for 8bit output
+ * @param lumFilterSize number of vertical luma/alpha input lines to scale
+ * @param chrFilter vertical chroma scaling coefficients, 12bit [0,4096]
+ * @param chrUSrc scaled chroma (U) source data, 15bit for 8bit output
+ * @param chrVSrc scaled chroma (V) source data, 15bit for 8bit output
+ * @param chrFilterSize number of vertical chroma input lines to scale
+ * @param alpSrc scaled alpha (A) source data, 15bit for 8bit output
+ * @param dest pointer to the 4 output planes (Y/U/V/A)
+ * @param dstW width of dest[0], dest[3], lumSrc and alpSrc in pixels
+ * @param chrDstW width of dest[1], dest[2], chrUSrc and chrVSrc
+ */
+typedef void (*yuv2planarX_fn) (struct SwsContext *c, const int16_t *lumFilter,
+ const int16_t **lumSrc, int lumFilterSize,
const int16_t *chrFilter, const int16_t **chrUSrc,
- const int16_t **chrVSrc, int chrFilterSize,
- const int16_t **alpSrc,
- uint8_t *dest,
- uint8_t *uDest, uint8_t *vDest, uint8_t *aDest,
+ const int16_t **chrVSrc, int chrFilterSize,
+ const int16_t **alpSrc, uint8_t *dest[4],
int dstW, int chrDstW, const uint8_t *lumDither, const uint8_t *chrDither);
-typedef void (*yuv2packed1_fn) (struct SwsContext *c,
- const uint16_t *buf0,
- const uint16_t *ubuf0, const uint16_t *ubuf1,
- const uint16_t *vbuf0, const uint16_t *vbuf1,
- const uint16_t *abuf0,
- uint8_t *dest,
- int dstW, int uvalpha, int dstFormat, int flags, int y);
-typedef void (*yuv2packed2_fn) (struct SwsContext *c,
- const uint16_t *buf0, const uint16_t *buf1,
- const uint16_t *ubuf0, const uint16_t *ubuf1,
- const uint16_t *vbuf0, const uint16_t *vbuf1,
- const uint16_t *abuf0, const uint16_t *abuf1,
- uint8_t *dest,
+/**
+ * Write one line of horizontally scaled Y/U/V/A to packed-pixel YUV/RGB
+ * output without any additional vertical scaling (or point-scaling). Note
+ * that this function may do chroma scaling, see the "uvalpha" argument.
+ *
+ * @param c SWS scaling context
+ * @param lumSrc scaled luma (Y) source data, 15bit for 8bit output
+ * @param chrUSrc scaled chroma (U) source data, 15bit for 8bit output
+ * @param chrVSrc scaled chroma (V) source data, 15bit for 8bit output
+ * @param alpSrc scaled alpha (A) source data, 15bit for 8bit output
+ * @param dest pointer to the output plane
+ * @param dstW width of lumSrc and alpSrc in pixels, number of pixels
+ * to write into dest[]
+ * @param uvalpha chroma scaling coefficient for the second line of chroma
+ * pixels, either 2048 or 0. If 0, one chroma input is used
+ * for 2 output pixels (or if the SWS_FLAG_FULL_CHR_INT flag
+ * is set, it generates 1 output pixel). If 2048, two chroma
+ * input pixels should be averaged for 2 output pixels (this
+ * only happens if SWS_FLAG_FULL_CHR_INT is not set)
+ * @param y vertical line number for this output. This does not need
+ * to be used to calculate the offset in the destination,
+ * but can be used to generate comfort noise using dithering
+ * for some output formats.
+ */
+typedef void (*yuv2packed1_fn) (struct SwsContext *c, const int16_t *lumSrc,
+ const int16_t *chrUSrc[2], const int16_t *chrVSrc[2],
+ const int16_t *alpSrc, uint8_t *dest,
+ int dstW, int uvalpha, int y);
+/**
+ * Write one line of horizontally scaled Y/U/V/A to packed-pixel YUV/RGB
+ * output by doing bilinear scaling between two input lines.
+ *
+ * @param c SWS scaling context
+ * @param lumSrc scaled luma (Y) source data, 15bit for 8bit output
+ * @param chrUSrc scaled chroma (U) source data, 15bit for 8bit output
+ * @param chrVSrc scaled chroma (V) source data, 15bit for 8bit output
+ * @param alpSrc scaled alpha (A) source data, 15bit for 8bit output
+ * @param dest pointer to the output plane
+ * @param dstW width of lumSrc and alpSrc in pixels, number of pixels
+ * to write into dest[]
+ * @param yalpha luma/alpha scaling coefficients for the second input line.
+ * The first line's coefficients can be calculated by using
+ * 4096 - yalpha
+ * @param uvalpha chroma scaling coefficient for the second input line. The
+ * first line's coefficients can be calculated by using
+ * 4096 - uvalpha
+ * @param y vertical line number for this output. This does not need
+ * to be used to calculate the offset in the destination,
+ * but can be used to generate comfort noise using dithering
+ * for some output formats.
+ */
+typedef void (*yuv2packed2_fn) (struct SwsContext *c, const int16_t *lumSrc[2],
+ const int16_t *chrUSrc[2], const int16_t *chrVSrc[2],
+ const int16_t *alpSrc[2], uint8_t *dest,
int dstW, int yalpha, int uvalpha, int y);
-typedef void (*yuv2packedX_fn) (struct SwsContext *c,
- const int16_t *lumFilter, const int16_t **lumSrc, int lumFilterSize,
+/**
+ * Write one line of horizontally scaled Y/U/V/A to packed-pixel YUV/RGB
+ * output by doing multi-point vertical scaling between input pixels.
+ *
+ * @param c SWS scaling context
+ * @param lumFilter vertical luma/alpha scaling coefficients, 12bit [0,4096]
+ * @param lumSrc scaled luma (Y) source data, 15bit for 8bit output
+ * @param lumFilterSize number of vertical luma/alpha input lines to scale
+ * @param chrFilter vertical chroma scaling coefficients, 12bit [0,4096]
+ * @param chrUSrc scaled chroma (U) source data, 15bit for 8bit output
+ * @param chrVSrc scaled chroma (V) source data, 15bit for 8bit output
+ * @param chrFilterSize number of vertical chroma input lines to scale
+ * @param alpSrc scaled alpha (A) source data, 15bit for 8bit output
+ * @param dest pointer to the output plane
+ * @param dstW width of lumSrc and alpSrc in pixels, number of pixels
+ * to write into dest[]
+ * @param y vertical line number for this output. This does not need
+ * to be used to calculate the offset in the destination,
+ * but can be used to generate comfort noise using dithering
+ * or some output formats.
+ */
+typedef void (*yuv2packedX_fn) (struct SwsContext *c, const int16_t *lumFilter,
+ const int16_t **lumSrc, int lumFilterSize,
const int16_t *chrFilter, const int16_t **chrUSrc,
const int16_t **chrVSrc, int chrFilterSize,
const int16_t **alpSrc, uint8_t *dest,
- int dstW, int dstY);
+ int dstW, int y);
/* This struct should be aligned on at least a 32-byte boundary. */
typedef struct SwsContext {
void (*chrToYV12)(uint8_t *dstU, uint8_t *dstV,
const uint8_t *src1, const uint8_t *src2,
int width, uint32_t *pal); ///< Unscaled conversion of chroma planes to YV12 for horizontal scaler.
+ /**
+ * Scale one horizontal line of input data using a bilinear filter
+ * to produce one line of output data. Compared to SwsContext->hScale(),
+ * please take note of the following caveats when using these:
+ * - Scaling is done using only 7bit instead of 14bit coefficients.
+ * - You can use no more than 5 input pixels to produce 4 output
+ * pixels. Therefore, this filter should not be used for downscaling
+ * by more than ~20% in width (because that equals more than 5/4th
+ * downscaling and thus more than 5 pixels input per 4 pixels output).
+ * - In general, bilinear filters create artifacts during downscaling
+ * (even when <20%), because one output pixel will span more than one
+ * input pixel, and thus some pixels will need edges of both neighbor
+ * pixels to interpolate the output pixel. Since you can use at most
+ * two input pixels per output pixel in bilinear scaling, this is
+ * impossible and thus downscaling by any size will create artifacts.
+ * To enable this type of scaling, set SWS_FLAG_FAST_BILINEAR
+ * in SwsContext->flags.
+ */
+ /** @{ */
void (*hyscale_fast)(struct SwsContext *c,
int16_t *dst, int dstWidth,
const uint8_t *src, int srcW, int xInc);
int16_t *dst1, int16_t *dst2, int dstWidth,
const uint8_t *src1, const uint8_t *src2,
int srcW, int xInc);
+ /** @} */
+ /**
+ * Scale one horizontal line of input data using a filter over the input
+ * lines, to produce one (differently sized) line of output data.
+ *
+ * @param dst pointer to destination buffer for horizontally scaled
+ * data. If the scaling depth (SwsContext->scalingBpp) is
+ * 8, data will be 15bpp in 16bits (int16_t) width. If
+ * scaling depth is 16, data will be 19bpp in 32bpp
+ * (int32_t) width.
+ * @param dstW width of destination image
+ * @param src pointer to source data to be scaled. If scaling depth
+ * is 8, this is 8bpp in 8bpp (uint8_t) width. If scaling
+ * depth is 16, this is 16bpp in 16bpp (uint16_t) depth.
+ * @param filter filter coefficients to be used per output pixel for
+ * scaling. This contains 14bpp filtering coefficients.
+ * Guaranteed to contain dstW * filterSize entries.
+ * @param filterPos position of the first input pixel to be used for
+ * each output pixel during scaling. Guaranteed to
+ * contain dstW entries.
+ * @param filterSize the number of input coefficients to be used (and
+ * thus the number of input pixels to be used) for
+ * creating a single output pixel. Is aligned to 4
+ * (and input coefficients thus padded with zeroes)
+ * to simplify creating SIMD code.
+ */
void (*hScale)(int16_t *dst, int dstW, const uint8_t *src,
const int16_t *filter, const int16_t *filterPos,
int filterSize);
const int16_t *chrFilter, const int16_t **chrUSrc,
const int16_t **chrVSrc,
int chrFilterSize, const int16_t **alpSrc,
- uint8_t *dest, uint8_t *uDest, uint8_t *vDest,
- uint8_t *aDest, int dstW, int chrDstW,
+ uint8_t *dest[4], int dstW, int chrDstW,
const uint8_t *lumDither, const uint8_t *chrDither)
{
int i;
+ uint8_t *yDest = dest[0], *uDest = dest[1], *vDest = dest[2],
+ *aDest = CONFIG_SWSCALE_ALPHA ? dest[3] : NULL;
+
if (uDest) {
x86_reg uv_off = c->uv_off;
for(i=0; i<8; i++) c->dither16[i] = chrDither[i]>>4;
YSCALEYUV2YV12X(ALP_MMX_FILTER_OFFSET, aDest, dstW, 0)
}
- YSCALEYUV2YV12X(LUM_MMX_FILTER_OFFSET, dest, dstW, 0)
+ YSCALEYUV2YV12X(LUM_MMX_FILTER_OFFSET, yDest, dstW, 0)
}
#define YSCALEYUV2YV12X_ACCURATE(offset, dest, end, pos) \
const int16_t *chrFilter, const int16_t **chrUSrc,
const int16_t **chrVSrc,
int chrFilterSize, const int16_t **alpSrc,
- uint8_t *dest, uint8_t *uDest, uint8_t *vDest,
- uint8_t *aDest, int dstW, int chrDstW,
+ uint8_t *dest[4], int dstW, int chrDstW,
const uint8_t *lumDither, const uint8_t *chrDither)
{
int i;
+ uint8_t *yDest = dest[0], *uDest = dest[1], *vDest = dest[2],
+ *aDest = CONFIG_SWSCALE_ALPHA ? dest[3] : NULL;
+
if (uDest) {
x86_reg uv_off = c->uv_off;
for(i=0; i<8; i++) c->dither32[i] = chrDither[i]<<12;
YSCALEYUV2YV12X_ACCURATE(ALP_MMX_FILTER_OFFSET, aDest, dstW, 0)
}
- YSCALEYUV2YV12X_ACCURATE(LUM_MMX_FILTER_OFFSET, dest, dstW, 0)
+ YSCALEYUV2YV12X_ACCURATE(LUM_MMX_FILTER_OFFSET, yDest, dstW, 0)
}
static void RENAME(yuv2yuv1)(SwsContext *c, const int16_t *lumSrc,
const int16_t *chrUSrc, const int16_t *chrVSrc,
const int16_t *alpSrc,
- uint8_t *dest, uint8_t *uDest, uint8_t *vDest,
- uint8_t *aDest, int dstW, int chrDstW,
+ uint8_t *dst[4], int dstW, int chrDstW,
const uint8_t *lumDither, const uint8_t *chrDither)
{
int p= 4;
- const int16_t *src[4]= { alpSrc + dstW, lumSrc + dstW, chrUSrc + chrDstW, chrVSrc + chrDstW };
- uint8_t *dst[4]= { aDest, dest, uDest, vDest };
- x86_reg counter[4]= { dstW, dstW, chrDstW, chrDstW };
+ const int16_t *src[4]= {
+ lumSrc + dstW, chrUSrc + chrDstW,
+ chrVSrc + chrDstW, alpSrc + dstW
+ };
+ x86_reg counter[4]= { dstW, chrDstW, chrDstW, dstW };
while (p--) {
if (dst[p]) {
static void RENAME(yuv2yuv1_ar)(SwsContext *c, const int16_t *lumSrc,
const int16_t *chrUSrc, const int16_t *chrVSrc,
const int16_t *alpSrc,
- uint8_t *dest, uint8_t *uDest, uint8_t *vDest,
- uint8_t *aDest, int dstW, int chrDstW,
+ uint8_t *dst[4], int dstW, int chrDstW,
const uint8_t *lumDither, const uint8_t *chrDither)
{
int p= 4;
- const int16_t *src[4]= { alpSrc + dstW, lumSrc + dstW, chrUSrc + chrDstW, chrVSrc + chrDstW };
- uint8_t *dst[4]= { aDest, dest, uDest, vDest };
- x86_reg counter[4]= { dstW, dstW, chrDstW, chrDstW };
+ const int16_t *src[4]= {
+ lumSrc + dstW, chrUSrc + chrDstW,
+ chrVSrc + chrDstW, alpSrc + dstW
+ };
+ x86_reg counter[4]= { dstW, chrDstW, chrDstW, dstW };
while (p--) {
if (dst[p]) {
/**
* vertical bilinear scale YV12 to RGB
*/
-static void RENAME(yuv2rgb32_2)(SwsContext *c, const uint16_t *buf0,
- const uint16_t *buf1, const uint16_t *ubuf0,
- const uint16_t *ubuf1, const uint16_t *vbuf0,
- const uint16_t *vbuf1, const uint16_t *abuf0,
- const uint16_t *abuf1, uint8_t *dest,
+static void RENAME(yuv2rgb32_2)(SwsContext *c, const int16_t *buf[2],
+ const int16_t *ubuf[2], const int16_t *vbuf[2],
+ const int16_t *abuf[2], uint8_t *dest,
int dstW, int yalpha, int uvalpha, int y)
{
+ const int16_t *buf0 = buf[0], *buf1 = buf[1],
+ *ubuf0 = ubuf[0], *ubuf1 = ubuf[1];
+
if (CONFIG_SWSCALE_ALPHA && c->alpPixBuf) {
+ const int16_t *abuf0 = abuf[0], *abuf1 = abuf[1];
#if ARCH_X86_64
__asm__ volatile(
YSCALEYUV2RGB(%%r8, %5)
}
}
-static void RENAME(yuv2bgr24_2)(SwsContext *c, const uint16_t *buf0,
- const uint16_t *buf1, const uint16_t *ubuf0,
- const uint16_t *ubuf1, const uint16_t *vbuf0,
- const uint16_t *vbuf1, const uint16_t *abuf0,
- const uint16_t *abuf1, uint8_t *dest,
+static void RENAME(yuv2bgr24_2)(SwsContext *c, const int16_t *buf[2],
+ const int16_t *ubuf[2], const int16_t *vbuf[2],
+ const int16_t *abuf[2], uint8_t *dest,
int dstW, int yalpha, int uvalpha, int y)
{
+ const int16_t *buf0 = buf[0], *buf1 = buf[1],
+ *ubuf0 = ubuf[0], *ubuf1 = ubuf[1];
+
//Note 8280 == DSTW_OFFSET but the preprocessor can't handle that there :(
__asm__ volatile(
"mov %%"REG_b", "ESP_OFFSET"(%5) \n\t"
);
}
-static void RENAME(yuv2rgb555_2)(SwsContext *c, const uint16_t *buf0,
- const uint16_t *buf1, const uint16_t *ubuf0,
- const uint16_t *ubuf1, const uint16_t *vbuf0,
- const uint16_t *vbuf1, const uint16_t *abuf0,
- const uint16_t *abuf1, uint8_t *dest,
+static void RENAME(yuv2rgb555_2)(SwsContext *c, const int16_t *buf[2],
+ const int16_t *ubuf[2], const int16_t *vbuf[2],
+ const int16_t *abuf[2], uint8_t *dest,
int dstW, int yalpha, int uvalpha, int y)
{
+ const int16_t *buf0 = buf[0], *buf1 = buf[1],
+ *ubuf0 = ubuf[0], *ubuf1 = ubuf[1];
+
//Note 8280 == DSTW_OFFSET but the preprocessor can't handle that there :(
__asm__ volatile(
"mov %%"REG_b", "ESP_OFFSET"(%5) \n\t"
);
}
-static void RENAME(yuv2rgb565_2)(SwsContext *c, const uint16_t *buf0,
- const uint16_t *buf1, const uint16_t *ubuf0,
- const uint16_t *ubuf1, const uint16_t *vbuf0,
- const uint16_t *vbuf1, const uint16_t *abuf0,
- const uint16_t *abuf1, uint8_t *dest,
+static void RENAME(yuv2rgb565_2)(SwsContext *c, const int16_t *buf[2],
+ const int16_t *ubuf[2], const int16_t *vbuf[2],
+ const int16_t *abuf[2], uint8_t *dest,
int dstW, int yalpha, int uvalpha, int y)
{
+ const int16_t *buf0 = buf[0], *buf1 = buf[1],
+ *ubuf0 = ubuf[0], *ubuf1 = ubuf[1];
+
//Note 8280 == DSTW_OFFSET but the preprocessor can't handle that there :(
__asm__ volatile(
"mov %%"REG_b", "ESP_OFFSET"(%5) \n\t"
#define YSCALEYUV2PACKED(index, c) REAL_YSCALEYUV2PACKED(index, c)
-static void RENAME(yuv2yuyv422_2)(SwsContext *c, const uint16_t *buf0,
- const uint16_t *buf1, const uint16_t *ubuf0,
- const uint16_t *ubuf1, const uint16_t *vbuf0,
- const uint16_t *vbuf1, const uint16_t *abuf0,
- const uint16_t *abuf1, uint8_t *dest,
+static void RENAME(yuv2yuyv422_2)(SwsContext *c, const int16_t *buf[2],
+ const int16_t *ubuf[2], const int16_t *vbuf[2],
+ const int16_t *abuf[2], uint8_t *dest,
int dstW, int yalpha, int uvalpha, int y)
{
+ const int16_t *buf0 = buf[0], *buf1 = buf[1],
+ *ubuf0 = ubuf[0], *ubuf1 = ubuf[1];
+
//Note 8280 == DSTW_OFFSET but the preprocessor can't handle that there :(
__asm__ volatile(
"mov %%"REG_b", "ESP_OFFSET"(%5) \n\t"
/**
* YV12 to RGB without scaling or interpolating
*/
-static void RENAME(yuv2rgb32_1)(SwsContext *c, const uint16_t *buf0,
- const uint16_t *ubuf0, const uint16_t *ubuf1,
- const uint16_t *vbuf0, const uint16_t *vbuf1,
- const uint16_t *abuf0, uint8_t *dest,
- int dstW, int uvalpha, enum PixelFormat dstFormat,
- int flags, int y)
+static void RENAME(yuv2rgb32_1)(SwsContext *c, const int16_t *buf0,
+ const int16_t *ubuf[2], const int16_t *bguf[2],
+ const int16_t *abuf0, uint8_t *dest,
+ int dstW, int uvalpha, int y)
{
- const uint16_t *buf1= buf0; //FIXME needed for RGB1/BGR1
+ const int16_t *ubuf0 = ubuf[0], *ubuf1 = ubuf[1];
+ const int16_t *buf1= buf0; //FIXME needed for RGB1/BGR1
if (uvalpha < 2048) { // note this is not correct (shifts chrominance by 0.5 pixels) but it is a bit faster
if (CONFIG_SWSCALE_ALPHA && c->alpPixBuf) {
}
}
-static void RENAME(yuv2bgr24_1)(SwsContext *c, const uint16_t *buf0,
- const uint16_t *ubuf0, const uint16_t *ubuf1,
- const uint16_t *vbuf0, const uint16_t *vbuf1,
- const uint16_t *abuf0, uint8_t *dest,
- int dstW, int uvalpha, enum PixelFormat dstFormat,
- int flags, int y)
+static void RENAME(yuv2bgr24_1)(SwsContext *c, const int16_t *buf0,
+ const int16_t *ubuf[2], const int16_t *bguf[2],
+ const int16_t *abuf0, uint8_t *dest,
+ int dstW, int uvalpha, int y)
{
- const uint16_t *buf1= buf0; //FIXME needed for RGB1/BGR1
+ const int16_t *ubuf0 = ubuf[0], *ubuf1 = ubuf[1];
+ const int16_t *buf1= buf0; //FIXME needed for RGB1/BGR1
if (uvalpha < 2048) { // note this is not correct (shifts chrominance by 0.5 pixels) but it is a bit faster
__asm__ volatile(
}
}
-static void RENAME(yuv2rgb555_1)(SwsContext *c, const uint16_t *buf0,
- const uint16_t *ubuf0, const uint16_t *ubuf1,
- const uint16_t *vbuf0, const uint16_t *vbuf1,
- const uint16_t *abuf0, uint8_t *dest,
- int dstW, int uvalpha, enum PixelFormat dstFormat,
- int flags, int y)
+static void RENAME(yuv2rgb555_1)(SwsContext *c, const int16_t *buf0,
+ const int16_t *ubuf[2], const int16_t *bguf[2],
+ const int16_t *abuf0, uint8_t *dest,
+ int dstW, int uvalpha, int y)
{
- const uint16_t *buf1= buf0; //FIXME needed for RGB1/BGR1
+ const int16_t *ubuf0 = ubuf[0], *ubuf1 = ubuf[1];
+ const int16_t *buf1= buf0; //FIXME needed for RGB1/BGR1
if (uvalpha < 2048) { // note this is not correct (shifts chrominance by 0.5 pixels) but it is a bit faster
__asm__ volatile(
}
}
-static void RENAME(yuv2rgb565_1)(SwsContext *c, const uint16_t *buf0,
- const uint16_t *ubuf0, const uint16_t *ubuf1,
- const uint16_t *vbuf0, const uint16_t *vbuf1,
- const uint16_t *abuf0, uint8_t *dest,
- int dstW, int uvalpha, enum PixelFormat dstFormat,
- int flags, int y)
+static void RENAME(yuv2rgb565_1)(SwsContext *c, const int16_t *buf0,
+ const int16_t *ubuf[2], const int16_t *bguf[2],
+ const int16_t *abuf0, uint8_t *dest,
+ int dstW, int uvalpha, int y)
{
- const uint16_t *buf1= buf0; //FIXME needed for RGB1/BGR1
+ const int16_t *ubuf0 = ubuf[0], *ubuf1 = ubuf[1];
+ const int16_t *buf1= buf0; //FIXME needed for RGB1/BGR1
if (uvalpha < 2048) { // note this is not correct (shifts chrominance by 0.5 pixels) but it is a bit faster
__asm__ volatile(
"psraw $7, %%mm7 \n\t"
#define YSCALEYUV2PACKED1b(index, c) REAL_YSCALEYUV2PACKED1b(index, c)
-static void RENAME(yuv2yuyv422_1)(SwsContext *c, const uint16_t *buf0,
- const uint16_t *ubuf0, const uint16_t *ubuf1,
- const uint16_t *vbuf0, const uint16_t *vbuf1,
- const uint16_t *abuf0, uint8_t *dest,
- int dstW, int uvalpha, enum PixelFormat dstFormat,
- int flags, int y)
+static void RENAME(yuv2yuyv422_1)(SwsContext *c, const int16_t *buf0,
+ const int16_t *ubuf[2], const int16_t *bguf[2],
+ const int16_t *abuf0, uint8_t *dest,
+ int dstW, int uvalpha, int y)
{
- const uint16_t *buf1= buf0; //FIXME needed for RGB1/BGR1
+ const int16_t *ubuf0 = ubuf[0], *ubuf1 = ubuf[1];
+ const int16_t *buf1= buf0; //FIXME needed for RGB1/BGR1
if (uvalpha < 2048) { // note this is not correct (shifts chrominance by 0.5 pixels) but it is a bit faster
__asm__ volatile(
OSDN Git Service
