/*
- * Copyright (C) 2001-2003 Michael Niedermayer <michaelni@gmx.at>
+ * Copyright (C) 2001-2011 Michael Niedermayer <michaelni@gmx.at>
*
- * This file is part of Libav.
+ * This file is part of FFmpeg.
*
- * Libav is free software; you can redistribute it and/or
+ * FFmpeg is free software; you can redistribute it and/or
* modify it under the terms of the GNU Lesser General Public
* License as published by the Free Software Foundation; either
* version 2.1 of the License, or (at your option) any later version.
*
- * Libav is distributed in the hope that it will be useful,
+ * FFmpeg is distributed in the hope that it will be useful,
* but WITHOUT ANY WARRANTY; without even the implied warranty of
* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
* Lesser General Public License for more details.
*
* You should have received a copy of the GNU Lesser General Public
- * License along with Libav; if not, write to the Free Software
+ * License along with FFmpeg; if not, write to the Free Software
* Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
*/
#include "swscale.h"
#include "swscale_internal.h"
#include "rgb2rgb.h"
+#include "libavutil/avassert.h"
#include "libavutil/intreadwrite.h"
#include "libavutil/cpu.h"
#include "libavutil/avutil.h"
#include "libavutil/bswap.h"
#include "libavutil/pixdesc.h"
-#define DITHER1XBPP
#define RGB2YUV_SHIFT 15
#define BY ( (int)(0.114*219/255*(1<<RGB2YUV_SHIFT)+0.5))
DECLARE_ALIGNED(8, const uint8_t, ff_sws_pb_64)[8] =
{ 64, 64, 64, 64, 64, 64, 64, 64 };
+DECLARE_ALIGNED(8, const uint8_t, dithers)[8][8][8]={
+{
+ { 0, 1, 0, 1, 0, 1, 0, 1,},
+ { 1, 0, 1, 0, 1, 0, 1, 0,},
+ { 0, 1, 0, 1, 0, 1, 0, 1,},
+ { 1, 0, 1, 0, 1, 0, 1, 0,},
+ { 0, 1, 0, 1, 0, 1, 0, 1,},
+ { 1, 0, 1, 0, 1, 0, 1, 0,},
+ { 0, 1, 0, 1, 0, 1, 0, 1,},
+ { 1, 0, 1, 0, 1, 0, 1, 0,},
+},{
+ { 1, 2, 1, 2, 1, 2, 1, 2,},
+ { 3, 0, 3, 0, 3, 0, 3, 0,},
+ { 1, 2, 1, 2, 1, 2, 1, 2,},
+ { 3, 0, 3, 0, 3, 0, 3, 0,},
+ { 1, 2, 1, 2, 1, 2, 1, 2,},
+ { 3, 0, 3, 0, 3, 0, 3, 0,},
+ { 1, 2, 1, 2, 1, 2, 1, 2,},
+ { 3, 0, 3, 0, 3, 0, 3, 0,},
+},{
+ { 2, 4, 3, 5, 2, 4, 3, 5,},
+ { 6, 0, 7, 1, 6, 0, 7, 1,},
+ { 3, 5, 2, 4, 3, 5, 2, 4,},
+ { 7, 1, 6, 0, 7, 1, 6, 0,},
+ { 2, 4, 3, 5, 2, 4, 3, 5,},
+ { 6, 0, 7, 1, 6, 0, 7, 1,},
+ { 3, 5, 2, 4, 3, 5, 2, 4,},
+ { 7, 1, 6, 0, 7, 1, 6, 0,},
+},{
+ { 4, 8, 7, 11, 4, 8, 7, 11,},
+ { 12, 0, 15, 3, 12, 0, 15, 3,},
+ { 6, 10, 5, 9, 6, 10, 5, 9,},
+ { 14, 2, 13, 1, 14, 2, 13, 1,},
+ { 4, 8, 7, 11, 4, 8, 7, 11,},
+ { 12, 0, 15, 3, 12, 0, 15, 3,},
+ { 6, 10, 5, 9, 6, 10, 5, 9,},
+ { 14, 2, 13, 1, 14, 2, 13, 1,},
+},{
+ { 9, 17, 15, 23, 8, 16, 14, 22,},
+ { 25, 1, 31, 7, 24, 0, 30, 6,},
+ { 13, 21, 11, 19, 12, 20, 10, 18,},
+ { 29, 5, 27, 3, 28, 4, 26, 2,},
+ { 8, 16, 14, 22, 9, 17, 15, 23,},
+ { 24, 0, 30, 6, 25, 1, 31, 7,},
+ { 12, 20, 10, 18, 13, 21, 11, 19,},
+ { 28, 4, 26, 2, 29, 5, 27, 3,},
+},{
+ { 18, 34, 30, 46, 17, 33, 29, 45,},
+ { 50, 2, 62, 14, 49, 1, 61, 13,},
+ { 26, 42, 22, 38, 25, 41, 21, 37,},
+ { 58, 10, 54, 6, 57, 9, 53, 5,},
+ { 16, 32, 28, 44, 19, 35, 31, 47,},
+ { 48, 0, 60, 12, 51, 3, 63, 15,},
+ { 24, 40, 20, 36, 27, 43, 23, 39,},
+ { 56, 8, 52, 4, 59, 11, 55, 7,},
+},{
+ { 18, 34, 30, 46, 17, 33, 29, 45,},
+ { 50, 2, 62, 14, 49, 1, 61, 13,},
+ { 26, 42, 22, 38, 25, 41, 21, 37,},
+ { 58, 10, 54, 6, 57, 9, 53, 5,},
+ { 16, 32, 28, 44, 19, 35, 31, 47,},
+ { 48, 0, 60, 12, 51, 3, 63, 15,},
+ { 24, 40, 20, 36, 27, 43, 23, 39,},
+ { 56, 8, 52, 4, 59, 11, 55, 7,},
+},{
+ { 36, 68, 60, 92, 34, 66, 58, 90,},
+ { 100, 4,124, 28, 98, 2,122, 26,},
+ { 52, 84, 44, 76, 50, 82, 42, 74,},
+ { 116, 20,108, 12,114, 18,106, 10,},
+ { 32, 64, 56, 88, 38, 70, 62, 94,},
+ { 96, 0,120, 24,102, 6,126, 30,},
+ { 48, 80, 40, 72, 54, 86, 46, 78,},
+ { 112, 16,104, 8,118, 22,110, 14,},
+}};
+
+static const uint8_t flat64[8]={64,64,64,64,64,64,64,64};
+
+const uint16_t dither_scale[15][16]={
+{ 2, 3, 3, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5,},
+{ 2, 3, 7, 7, 13, 13, 25, 25, 25, 25, 25, 25, 25, 25, 25, 25,},
+{ 3, 3, 4, 15, 15, 29, 57, 57, 57, 113, 113, 113, 113, 113, 113, 113,},
+{ 3, 4, 4, 5, 31, 31, 61, 121, 241, 241, 241, 241, 481, 481, 481, 481,},
+{ 3, 4, 5, 5, 6, 63, 63, 125, 249, 497, 993, 993, 993, 993, 993, 1985,},
+{ 3, 5, 6, 6, 6, 7, 127, 127, 253, 505, 1009, 2017, 4033, 4033, 4033, 4033,},
+{ 3, 5, 6, 7, 7, 7, 8, 255, 255, 509, 1017, 2033, 4065, 8129,16257,16257,},
+{ 3, 5, 6, 8, 8, 8, 8, 9, 511, 511, 1021, 2041, 4081, 8161,16321,32641,},
+{ 3, 5, 7, 8, 9, 9, 9, 9, 10, 1023, 1023, 2045, 4089, 8177,16353,32705,},
+{ 3, 5, 7, 8, 10, 10, 10, 10, 10, 11, 2047, 2047, 4093, 8185,16369,32737,},
+{ 3, 5, 7, 8, 10, 11, 11, 11, 11, 11, 12, 4095, 4095, 8189,16377,32753,},
+{ 3, 5, 7, 9, 10, 12, 12, 12, 12, 12, 12, 13, 8191, 8191,16381,32761,},
+{ 3, 5, 7, 9, 10, 12, 13, 13, 13, 13, 13, 13, 14,16383,16383,32765,},
+{ 3, 5, 7, 9, 10, 12, 14, 14, 14, 14, 14, 14, 14, 15,32767,32767,},
+{ 3, 5, 7, 9, 11, 12, 14, 15, 15, 15, 15, 15, 15, 15, 16,65535,},
+};
+
#define output_pixel(pos, val, bias, signedness) \
if (big_endian) { \
AV_WB16(pos, bias + av_clip_ ## signedness ## 16(val >> shift)); \
int big_endian, int output_bits)
{
int i;
- int shift = 19 - output_bits;
+ int shift = 3;
+ av_assert0(output_bits == 16);
for (i = 0; i < dstW; i++) {
int val = src[i] + (1 << (shift - 1));
int big_endian, int output_bits)
{
int i;
- int shift = 15 + 16 - output_bits;
+ int shift = 15;
+ av_assert0(output_bits == 16);
for (i = 0; i < dstW; i++) {
- int val = 1 << (30-output_bits);
+ int val = 1 << (shift - 1);
int j;
/* range of val is [0,0x7FFFFFFF], so 31 bits, but with lanczos/spline
int shift = 11 + 16 - output_bits;
for (i = 0; i < dstW; i++) {
- int val = 1 << (26-output_bits);
+ int val = 1 << (shift - 1);
int j;
for (j = 0; j < filterSize; j++)
int i;
for (i = 0; i < (dstW >> 1); i++) {
- int Y1 = buf0[i * 2 ] << 1;
- int Y2 = buf0[i * 2 + 1] << 1;
+ int Y1 = (buf0[i * 2 ]+4)>>3;
+ int Y2 = (buf0[i * 2 + 1]+4)>>3;
output_pixel(&dest[i * 2 + 0], Y1);
output_pixel(&dest[i * 2 + 1], Y2);
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];
+ uint8_t *g = c->table_gU[128 + YUVRGB_TABLE_HEADROOM] + c->table_gV[128 + YUVRGB_TABLE_HEADROOM];
int i;
unsigned acc = 0;
{
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];
+ uint8_t *g = c->table_gU[128 + YUVRGB_TABLE_HEADROOM] + c->table_gV[128 + YUVRGB_TABLE_HEADROOM];
int yalpha1 = 4095 - yalpha;
int i;
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];
+ uint8_t *g = c->table_gU[128 + YUVRGB_TABLE_HEADROOM] + c->table_gV[128 + YUVRGB_TABLE_HEADROOM];
int i;
for (i = 0; i < dstW - 7; i += 8) {
for (i = 0; i < (dstW >> 1); i++) {
int Y1 = (buf0[i * 2] ) >> 2;
int Y2 = (buf0[i * 2 + 1]) >> 2;
- int U = (ubuf0[i] + ubuf1[i] + (-128 << 11)) >> 3;
- int V = (vbuf0[i] + vbuf1[i] + (-128 << 11)) >> 3;
+ int U = (ubuf0[i] + ubuf1[i] + (-128 << 12)) >> 3;
+ int V = (vbuf0[i] + vbuf1[i] + (-128 << 12)) >> 3;
int R, G, B;
Y1 -= c->yuv2rgb_y_offset;
* correct RGB values into the destination buffer.
*/
static av_always_inline void
-yuv2rgb_write(uint8_t *_dest, int i, unsigned Y1, unsigned Y2,
+yuv2rgb_write(uint8_t *_dest, int i, int Y1, int Y2,
unsigned A1, unsigned A2,
const void *_r, const void *_g, const void *_b, int y,
enum PixelFormat target, int hasAlpha)
#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];
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 (hasAlpha) {
A1 = 1 << 18;
A2 = 1 << 18;
}
}
- /* 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];
+ r = c->table_rV[V + YUVRGB_TABLE_HEADROOM];
+ g = (c->table_gU[U + YUVRGB_TABLE_HEADROOM] + c->table_gV[V + YUVRGB_TABLE_HEADROOM]);
+ b = c->table_bU[U + YUVRGB_TABLE_HEADROOM];
yuv2rgb_write(dest, i, Y1, Y2, hasAlpha ? A1 : 0, hasAlpha ? A2 : 0,
r, g, b, y, target, hasAlpha);
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];
+ const void *r = c->table_rV[V + YUVRGB_TABLE_HEADROOM],
+ *g = (c->table_gU[U + YUVRGB_TABLE_HEADROOM] + c->table_gV[V + YUVRGB_TABLE_HEADROOM]),
+ *b = c->table_bU[U + YUVRGB_TABLE_HEADROOM];
if (hasAlpha) {
A1 = (abuf0[i * 2 ] * yalpha1 + abuf1[i * 2 ] * yalpha) >> 19;
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];
+ const void *r = c->table_rV[V + YUVRGB_TABLE_HEADROOM],
+ *g = (c->table_gU[U + YUVRGB_TABLE_HEADROOM] + c->table_gV[V + YUVRGB_TABLE_HEADROOM]),
+ *b = c->table_bU[U + YUVRGB_TABLE_HEADROOM];
if (hasAlpha) {
A1 = abuf0[i * 2 ] >> 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];
+ const void *r = c->table_rV[V + YUVRGB_TABLE_HEADROOM],
+ *g = (c->table_gU[U + YUVRGB_TABLE_HEADROOM] + c->table_gV[V + YUVRGB_TABLE_HEADROOM]),
+ *b = c->table_bU[U + YUVRGB_TABLE_HEADROOM];
if (hasAlpha) {
A1 = abuf0[i * 2 ] >> 7;
for (i = 0; i < dstW; i++) {
int j;
- int Y = 0;
- int U = -128 << 19;
- int V = -128 << 19;
+ int Y = 1<<9;
+ int U = (1<<9)-(128 << 19);
+ int V = (1<<9)-(128 << 19);
int av_unused A;
int R, G, B;
U >>= 10;
V >>= 10;
if (hasAlpha) {
- A = 1 << 21;
+ A = 1 << 18;
for (j = 0; j < lumFilterSize; j++) {
A += alpSrc[j][i] * lumFilter[j];
}
dest[1] = B >> 22;
dest[2] = G >> 22;
dest[3] = R >> 22;
- dest += 4;
break;
case PIX_FMT_BGR24:
dest[0] = B >> 22;
#undef input_pixel
#define rgb48funcs(pattern, BE_LE, origin) \
-static void pattern ## 48 ## BE_LE ## ToY_c(uint8_t *_dst, const uint8_t *_src, \
+static void pattern ## 48 ## BE_LE ## ToY_c(uint8_t *_dst, const uint8_t *_src, const uint8_t *unused0, const uint8_t *unused1,\
int width, uint32_t *unused) \
{ \
const uint16_t *src = (const uint16_t *) _src; \
} \
\
static void pattern ## 48 ## BE_LE ## ToUV_c(uint8_t *_dstU, uint8_t *_dstV, \
- const uint8_t *_src1, const uint8_t *_src2, \
+ const uint8_t *unused0, const uint8_t *_src1, const uint8_t *_src2, \
int width, uint32_t *unused) \
{ \
const uint16_t *src1 = (const uint16_t *) _src1, \
} \
\
static void pattern ## 48 ## BE_LE ## ToUV_half_c(uint8_t *_dstU, uint8_t *_dstV, \
- const uint8_t *_src1, const uint8_t *_src2, \
+ const uint8_t *unused0, const uint8_t *_src1, const uint8_t *_src2, \
int width, uint32_t *unused) \
{ \
const uint16_t *src1 = (const uint16_t *) _src1, \
(isBE(origin) ? AV_RB16(&src[(i)*2]) : AV_RL16(&src[(i)*2])))
static av_always_inline void
-rgb16_32ToY_c_template(uint8_t *dst, const uint8_t *src,
+rgb16_32ToY_c_template(int16_t *dst, const uint8_t *src,
int width, enum PixelFormat origin,
int shr, int shg, int shb, int shp,
int maskr, int maskg, int maskb,
int rsh, int gsh, int bsh, int S)
{
const int ry = RY << rsh, gy = GY << gsh, by = BY << bsh;
- const unsigned rnd = 33u << (S - 1);
+ const unsigned rnd = (32<<((S)-1)) + (1<<(S-7));
int i;
for (i = 0; i < width; i++) {
int g = (px & maskg) >> shg;
int r = (px & maskr) >> shr;
- dst[i] = (ry * r + gy * g + by * b + rnd) >> S;
+ dst[i] = (ry * r + gy * g + by * b + rnd) >> ((S)-6);
}
}
static av_always_inline void
-rgb16_32ToUV_c_template(uint8_t *dstU, uint8_t *dstV,
+rgb16_32ToUV_c_template(int16_t *dstU, int16_t *dstV,
const uint8_t *src, int width,
enum PixelFormat origin,
int shr, int shg, int shb, int shp,
{
const int ru = RU << rsh, gu = GU << gsh, bu = BU << bsh,
rv = RV << rsh, gv = GV << gsh, bv = BV << bsh;
- const unsigned rnd = 257u << (S - 1);
+ const unsigned rnd = (256u<<((S)-1)) + (1<<(S-7));
int i;
for (i = 0; i < width; i++) {
int g = (px & maskg) >> shg;
int r = (px & maskr) >> shr;
- dstU[i] = (ru * r + gu * g + bu * b + rnd) >> S;
- dstV[i] = (rv * r + gv * g + bv * b + rnd) >> S;
+ dstU[i] = (ru * r + gu * g + bu * b + rnd) >> ((S)-6);
+ dstV[i] = (rv * r + gv * g + bv * b + rnd) >> ((S)-6);
}
}
static av_always_inline void
-rgb16_32ToUV_half_c_template(uint8_t *dstU, uint8_t *dstV,
+rgb16_32ToUV_half_c_template(int16_t *dstU, int16_t *dstV,
const uint8_t *src, int width,
enum PixelFormat origin,
int shr, int shg, int shb, int shp,
const int ru = RU << rsh, gu = GU << gsh, bu = BU << bsh,
rv = RV << rsh, gv = GV << gsh, bv = BV << bsh,
maskgx = ~(maskr | maskb);
- const unsigned rnd = 257u << S;
+ const unsigned rnd = (256U<<(S)) + (1<<(S-6));
int i;
maskr |= maskr << 1; maskb |= maskb << 1; maskg |= maskg << 1;
}
r = (rb & maskr) >> shr;
- dstU[i] = (ru * r + gu * g + bu * b + rnd) >> (S + 1);
- dstV[i] = (rv * r + gv * g + bv * b + rnd) >> (S + 1);
+ dstU[i] = (ru * r + gu * g + bu * b + (unsigned)rnd) >> ((S)-6+1);
+ dstV[i] = (rv * r + gv * g + bv * b + (unsigned)rnd) >> ((S)-6+1);
}
}
#define rgb16_32_wrapper(fmt, name, shr, shg, shb, shp, maskr, \
maskg, maskb, rsh, gsh, bsh, S) \
-static void name ## ToY_c(uint8_t *dst, const uint8_t *src, \
+static void name ## ToY_c(uint8_t *dst, const uint8_t *src, const uint8_t *unused1, const uint8_t *unused2, \
int width, uint32_t *unused) \
{ \
- rgb16_32ToY_c_template(dst, src, width, fmt, shr, shg, shb, shp, \
+ rgb16_32ToY_c_template((int16_t*)dst, src, width, fmt, \
+ shr, shg, shb, shp, \
maskr, maskg, maskb, rsh, gsh, bsh, S); \
} \
\
static void name ## ToUV_c(uint8_t *dstU, uint8_t *dstV, \
- const uint8_t *src, const uint8_t *dummy, \
+ const uint8_t *unused0, const uint8_t *src, const uint8_t *dummy, \
int width, uint32_t *unused) \
{ \
- rgb16_32ToUV_c_template(dstU, dstV, src, width, fmt, shr, shg, shb, shp, \
+ rgb16_32ToUV_c_template((int16_t*)dstU, (int16_t*)dstV, src, width, fmt, \
+ shr, shg, shb, shp, \
maskr, maskg, maskb, rsh, gsh, bsh, S); \
} \
\
static void name ## ToUV_half_c(uint8_t *dstU, uint8_t *dstV, \
- const uint8_t *src, const uint8_t *dummy, \
+ const uint8_t *unused0, const uint8_t *src, const uint8_t *dummy, \
int width, uint32_t *unused) \
{ \
- rgb16_32ToUV_half_c_template(dstU, dstV, src, width, fmt, shr, shg, shb, shp, \
+ rgb16_32ToUV_half_c_template((int16_t*)dstU, (int16_t*)dstV, src, width, fmt, \
+ shr, shg, shb, shp, \
maskr, maskg, maskb, rsh, gsh, bsh, S); \
}
rgb16_32_wrapper(PIX_FMT_RGB555BE, rgb15be, 0, 0, 0, 0, 0x7C00, 0x03E0, 0x001F, 0, 5, 10, RGB2YUV_SHIFT+7)
rgb16_32_wrapper(PIX_FMT_RGB444BE, rgb12be, 0, 0, 0, 0, 0x0F00, 0x00F0, 0x000F, 0, 4, 8, RGB2YUV_SHIFT+4)
-static void abgrToA_c(uint8_t *dst, const uint8_t *src, int width, uint32_t *unused)
+static void gbr24pToUV_half_c(uint16_t *dstU, uint16_t *dstV,
+ const uint8_t *gsrc, const uint8_t *bsrc, const uint8_t *rsrc,
+ int width, enum PixelFormat origin)
+{
+ int i;
+ for (i = 0; i < width; i++) {
+ unsigned int g = gsrc[2*i] + gsrc[2*i+1];
+ unsigned int b = bsrc[2*i] + bsrc[2*i+1];
+ unsigned int r = rsrc[2*i] + rsrc[2*i+1];
+
+ dstU[i] = (RU*r + GU*g + BU*b + (0x4001<<(RGB2YUV_SHIFT-6))) >> (RGB2YUV_SHIFT-6+1);
+ dstV[i] = (RV*r + GV*g + BV*b + (0x4001<<(RGB2YUV_SHIFT-6))) >> (RGB2YUV_SHIFT-6+1);
+ }
+}
+
+static void abgrToA_c(int16_t *dst, const uint8_t *src, const uint8_t *unused1, const uint8_t *unused2, int width, uint32_t *unused)
+{
+ int i;
+ for (i=0; i<width; i++) {
+ dst[i]= src[4*i]<<6;
+ }
+}
+
+static void rgbaToA_c(int16_t *dst, const uint8_t *src, const uint8_t *unused1, const uint8_t *unused2, int width, uint32_t *unused)
{
int i;
for (i=0; i<width; i++) {
- dst[i]= src[4*i];
+ dst[i]= src[4*i+3]<<6;
}
}
-static void rgbaToA_c(uint8_t *dst, const uint8_t *src, int width, uint32_t *unused)
+static void palToA_c(int16_t *dst, const uint8_t *src, const uint8_t *unused1, const uint8_t *unused2, int width, uint32_t *pal)
{
int i;
for (i=0; i<width; i++) {
- dst[i]= src[4*i+3];
+ int d= src[i];
+
+ dst[i]= (pal[d] >> 24)<<6;
}
}
-static void palToY_c(uint8_t *dst, const uint8_t *src, int width, uint32_t *pal)
+static void palToY_c(int16_t *dst, const uint8_t *src, const uint8_t *unused1, const uint8_t *unused2, long width, uint32_t *pal)
{
int i;
for (i=0; i<width; i++) {
int d= src[i];
- dst[i]= pal[d] & 0xFF;
+ dst[i]= (pal[d] & 0xFF)<<6;
}
}
-static void palToUV_c(uint8_t *dstU, uint8_t *dstV,
- const uint8_t *src1, const uint8_t *src2,
- int width, uint32_t *pal)
+static void palToUV_c(uint16_t *dstU, int16_t *dstV,
+ const uint8_t *unused0, const uint8_t *src1, const uint8_t *src2,
+ int width, uint32_t *pal)
{
int i;
assert(src1 == src2);
for (i=0; i<width; i++) {
int p= pal[src1[i]];
- dstU[i]= p>>8;
- dstV[i]= p>>16;
+ dstU[i]= (uint8_t)(p>> 8)<<6;
+ dstV[i]= (uint8_t)(p>>16)<<6;
}
}
-static void monowhite2Y_c(uint8_t *dst, const uint8_t *src,
- int width, uint32_t *unused)
+static void monowhite2Y_c(int16_t *dst, const uint8_t *src, const uint8_t *unused1, const uint8_t *unused2, int width, uint32_t *unused)
{
int i, j;
for (i=0; i<width/8; i++) {
int d= ~src[i];
for(j=0; j<8; j++)
- dst[8*i+j]= ((d>>(7-j))&1)*255;
+ dst[8*i+j]= ((d>>(7-j))&1)*16383;
+ }
+ if(width&7){
+ int d= ~src[i];
+ for(j=0; j<(width&7); j++)
+ dst[8*i+j]= ((d>>(7-j))&1)*16383;
}
}
-static void monoblack2Y_c(uint8_t *dst, const uint8_t *src,
- int width, uint32_t *unused)
+static void monoblack2Y_c(int16_t *dst, const uint8_t *src, const uint8_t *unused1, const uint8_t *unused2, int width, uint32_t *unused)
{
int i, j;
for (i=0; i<width/8; i++) {
int d= src[i];
for(j=0; j<8; j++)
- dst[8*i+j]= ((d>>(7-j))&1)*255;
+ dst[8*i+j]= ((d>>(7-j))&1)*16383;
+ }
+ if(width&7){
+ int d= src[i];
+ for(j=0; j<(width&7); j++)
+ dst[8*i+j]= ((d>>(7-j))&1)*16383;
}
}
-static void yuy2ToY_c(uint8_t *dst, const uint8_t *src, int width,
+static void yuy2ToY_c(uint8_t *dst, const uint8_t *src, const uint8_t *unused1, const uint8_t *unused2, int width,
uint32_t *unused)
{
int i;
dst[i]= src[2*i];
}
-static void yuy2ToUV_c(uint8_t *dstU, uint8_t *dstV, const uint8_t *src1,
+static void yuy2ToUV_c(uint8_t *dstU, uint8_t *dstV, const uint8_t *unused0, const uint8_t *src1,
const uint8_t *src2, int width, uint32_t *unused)
{
int i;
assert(src1 == src2);
}
-static void bswap16Y_c(uint8_t *_dst, const uint8_t *_src, int width, uint32_t *unused)
+static void bswap16Y_c(uint8_t *_dst, const uint8_t *_src, const uint8_t *unused1, const uint8_t *unused2, int width, uint32_t *unused)
{
int i;
const uint16_t *src = (const uint16_t *) _src;
}
}
-static void bswap16UV_c(uint8_t *_dstU, uint8_t *_dstV, const uint8_t *_src1,
+static void bswap16UV_c(uint8_t *_dstU, uint8_t *_dstV, const uint8_t *unused0, const uint8_t *_src1,
const uint8_t *_src2, int width, uint32_t *unused)
{
int i;
/* This is almost identical to the previous, end exists only because
* yuy2ToY/UV)(dst, src+1, ...) would have 100% unaligned accesses. */
-static void uyvyToY_c(uint8_t *dst, const uint8_t *src, int width,
+static void uyvyToY_c(uint8_t *dst, const uint8_t *src, const uint8_t *unused1, const uint8_t *unused2, int width,
uint32_t *unused)
{
int i;
dst[i]= src[2*i+1];
}
-static void uyvyToUV_c(uint8_t *dstU, uint8_t *dstV, const uint8_t *src1,
+static void uyvyToUV_c(uint8_t *dstU, uint8_t *dstV, const uint8_t *unused0, const uint8_t *src1,
const uint8_t *src2, int width, uint32_t *unused)
{
int i;
}
static void nv12ToUV_c(uint8_t *dstU, uint8_t *dstV,
- const uint8_t *src1, const uint8_t *src2,
+ const uint8_t *unused0, const uint8_t *src1, const uint8_t *src2,
int width, uint32_t *unused)
{
nvXXtoUV_c(dstU, dstV, src1, width);
}
static void nv21ToUV_c(uint8_t *dstU, uint8_t *dstV,
- const uint8_t *src1, const uint8_t *src2,
+ const uint8_t *unused0, const uint8_t *src1, const uint8_t *src2,
int width, uint32_t *unused)
{
nvXXtoUV_c(dstV, dstU, src1, width);
#define input_pixel(pos) (isBE(origin) ? AV_RB16(pos) : AV_RL16(pos))
-static void bgr24ToY_c(uint8_t *dst, const uint8_t *src,
+static void bgr24ToY_c(int16_t *dst, const uint8_t *src, const uint8_t *unused1, const uint8_t *unused2,
int width, uint32_t *unused)
{
int i;
int g= src[i*3+1];
int r= src[i*3+2];
- dst[i]= ((RY*r + GY*g + BY*b + (33<<(RGB2YUV_SHIFT-1)))>>RGB2YUV_SHIFT);
+ dst[i]= ((RY*r + GY*g + BY*b + (32<<(RGB2YUV_SHIFT-1)) + (1<<(RGB2YUV_SHIFT-7)))>>(RGB2YUV_SHIFT-6));
}
}
-static void bgr24ToUV_c(uint8_t *dstU, uint8_t *dstV, const uint8_t *src1,
+static void bgr24ToUV_c(int16_t *dstU, int16_t *dstV, const uint8_t *unused0, const uint8_t *src1,
const uint8_t *src2, int width, uint32_t *unused)
{
int i;
int g= src1[3*i + 1];
int r= src1[3*i + 2];
- dstU[i]= (RU*r + GU*g + BU*b + (257<<(RGB2YUV_SHIFT-1)))>>RGB2YUV_SHIFT;
- dstV[i]= (RV*r + GV*g + BV*b + (257<<(RGB2YUV_SHIFT-1)))>>RGB2YUV_SHIFT;
+ dstU[i]= (RU*r + GU*g + BU*b + (256<<(RGB2YUV_SHIFT-1)) + (1<<(RGB2YUV_SHIFT-7)))>>(RGB2YUV_SHIFT-6);
+ dstV[i]= (RV*r + GV*g + BV*b + (256<<(RGB2YUV_SHIFT-1)) + (1<<(RGB2YUV_SHIFT-7)))>>(RGB2YUV_SHIFT-6);
}
assert(src1 == src2);
}
-static void bgr24ToUV_half_c(uint8_t *dstU, uint8_t *dstV, const uint8_t *src1,
+static void bgr24ToUV_half_c(int16_t *dstU, int16_t *dstV, const uint8_t *unused0, const uint8_t *src1,
const uint8_t *src2, int width, uint32_t *unused)
{
int i;
int g= src1[6*i + 1] + src1[6*i + 4];
int r= src1[6*i + 2] + src1[6*i + 5];
- dstU[i]= (RU*r + GU*g + BU*b + (257<<RGB2YUV_SHIFT))>>(RGB2YUV_SHIFT+1);
- dstV[i]= (RV*r + GV*g + BV*b + (257<<RGB2YUV_SHIFT))>>(RGB2YUV_SHIFT+1);
+ dstU[i]= (RU*r + GU*g + BU*b + (256<<RGB2YUV_SHIFT) + (1<<(RGB2YUV_SHIFT-6)))>>(RGB2YUV_SHIFT-5);
+ dstV[i]= (RV*r + GV*g + BV*b + (256<<RGB2YUV_SHIFT) + (1<<(RGB2YUV_SHIFT-6)))>>(RGB2YUV_SHIFT-5);
}
assert(src1 == src2);
}
-static void rgb24ToY_c(uint8_t *dst, const uint8_t *src, int width,
+static void rgb24ToY_c(int16_t *dst, const uint8_t *src, const uint8_t *unused1, const uint8_t *unused2, int width,
uint32_t *unused)
{
int i;
int g= src[i*3+1];
int b= src[i*3+2];
- dst[i]= ((RY*r + GY*g + BY*b + (33<<(RGB2YUV_SHIFT-1)))>>RGB2YUV_SHIFT);
+ dst[i]= ((RY*r + GY*g + BY*b + (32<<(RGB2YUV_SHIFT-1)) + (1<<(RGB2YUV_SHIFT-7)))>>(RGB2YUV_SHIFT-6));
}
}
-static void rgb24ToUV_c(uint8_t *dstU, uint8_t *dstV, const uint8_t *src1,
+static void rgb24ToUV_c(int16_t *dstU, int16_t *dstV, const uint8_t *unused0, const uint8_t *src1,
const uint8_t *src2, int width, uint32_t *unused)
{
int i;
int g= src1[3*i + 1];
int b= src1[3*i + 2];
- dstU[i]= (RU*r + GU*g + BU*b + (257<<(RGB2YUV_SHIFT-1)))>>RGB2YUV_SHIFT;
- dstV[i]= (RV*r + GV*g + BV*b + (257<<(RGB2YUV_SHIFT-1)))>>RGB2YUV_SHIFT;
+ dstU[i]= (RU*r + GU*g + BU*b + (256<<(RGB2YUV_SHIFT-1)) + (1<<(RGB2YUV_SHIFT-7)))>>(RGB2YUV_SHIFT-6);
+ dstV[i]= (RV*r + GV*g + BV*b + (256<<(RGB2YUV_SHIFT-1)) + (1<<(RGB2YUV_SHIFT-7)))>>(RGB2YUV_SHIFT-6);
}
}
-static void rgb24ToUV_half_c(uint8_t *dstU, uint8_t *dstV, const uint8_t *src1,
- const uint8_t *src2, int width, uint32_t *unused)
+static void rgb24ToUV_half_c(int16_t *dstU, int16_t *dstV, const uint8_t *unused0, const uint8_t *src1,
+ const uint8_t *src2, int width, uint32_t *unused)
{
int i;
assert(src1==src2);
int g= src1[6*i + 1] + src1[6*i + 4];
int b= src1[6*i + 2] + src1[6*i + 5];
- dstU[i]= (RU*r + GU*g + BU*b + (257<<RGB2YUV_SHIFT))>>(RGB2YUV_SHIFT+1);
- dstV[i]= (RV*r + GV*g + BV*b + (257<<RGB2YUV_SHIFT))>>(RGB2YUV_SHIFT+1);
+ dstU[i]= (RU*r + GU*g + BU*b + (256<<RGB2YUV_SHIFT) + (1<<(RGB2YUV_SHIFT-6)))>>(RGB2YUV_SHIFT-5);
+ dstV[i]= (RV*r + GV*g + BV*b + (256<<RGB2YUV_SHIFT) + (1<<(RGB2YUV_SHIFT-6)))>>(RGB2YUV_SHIFT-5);
}
}
-static void planar_rgb_to_y(uint8_t *dst, const uint8_t *src[4], int width)
+static void planar_rgb_to_y(uint16_t *dst, const uint8_t *src[4], int width)
{
int i;
for (i = 0; i < width; i++) {
int b = src[1][i];
int r = src[2][i];
- dst[i] = ((RY * r + GY * g + BY * b + (33 << (RGB2YUV_SHIFT - 1))) >> RGB2YUV_SHIFT);
+ dst[i] = (RY*r + GY*g + BY*b + (0x801<<(RGB2YUV_SHIFT-7))) >> (RGB2YUV_SHIFT-6);
}
}
}
}
-static void planar_rgb_to_uv(uint8_t *dstU, uint8_t *dstV, const uint8_t *src[4], int width)
+static void planar_rgb_to_uv(uint16_t *dstU, uint16_t *dstV, const uint8_t *src[4], int width)
{
int i;
for (i = 0; i < width; i++) {
int b = src[1][i];
int r = src[2][i];
- dstU[i] = (RU * r + GU * g + BU * b + (257 << RGB2YUV_SHIFT)) >> (RGB2YUV_SHIFT + 1);
- dstV[i] = (RV * r + GV * g + BV * b + (257 << RGB2YUV_SHIFT)) >> (RGB2YUV_SHIFT + 1);
+ dstU[i] = (RU*r + GU*g + BU*b + (0x4001<<(RGB2YUV_SHIFT-7))) >> (RGB2YUV_SHIFT-6);
+ dstV[i] = (RV*r + GV*g + BV*b + (0x4001<<(RGB2YUV_SHIFT-7))) >> (RGB2YUV_SHIFT-6);
}
}
int bits = av_pix_fmt_descriptors[c->srcFormat].comp[0].depth_minus1;
int sh = bits - 4;
+ if((isAnyRGB(c->srcFormat) || c->srcFormat==PIX_FMT_PAL8) && av_pix_fmt_descriptors[c->srcFormat].comp[0].depth_minus1<15)
+ sh= 9;
+
for (i = 0; i < dstW; i++) {
int j;
int srcPos = filterPos[i];
const uint16_t *src = (const uint16_t *) _src;
int sh = av_pix_fmt_descriptors[c->srcFormat].comp[0].depth_minus1;
+ if(sh<15)
+ sh= isAnyRGB(c->srcFormat) || c->srcFormat==PIX_FMT_PAL8 ? 13 : av_pix_fmt_descriptors[c->srcFormat].comp[0].depth_minus1;
+
for (i = 0; i < dstW; i++) {
int j;
int srcPos = filterPos[i];
int i;
int32_t *dst = (int32_t *) _dst;
for (i = 0; i < width; i++)
- dst[i] = (dst[i]*14071 + (33561947<<4))>>14;
+ dst[i] = (dst[i]*(14071/4) + (33561947<<4)/4)>>12;
}
static void hyscale_fast_c(SwsContext *c, int16_t *dst, int dstWidth,
dst[i]= (src[xx]<<7) + (src[xx+1] - src[xx])*xalpha;
xpos+=xInc;
}
+ for (i=dstWidth-1; (i*xInc)>>16 >=srcW-1; i--)
+ dst[i] = src[srcW-1]*128;
}
// *** horizontal scale Y line to temp buffer
uint8_t *formatConvBuffer,
uint32_t *pal, int isAlpha)
{
- void (*toYV12)(uint8_t *, const uint8_t *, int, uint32_t *) = isAlpha ? c->alpToYV12 : c->lumToYV12;
+ void (*toYV12)(uint8_t *, const uint8_t *, const uint8_t *, const uint8_t *, int, uint32_t *) = isAlpha ? c->alpToYV12 : c->lumToYV12;
void (*convertRange)(int16_t *, int) = isAlpha ? NULL : c->lumConvertRange;
const uint8_t *src = src_in[isAlpha ? 3 : 0];
if (toYV12) {
- toYV12(formatConvBuffer, src, srcW, pal);
+ toYV12(formatConvBuffer, src, src_in[1], src_in[2], srcW, pal);
src= formatConvBuffer;
} else if (c->readLumPlanar && !isAlpha) {
c->readLumPlanar(formatConvBuffer, src_in, srcW);
dst2[i]=(src2[xx]*(xalpha^127)+src2[xx+1]*xalpha);
xpos+=xInc;
}
+ for (i=dstWidth-1; (i*xInc)>>16 >=srcW-1; i--) {
+ dst1[i] = src1[srcW-1]*128;
+ dst2[i] = src2[srcW-1]*128;
+ }
}
static av_always_inline void hcscale(SwsContext *c, int16_t *dst1, int16_t *dst2, int dstWidth,
{
const uint8_t *src1 = src_in[1], *src2 = src_in[2];
if (c->chrToYV12) {
- uint8_t *buf2 = formatConvBuffer + FFALIGN(srcW * FFALIGN(c->srcBpc, 8) >> 3, 16);
- c->chrToYV12(formatConvBuffer, buf2, src1, src2, srcW, pal);
+ uint8_t *buf2 = formatConvBuffer + FFALIGN(srcW*2+78, 16);
+ c->chrToYV12(formatConvBuffer, buf2, src_in[0], src1, src2, srcW, pal);
src1= formatConvBuffer;
src2= buf2;
} else if (c->readChrPlanar) {
- uint8_t *buf2 = formatConvBuffer + FFALIGN(srcW * FFALIGN(c->srcBpc, 8) >> 3, 16);
+ uint8_t *buf2 = formatConvBuffer + FFALIGN(srcW*2+78, 16);
c->readChrPlanar(formatConvBuffer, buf2, src_in, srcW);
src1= formatConvBuffer;
src2= buf2;
*yuv2packedX = yuv2bgr24_full_X_c;
break;
}
+ if(!*yuv2packedX)
+ goto YUV_PACKED;
} else {
+ YUV_PACKED:
switch (dstFormat) {
case PIX_FMT_RGB48LE:
*yuv2packed1 = yuv2rgb48le_1_c;
int16_t *vChrFilterPos= c->vChrFilterPos;
int16_t *hLumFilterPos= c->hLumFilterPos;
int16_t *hChrFilterPos= c->hChrFilterPos;
- int16_t *vLumFilter= c->vLumFilter;
- int16_t *vChrFilter= c->vChrFilter;
int16_t *hLumFilter= c->hLumFilter;
int16_t *hChrFilter= c->hChrFilter;
int32_t *lumMmxFilter= c->lumMmxFilter;
const int chrSrcSliceH= -((-srcSliceH) >> c->chrSrcVSubSample);
int lastDstY;
uint32_t *pal=c->pal_yuv;
+ int should_dither= isNBPS(c->srcFormat) || is16BPS(c->srcFormat);
+
yuv2planar1_fn yuv2plane1 = c->yuv2plane1;
yuv2planarX_fn yuv2planeX = c->yuv2planeX;
yuv2interleavedX_fn yuv2nv12cX = c->yuv2nv12cX;
yuv2packed1_fn yuv2packed1 = c->yuv2packed1;
yuv2packed2_fn yuv2packed2 = c->yuv2packed2;
yuv2packedX_fn yuv2packedX = c->yuv2packedX;
- int should_dither = is9_OR_10BPS(c->srcFormat) || is16BPS(c->srcFormat);
/* vars which will change and which we need to store back in the context */
int dstY= c->dstY;
DEBUG_BUFFERS("vLumFilterSize: %d vLumBufSize: %d vChrFilterSize: %d vChrBufSize: %d\n",
vLumFilterSize, vLumBufSize, vChrFilterSize, vChrBufSize);
- if (dstStride[0]%8 !=0 || dstStride[1]%8 !=0 || dstStride[2]%8 !=0 || dstStride[3]%8 != 0) {
+ if (dstStride[0]%16 !=0 || dstStride[1]%16 !=0 || dstStride[2]%16 !=0 || dstStride[3]%16 != 0) {
static int warnedAlready=0; //FIXME move this into the context perhaps
if (flags & SWS_PRINT_INFO && !warnedAlready) {
av_log(c, AV_LOG_WARNING, "Warning: dstStride is not aligned!\n"
}
}
+ if ((int)dst[0]%16 || (int)dst[1]%16 || (int)dst[2]%16 || (int)src[0]%16 || (int)src[1]%16 || (int)src[2]%16
+ || dstStride[0]%16 || dstStride[1]%16 || dstStride[2]%16 || dstStride[3]%16
+ || srcStride[0]%16 || srcStride[1]%16 || srcStride[2]%16 || srcStride[3]%16
+ ) {
+ static int warnedAlready=0;
+ int cpu_flags = av_get_cpu_flags();
+ if (HAVE_MMX2 && (cpu_flags & AV_CPU_FLAG_SSE2) && !warnedAlready){
+ av_log(c, AV_LOG_WARNING, "Warning: data is not aligned! This can lead to a speedloss\n");
+ warnedAlready=1;
+ }
+ }
+
/* Note the user might start scaling the picture in the middle so this
will not get executed. This is not really intended but works
currently, so people might do it. */
dst[2] + dstStride[2] * chrDstY,
(CONFIG_SWSCALE_ALPHA && alpPixBuf) ? dst[3] + dstStride[3] * dstY : NULL,
};
+ int use_mmx_vfilter= c->use_mmx_vfilter;
const int firstLumSrcY= vLumFilterPos[dstY]; //First line needed as input
const int firstLumSrcY2= vLumFilterPos[FFMIN(dstY | ((1<<c->chrDstVSubSample) - 1), dstH-1)];
const int firstChrSrcY= vChrFilterPos[chrDstY]; //First line needed as input
-
- // Last line needed as input
- int lastLumSrcY = FFMIN(c->srcH, firstLumSrcY + vLumFilterSize) - 1;
- int lastLumSrcY2 = FFMIN(c->srcH, firstLumSrcY2 + vLumFilterSize) - 1;
- int lastChrSrcY = FFMIN(c->chrSrcH, firstChrSrcY + vChrFilterSize) - 1;
+ int lastLumSrcY= firstLumSrcY + vLumFilterSize -1; // Last line needed as input
+ int lastLumSrcY2=firstLumSrcY2+ vLumFilterSize -1; // Last line needed as input
+ int lastChrSrcY= firstChrSrcY + vChrFilterSize -1; // Last line needed as input
int enough_lines;
//handle holes (FAST_BILINEAR & weird filters)
// hmm looks like we can't use MMX here without overwriting this array's tail
find_c_packed_planar_out_funcs(c, &yuv2plane1, &yuv2planeX, &yuv2nv12cX,
&yuv2packed1, &yuv2packed2, &yuv2packedX);
+ use_mmx_vfilter= 0;
}
{
- const int16_t **lumSrcPtr= (const int16_t **) lumPixBuf + lumBufIndex + firstLumSrcY - lastInLumBuf + vLumBufSize;
- const int16_t **chrUSrcPtr= (const int16_t **) chrUPixBuf + chrBufIndex + firstChrSrcY - lastInChrBuf + vChrBufSize;
- const int16_t **chrVSrcPtr= (const int16_t **) chrVPixBuf + chrBufIndex + firstChrSrcY - lastInChrBuf + vChrBufSize;
- const int16_t **alpSrcPtr= (CONFIG_SWSCALE_ALPHA && alpPixBuf) ? (const int16_t **) alpPixBuf + lumBufIndex + firstLumSrcY - lastInLumBuf + vLumBufSize : NULL;
-
- if (firstLumSrcY < 0 || firstLumSrcY + vLumFilterSize > c->srcH) {
- const int16_t **tmpY = (const int16_t **) lumPixBuf + 2 * vLumBufSize;
- int neg = -firstLumSrcY, i, end = FFMIN(c->srcH - firstLumSrcY, vLumFilterSize);
- for (i = 0; i < neg; i++)
- tmpY[i] = lumSrcPtr[neg];
- for ( ; i < end; i++)
- tmpY[i] = lumSrcPtr[i];
- for ( ; i < vLumFilterSize; i++)
- tmpY[i] = tmpY[i-1];
- lumSrcPtr = tmpY;
-
- if (alpSrcPtr) {
- const int16_t **tmpA = (const int16_t **) alpPixBuf + 2 * vLumBufSize;
- for (i = 0; i < neg; i++)
- tmpA[i] = alpSrcPtr[neg];
- for ( ; i < end; i++)
- tmpA[i] = alpSrcPtr[i];
- for ( ; i < vLumFilterSize; i++)
- tmpA[i] = tmpA[i - 1];
- alpSrcPtr = tmpA;
- }
- }
- if (firstChrSrcY < 0 || firstChrSrcY + vChrFilterSize > c->chrSrcH) {
- const int16_t **tmpU = (const int16_t **) chrUPixBuf + 2 * vChrBufSize,
- **tmpV = (const int16_t **) chrVPixBuf + 2 * vChrBufSize;
- int neg = -firstChrSrcY, i, end = FFMIN(c->chrSrcH - firstChrSrcY, vChrFilterSize);
- for (i = 0; i < neg; i++) {
- tmpU[i] = chrUSrcPtr[neg];
- tmpV[i] = chrVSrcPtr[neg];
- }
- for ( ; i < end; i++) {
- tmpU[i] = chrUSrcPtr[i];
- tmpV[i] = chrVSrcPtr[i];
- }
- for ( ; i < vChrFilterSize; i++) {
- tmpU[i] = tmpU[i - 1];
- tmpV[i] = tmpV[i - 1];
- }
- chrUSrcPtr = tmpU;
- chrVSrcPtr = tmpV;
- }
+ const int16_t **lumSrcPtr= (const int16_t **)(void*) lumPixBuf + lumBufIndex + firstLumSrcY - lastInLumBuf + vLumBufSize;
+ const int16_t **chrUSrcPtr= (const int16_t **)(void*) chrUPixBuf + chrBufIndex + firstChrSrcY - lastInChrBuf + vChrBufSize;
+ const int16_t **chrVSrcPtr= (const int16_t **)(void*) chrVPixBuf + chrBufIndex + firstChrSrcY - lastInChrBuf + vChrBufSize;
+ const int16_t **alpSrcPtr= (CONFIG_SWSCALE_ALPHA && alpPixBuf) ? (const int16_t **)(void*) alpPixBuf + lumBufIndex + firstLumSrcY - lastInLumBuf + vLumBufSize : NULL;
+ int16_t *vLumFilter= c->vLumFilter;
+ int16_t *vChrFilter= c->vChrFilter;
if (isPlanarYUV(dstFormat) || dstFormat==PIX_FMT_GRAY8) { //YV12 like
const int chrSkipMask= (1<<c->chrDstVSubSample)-1;
+ vLumFilter += dstY * vLumFilterSize;
+ vChrFilter += chrDstY * vChrFilterSize;
+
+ av_assert0(use_mmx_vfilter != (
+ yuv2planeX == yuv2planeX_10BE_c
+ || yuv2planeX == yuv2planeX_10LE_c
+ || yuv2planeX == yuv2planeX_9BE_c
+ || yuv2planeX == yuv2planeX_9LE_c
+ || yuv2planeX == yuv2planeX_16BE_c
+ || yuv2planeX == yuv2planeX_16LE_c
+ || yuv2planeX == yuv2planeX_8_c) || !ARCH_X86);
+
+ if(use_mmx_vfilter){
+ vLumFilter= c->lumMmxFilter;
+ vChrFilter= c->chrMmxFilter;
+ }
+
if (vLumFilterSize == 1) {
yuv2plane1(lumSrcPtr[0], dest[0], dstW, c->lumDither8, 0);
} else {
- yuv2planeX(vLumFilter + dstY * vLumFilterSize, vLumFilterSize,
+ yuv2planeX(vLumFilter, vLumFilterSize,
lumSrcPtr, dest[0], dstW, c->lumDither8, 0);
}
if (!((dstY&chrSkipMask) || isGray(dstFormat))) {
if (yuv2nv12cX) {
- yuv2nv12cX(c, vChrFilter + chrDstY * vChrFilterSize, vChrFilterSize, chrUSrcPtr, chrVSrcPtr, dest[1], chrDstW);
+ yuv2nv12cX(c, vChrFilter, vChrFilterSize, chrUSrcPtr, chrVSrcPtr, dest[1], chrDstW);
} else if (vChrFilterSize == 1) {
yuv2plane1(chrUSrcPtr[0], dest[1], chrDstW, c->chrDither8, 0);
yuv2plane1(chrVSrcPtr[0], dest[2], chrDstW, c->chrDither8, 3);
} else {
- yuv2planeX(vChrFilter + chrDstY * vChrFilterSize, vChrFilterSize,
+ yuv2planeX(vChrFilter, vChrFilterSize,
chrUSrcPtr, dest[1], chrDstW, c->chrDither8, 0);
- yuv2planeX(vChrFilter + chrDstY * vChrFilterSize, vChrFilterSize,
- chrVSrcPtr, dest[2], chrDstW, c->chrDither8, 3);
+ yuv2planeX(vChrFilter, vChrFilterSize,
+ chrVSrcPtr, dest[2], chrDstW, c->chrDither8, use_mmx_vfilter ? (c->uv_offx2 >> 1) : 3);
}
}
if (CONFIG_SWSCALE_ALPHA && alpPixBuf){
+ if(use_mmx_vfilter){
+ vLumFilter= c->alpMmxFilter;
+ }
if (vLumFilterSize == 1) {
yuv2plane1(alpSrcPtr[0], dest[3], dstW, c->lumDither8, 0);
} else {
- yuv2planeX(vLumFilter + dstY * vLumFilterSize, vLumFilterSize,
+ yuv2planeX(vLumFilter, vLumFilterSize,
alpSrcPtr, dest[3], dstW, c->lumDither8, 0);
}
}
case PIX_FMT_YUV422P9LE:
case PIX_FMT_YUV420P9LE:
case PIX_FMT_YUV422P10LE:
- case PIX_FMT_YUV444P10LE:
case PIX_FMT_YUV420P10LE:
+ case PIX_FMT_YUV444P10LE:
case PIX_FMT_YUV420P16LE:
case PIX_FMT_YUV422P16LE:
case PIX_FMT_YUV444P16LE: c->chrToYV12 = bswap16UV_c; break;
case PIX_FMT_RGB565BE: c->chrToYV12 = rgb16beToUV_half_c; break;
case PIX_FMT_RGB555LE: c->chrToYV12 = rgb15leToUV_half_c; break;
case PIX_FMT_RGB555BE: c->chrToYV12 = rgb15beToUV_half_c; break;
+ case PIX_FMT_GBR24P : c->chrToYV12 = gbr24pToUV_half_c; break;
case PIX_FMT_RGB444LE: c->chrToYV12 = rgb12leToUV_half_c; break;
case PIX_FMT_RGB444BE: c->chrToYV12 = rgb12beToUV_half_c; break;
}
case PIX_FMT_YUV444P9LE:
case PIX_FMT_YUV422P9LE:
case PIX_FMT_YUV420P9LE:
- case PIX_FMT_YUV444P10LE:
case PIX_FMT_YUV422P10LE:
case PIX_FMT_YUV420P10LE:
+ case PIX_FMT_YUV444P10LE:
case PIX_FMT_YUV420P16LE:
case PIX_FMT_YUV422P16LE:
case PIX_FMT_YUV444P16LE:
case PIX_FMT_ABGR:
case PIX_FMT_ARGB: c->alpToYV12 = abgrToA_c; break;
case PIX_FMT_Y400A: c->alpToYV12 = uyvyToY_c; break;
+ case PIX_FMT_PAL8 : c->alpToYV12 = palToA_c; break;
}
}
+
if (c->srcBpc == 8) {
if (c->dstBpc <= 10) {
c->hyScale = c->hcScale = hScale8To15_c;
projects / ffmpeg / blobdiff