diff options
Diffstat (limited to 'libswscale/swscale.c')
-rw-r--r-- | libswscale/swscale.c | 665 |
1 files changed, 491 insertions, 174 deletions
diff --git a/libswscale/swscale.c b/libswscale/swscale.c index f6427bf..dff44dd 100644 --- a/libswscale/swscale.c +++ b/libswscale/swscale.c @@ -1,32 +1,33 @@ /* - * 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 <assert.h> #include <inttypes.h> #include <math.h> #include <stdio.h> #include <string.h> +#include "libavutil/avassert.h" #include "libavutil/avutil.h" #include "libavutil/bswap.h" #include "libavutil/cpu.h" +#include "libavutil/imgutils.h" #include "libavutil/intreadwrite.h" #include "libavutil/mathematics.h" #include "libavutil/pixdesc.h" @@ -35,7 +36,7 @@ #include "swscale_internal.h" #include "swscale.h" -DECLARE_ALIGNED(8, const uint8_t, ff_dither_8x8_128)[8][8] = { +DECLARE_ALIGNED(8, const uint8_t, ff_dither_8x8_128)[9][8] = { { 36, 68, 60, 92, 34, 66, 58, 90, }, { 100, 4, 124, 28, 98, 2, 122, 26, }, { 52, 84, 44, 76, 50, 82, 42, 74, }, @@ -44,6 +45,7 @@ DECLARE_ALIGNED(8, const uint8_t, ff_dither_8x8_128)[8][8] = { { 96, 0, 120, 24, 102, 6, 126, 30, }, { 48, 80, 40, 72, 54, 86, 46, 78, }, { 112, 16, 104, 8, 118, 22, 110, 14, }, + { 36, 68, 60, 92, 34, 66, 58, 90, }, }; DECLARE_ALIGNED(8, static const uint8_t, sws_pb_64)[8] = { @@ -61,28 +63,6 @@ static av_always_inline void fillPlane(uint8_t *plane, int stride, int width, } } -static void fill_plane9or10(uint8_t *plane, int stride, int width, - int height, int y, uint8_t val, - const int dst_depth, const int big_endian) -{ - int i, j; - uint16_t *dst = (uint16_t *) (plane + stride * y); -#define FILL8TO9_OR_10(wfunc) \ - for (i = 0; i < height; i++) { \ - for (j = 0; j < width; j++) { \ - wfunc(&dst[j], (val << (dst_depth - 8)) | \ - (val >> (16 - dst_depth))); \ - } \ - dst += stride / 2; \ - } - if (big_endian) { - FILL8TO9_OR_10(AV_WB16); - } else { - FILL8TO9_OR_10(AV_WL16); - } -} - - static void hScale16To19_c(SwsContext *c, int16_t *_dst, int dstW, const uint8_t *_src, const int16_t *filter, const int32_t *filterPos, int filterSize) @@ -94,6 +74,9 @@ static void hScale16To19_c(SwsContext *c, int16_t *_dst, int dstW, int bits = desc->comp[0].depth_minus1; int sh = bits - 4; + if((isAnyRGB(c->srcFormat) || c->srcFormat==AV_PIX_FMT_PAL8) && desc->comp[0].depth_minus1<15) + sh= 9; + for (i = 0; i < dstW; i++) { int j; int srcPos = filterPos[i]; @@ -116,6 +99,9 @@ static void hScale16To15_c(SwsContext *c, int16_t *dst, int dstW, const uint16_t *src = (const uint16_t *) _src; int sh = desc->comp[0].depth_minus1; + if(sh<15) + sh= isAnyRGB(c->srcFormat) || c->srcFormat==AV_PIX_FMT_PAL8 ? 13 : desc->comp[0].depth_minus1; + for (i = 0; i < dstW; i++) { int j; int srcPos = filterPos[i]; @@ -223,8 +209,9 @@ static void lumRangeToJpeg16_c(int16_t *_dst, int width) { int i; int32_t *dst = (int32_t *) _dst; - for (i = 0; i < width; i++) - dst[i] = (FFMIN(dst[i], 30189 << 4) * 4769 - (39057361 << 2)) >> 12; + for (i = 0; i < width; i++) { + dst[i] = ((int)(FFMIN(dst[i], 30189 << 4) * 4769U - (39057361 << 2))) >> 12; + } } static void lumRangeFromJpeg16_c(int16_t *_dst, int width) @@ -232,20 +219,7 @@ static void lumRangeFromJpeg16_c(int16_t *_dst, int width) int i; int32_t *dst = (int32_t *) _dst; for (i = 0; i < width; i++) - dst[i] = (dst[i] * 14071 + (33561947 << 4)) >> 14; -} - -static void hyscale_fast_c(SwsContext *c, int16_t *dst, int dstWidth, - const uint8_t *src, int srcW, int xInc) -{ - int i; - unsigned int xpos = 0; - for (i = 0; i < dstWidth; i++) { - register unsigned int xx = xpos >> 16; - register unsigned int xalpha = (xpos & 0xFFFF) >> 9; - dst[i] = (src[xx] << 7) + (src[xx + 1] - src[xx]) * xalpha; - xpos += xInc; - } + dst[i] = (dst[i]*(14071/4) + (33561947<<4)/4)>>12; } // *** horizontal scale Y line to temp buffer @@ -258,19 +232,19 @@ static av_always_inline void hyscale(SwsContext *c, int16_t *dst, int dstWidth, uint8_t *formatConvBuffer, uint32_t *pal, int isAlpha) { - void (*toYV12)(uint8_t *, const uint8_t *, int, uint32_t *) = + 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); + c->readLumPlanar(formatConvBuffer, src_in, srcW, c->input_rgb2yuv_table); src = formatConvBuffer; } else if (c->readAlpPlanar && isAlpha) { - c->readAlpPlanar(formatConvBuffer, src_in, srcW); + c->readAlpPlanar(formatConvBuffer, src_in, srcW, NULL); src = formatConvBuffer; } @@ -285,21 +259,6 @@ static av_always_inline void hyscale(SwsContext *c, int16_t *dst, int dstWidth, convertRange(dst, dstWidth); } -static void hcscale_fast_c(SwsContext *c, int16_t *dst1, int16_t *dst2, - int dstWidth, const uint8_t *src1, - const uint8_t *src2, int srcW, int xInc) -{ - int i; - unsigned int xpos = 0; - for (i = 0; i < dstWidth; i++) { - register unsigned int xx = xpos >> 16; - register unsigned int xalpha = (xpos & 0xFFFF) >> 9; - dst1[i] = (src1[xx] * (xalpha ^ 127) + src1[xx + 1] * xalpha); - dst2[i] = (src2[xx] * (xalpha ^ 127) + src2[xx + 1] * xalpha); - xpos += xInc; - } -} - static av_always_inline void hcscale(SwsContext *c, int16_t *dst1, int16_t *dst2, int dstWidth, const uint8_t *src_in[4], @@ -312,14 +271,14 @@ static av_always_inline void hcscale(SwsContext *c, int16_t *dst1, 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); - src1 = formatConvBuffer; - src2 = buf2; + 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); - c->readChrPlanar(formatConvBuffer, buf2, src_in, srcW); + FFALIGN(srcW*2+78, 16); + c->readChrPlanar(formatConvBuffer, buf2, src_in, srcW, c->input_rgb2yuv_table); src1 = formatConvBuffer; src2 = buf2; } @@ -359,8 +318,6 @@ static int swscale(SwsContext *c, const uint8_t *src[], int32_t *vChrFilterPos = c->vChrFilterPos; int32_t *hLumFilterPos = c->hLumFilterPos; int32_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; @@ -384,8 +341,8 @@ static int swscale(SwsContext *c, const uint8_t *src[], yuv2packed2_fn yuv2packed2 = c->yuv2packed2; yuv2packedX_fn yuv2packedX = c->yuv2packedX; yuv2anyX_fn yuv2anyX = c->yuv2anyX; - const int chrSrcSliceY = srcSliceY >> c->chrSrcVSubSample; - const int chrSrcSliceH = -((-srcSliceH) >> c->chrSrcVSubSample); + const int chrSrcSliceY = srcSliceY >> c->chrSrcVSubSample; + const int chrSrcSliceH = FF_CEIL_RSHIFT(srcSliceH, c->chrSrcVSubSample); int should_dither = is9_OR_10BPS(c->srcFormat) || is16BPS(c->srcFormat); int lastDstY; @@ -397,6 +354,10 @@ static int swscale(SwsContext *c, const uint8_t *src[], int lastInLumBuf = c->lastInLumBuf; int lastInChrBuf = c->lastInChrBuf; + if (!usePal(c->srcFormat)) { + pal = c->input_rgb2yuv_table; + } + if (isPacked(c->srcFormat)) { src[0] = src[1] = @@ -420,8 +381,8 @@ static int swscale(SwsContext *c, const uint8_t *src[], 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]&15 || dstStride[1]&15 || + dstStride[2]&15 || dstStride[3]&15) { static int warnedAlready = 0; // FIXME maybe move this into the context if (flags & SWS_PRINT_INFO && !warnedAlready) { av_log(c, AV_LOG_WARNING, @@ -431,6 +392,19 @@ static int swscale(SwsContext *c, const uint8_t *src[], } } + if ( (uintptr_t)dst[0]&15 || (uintptr_t)dst[1]&15 || (uintptr_t)dst[2]&15 + || (uintptr_t)src[0]&15 || (uintptr_t)src[1]&15 || (uintptr_t)src[2]&15 + || dstStride[0]&15 || dstStride[1]&15 || dstStride[2]&15 || dstStride[3]&15 + || srcStride[0]&15 || srcStride[1]&15 || srcStride[2]&15 || srcStride[3]&15 + ) { + static int warnedAlready=0; + int cpu_flags = av_get_cpu_flags(); + if (HAVE_MMXEXT && (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. */ @@ -455,6 +429,7 @@ static int swscale(SwsContext *c, const uint8_t *src[], dst[2] + dstStride[2] * chrDstY, (CONFIG_SWSCALE_ALPHA && alpPixBuf) ? dst[3] + dstStride[3] * dstY : NULL, }; + int use_mmx_vfilter= c->use_mmx_vfilter; // First line needed as input const int firstLumSrcY = FFMAX(1 - vLumFilterSize, vLumFilterPos[dstY]); @@ -473,8 +448,8 @@ static int swscale(SwsContext *c, const uint8_t *src[], lastInLumBuf = firstLumSrcY - 1; if (firstChrSrcY > lastInChrBuf) lastInChrBuf = firstChrSrcY - 1; - assert(firstLumSrcY >= lastInLumBuf - vLumBufSize + 1); - assert(firstChrSrcY >= lastInChrBuf - vChrBufSize + 1); + av_assert0(firstLumSrcY >= lastInLumBuf - vLumBufSize + 1); + av_assert0(firstChrSrcY >= lastInChrBuf - vChrBufSize + 1); DEBUG_BUFFERS("dstY: %d\n", dstY); DEBUG_BUFFERS("\tfirstLumSrcY: %d lastLumSrcY: %d lastInLumBuf: %d\n", @@ -484,7 +459,7 @@ static int swscale(SwsContext *c, const uint8_t *src[], // Do we have enough lines in this slice to output the dstY line enough_lines = lastLumSrcY2 < srcSliceY + srcSliceH && - lastChrSrcY < -((-srcSliceY - srcSliceH) >> c->chrSrcVSubSample); + lastChrSrcY < FF_CEIL_RSHIFT(srcSliceY + srcSliceH, c->chrSrcVSubSample); if (!enough_lines) { lastLumSrcY = srcSliceY + srcSliceH - 1; @@ -502,9 +477,9 @@ static int swscale(SwsContext *c, const uint8_t *src[], src[3] + (lastInLumBuf + 1 - srcSliceY) * srcStride[3], }; lumBufIndex++; - assert(lumBufIndex < 2 * vLumBufSize); - assert(lastInLumBuf + 1 - srcSliceY < srcSliceH); - assert(lastInLumBuf + 1 - srcSliceY >= 0); + av_assert0(lumBufIndex < 2 * vLumBufSize); + av_assert0(lastInLumBuf + 1 - srcSliceY < srcSliceH); + av_assert0(lastInLumBuf + 1 - srcSliceY >= 0); hyscale(c, lumPixBuf[lumBufIndex], dstW, src1, srcW, lumXInc, hLumFilter, hLumFilterPos, hLumFilterSize, formatConvBuffer, pal, 0); @@ -524,9 +499,9 @@ static int swscale(SwsContext *c, const uint8_t *src[], src[3] + (lastInChrBuf + 1 - chrSrcSliceY) * srcStride[3], }; chrBufIndex++; - assert(chrBufIndex < 2 * vChrBufSize); - assert(lastInChrBuf + 1 - chrSrcSliceY < (chrSrcSliceH)); - assert(lastInChrBuf + 1 - chrSrcSliceY >= 0); + av_assert0(chrBufIndex < 2 * vChrBufSize); + av_assert0(lastInChrBuf + 1 - chrSrcSliceY < (chrSrcSliceH)); + av_assert0(lastInChrBuf + 1 - chrSrcSliceY >= 0); // FIXME replace parameters through context struct (some at least) if (c->needs_hcscale) @@ -547,7 +522,7 @@ static int swscale(SwsContext *c, const uint8_t *src[], break; // we can't output a dstY line so let's try with the next slice #if HAVE_MMX_INLINE - updateMMXDitherTables(c, dstY, lumBufIndex, chrBufIndex, + ff_updateMMXDitherTables(c, dstY, lumBufIndex, chrBufIndex, lastInLumBuf, lastInChrBuf); #endif if (should_dither) { @@ -559,103 +534,81 @@ static int swscale(SwsContext *c, const uint8_t *src[], * this array's tail */ ff_sws_init_output_funcs(c, &yuv2plane1, &yuv2planeX, &yuv2nv12cX, &yuv2packed1, &yuv2packed2, &yuv2packedX, &yuv2anyX); + 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 **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 **)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; - int 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; - int 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 **)(void*) alpPixBuf + lumBufIndex + firstLumSrcY - lastInLumBuf + vLumBufSize : NULL; + int16_t *vLumFilter = c->vLumFilter; + int16_t *vChrFilter = c->vChrFilter; if (isPlanarYUV(dstFormat) || (isGray(dstFormat) && !isALPHA(dstFormat))) { // 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= (int16_t *)c->lumMmxFilter; + vChrFilter= (int16_t *)c->chrMmxFilter; + } + if (vLumFilterSize == 1) { yuv2plane1(lumSrcPtr[0], dest[0], dstW, c->lumDither8, 0); } else { - yuv2planeX(vLumFilter + dstY * vLumFilterSize, - vLumFilterSize, lumSrcPtr, dest[0], + yuv2planeX(vLumFilter, vLumFilterSize, + lumSrcPtr, dest[0], dstW, c->lumDither8, 0); } if (!((dstY & chrSkipMask) || isGray(dstFormat))) { if (yuv2nv12cX) { - yuv2nv12cX(c, vChrFilter + chrDstY * vChrFilterSize, + 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, + yuv2planeX(vChrFilter, vChrFilterSize, chrUSrcPtr, dest[1], chrDstW, c->chrDither8, 0); - yuv2planeX(vChrFilter + chrDstY * vChrFilterSize, + yuv2planeX(vChrFilter, vChrFilterSize, chrVSrcPtr, dest[2], - chrDstW, c->chrDither8, 3); + chrDstW, c->chrDither8, use_mmx_vfilter ? (c->uv_offx2 >> 1) : 3); } } if (CONFIG_SWSCALE_ALPHA && alpPixBuf) { + if(use_mmx_vfilter){ + vLumFilter= (int16_t *)c->alpMmxFilter; + } if (vLumFilterSize == 1) { yuv2plane1(alpSrcPtr[0], dest[3], dstW, c->lumDither8, 0); } else { - yuv2planeX(vLumFilter + dstY * vLumFilterSize, + yuv2planeX(vLumFilter, vLumFilterSize, alpSrcPtr, dest[3], dstW, c->lumDither8, 0); } } } else if (yuv2packedX) { + av_assert1(lumSrcPtr + vLumFilterSize - 1 < (const int16_t **)lumPixBuf + vLumBufSize * 2); + av_assert1(chrUSrcPtr + vChrFilterSize - 1 < (const int16_t **)chrUPixBuf + vChrBufSize * 2); if (c->yuv2packed1 && vLumFilterSize == 1 && vChrFilterSize <= 2) { // unscaled RGB int chrAlpha = vChrFilterSize == 1 ? 0 : vChrFilter[2 * dstY + 1]; @@ -681,6 +634,7 @@ static int swscale(SwsContext *c, const uint8_t *src[], alpSrcPtr, dest[0], dstW, dstY); } } else { + av_assert1(!yuv2packed1 && !yuv2packed2); yuv2anyX(c, vLumFilter + dstY * vLumFilterSize, lumSrcPtr, vLumFilterSize, vChrFilter + dstY * vChrFilterSize, @@ -689,18 +643,15 @@ static int swscale(SwsContext *c, const uint8_t *src[], } } } - if (isPlanar(dstFormat) && isALPHA(dstFormat) && !alpPixBuf) { int length = dstW; int height = dstY - lastDstY; - if (is16BPS(c->dstFormat)) - length *= 2; - if (is9_OR_10BPS(dstFormat)) { + if (is16BPS(dstFormat) || isNBPS(dstFormat)) { const AVPixFmtDescriptor *desc = av_pix_fmt_desc_get(dstFormat); - fill_plane9or10(dst[3], dstStride[3], length, height, lastDstY, - 255, desc->comp[3].depth_minus1 + 1, - isBE(dstFormat)); + fillPlane16(dst[3], dstStride[3], length, height, lastDstY, + 1, desc->comp[3].depth_minus1, + isBE(dstFormat)); } else fillPlane(dst[3], dstStride[3], length, height, lastDstY, 255); } @@ -721,6 +672,31 @@ static int swscale(SwsContext *c, const uint8_t *src[], return dstY - lastDstY; } +av_cold void ff_sws_init_range_convert(SwsContext *c) +{ + c->lumConvertRange = NULL; + c->chrConvertRange = NULL; + if (c->srcRange != c->dstRange && !isAnyRGB(c->dstFormat)) { + if (c->dstBpc <= 14) { + if (c->srcRange) { + c->lumConvertRange = lumRangeFromJpeg_c; + c->chrConvertRange = chrRangeFromJpeg_c; + } else { + c->lumConvertRange = lumRangeToJpeg_c; + c->chrConvertRange = chrRangeToJpeg_c; + } + } else { + if (c->srcRange) { + c->lumConvertRange = lumRangeFromJpeg16_c; + c->chrConvertRange = chrRangeFromJpeg16_c; + } else { + c->lumConvertRange = lumRangeToJpeg16_c; + c->chrConvertRange = chrRangeToJpeg16_c; + } + } + } +} + static av_cold void sws_init_swscale(SwsContext *c) { enum AVPixelFormat srcFormat = c->srcFormat; @@ -731,40 +707,23 @@ static av_cold void sws_init_swscale(SwsContext *c) ff_sws_init_input_funcs(c); + if (c->srcBpc == 8) { - if (c->dstBpc <= 10) { + if (c->dstBpc <= 14) { c->hyScale = c->hcScale = hScale8To15_c; if (c->flags & SWS_FAST_BILINEAR) { - c->hyscale_fast = hyscale_fast_c; - c->hcscale_fast = hcscale_fast_c; + c->hyscale_fast = ff_hyscale_fast_c; + c->hcscale_fast = ff_hcscale_fast_c; } } else { c->hyScale = c->hcScale = hScale8To19_c; } } else { - c->hyScale = c->hcScale = c->dstBpc > 10 ? hScale16To19_c + c->hyScale = c->hcScale = c->dstBpc > 14 ? hScale16To19_c : hScale16To15_c; } - if (c->srcRange != c->dstRange && !isAnyRGB(c->dstFormat)) { - if (c->dstBpc <= 10) { - if (c->srcRange) { - c->lumConvertRange = lumRangeFromJpeg_c; - c->chrConvertRange = chrRangeFromJpeg_c; - } else { - c->lumConvertRange = lumRangeToJpeg_c; - c->chrConvertRange = chrRangeToJpeg_c; - } - } else { - if (c->srcRange) { - c->lumConvertRange = lumRangeFromJpeg16_c; - c->chrConvertRange = chrRangeFromJpeg16_c; - } else { - c->lumConvertRange = lumRangeToJpeg16_c; - c->chrConvertRange = chrRangeToJpeg16_c; - } - } - } + ff_sws_init_range_convert(c); if (!(isGray(srcFormat) || isGray(c->dstFormat) || srcFormat == AV_PIX_FMT_MONOBLACK || srcFormat == AV_PIX_FMT_MONOWHITE)) @@ -782,3 +741,361 @@ SwsFunc ff_getSwsFunc(SwsContext *c) return swscale; } + +static void reset_ptr(const uint8_t *src[], enum AVPixelFormat format) +{ + if (!isALPHA(format)) + src[3] = NULL; + if (!isPlanar(format)) { + src[3] = src[2] = NULL; + + if (!usePal(format)) + src[1] = NULL; + } +} + +static int check_image_pointers(const uint8_t * const data[4], enum AVPixelFormat pix_fmt, + const int linesizes[4]) +{ + const AVPixFmtDescriptor *desc = av_pix_fmt_desc_get(pix_fmt); + int i; + + for (i = 0; i < 4; i++) { + int plane = desc->comp[i].plane; + if (!data[plane] || !linesizes[plane]) + return 0; + } + + return 1; +} + +static void xyz12Torgb48(struct SwsContext *c, uint16_t *dst, + const uint16_t *src, int stride, int h) +{ + int xp,yp; + const AVPixFmtDescriptor *desc = av_pix_fmt_desc_get(c->srcFormat); + + for (yp=0; yp<h; yp++) { + for (xp=0; xp+2<stride; xp+=3) { + int x, y, z, r, g, b; + + if (desc->flags & AV_PIX_FMT_FLAG_BE) { + x = AV_RB16(src + xp + 0); + y = AV_RB16(src + xp + 1); + z = AV_RB16(src + xp + 2); + } else { + x = AV_RL16(src + xp + 0); + y = AV_RL16(src + xp + 1); + z = AV_RL16(src + xp + 2); + } + + x = c->xyzgamma[x>>4]; + y = c->xyzgamma[y>>4]; + z = c->xyzgamma[z>>4]; + + // convert from XYZlinear to sRGBlinear + r = c->xyz2rgb_matrix[0][0] * x + + c->xyz2rgb_matrix[0][1] * y + + c->xyz2rgb_matrix[0][2] * z >> 12; + g = c->xyz2rgb_matrix[1][0] * x + + c->xyz2rgb_matrix[1][1] * y + + c->xyz2rgb_matrix[1][2] * z >> 12; + b = c->xyz2rgb_matrix[2][0] * x + + c->xyz2rgb_matrix[2][1] * y + + c->xyz2rgb_matrix[2][2] * z >> 12; + + // limit values to 12-bit depth + r = av_clip_uintp2(r, 12); + g = av_clip_uintp2(g, 12); + b = av_clip_uintp2(b, 12); + + // convert from sRGBlinear to RGB and scale from 12bit to 16bit + if (desc->flags & AV_PIX_FMT_FLAG_BE) { + AV_WB16(dst + xp + 0, c->rgbgamma[r] << 4); + AV_WB16(dst + xp + 1, c->rgbgamma[g] << 4); + AV_WB16(dst + xp + 2, c->rgbgamma[b] << 4); + } else { + AV_WL16(dst + xp + 0, c->rgbgamma[r] << 4); + AV_WL16(dst + xp + 1, c->rgbgamma[g] << 4); + AV_WL16(dst + xp + 2, c->rgbgamma[b] << 4); + } + } + src += stride; + dst += stride; + } +} + +static void rgb48Toxyz12(struct SwsContext *c, uint16_t *dst, + const uint16_t *src, int stride, int h) +{ + int xp,yp; + const AVPixFmtDescriptor *desc = av_pix_fmt_desc_get(c->dstFormat); + + for (yp=0; yp<h; yp++) { + for (xp=0; xp+2<stride; xp+=3) { + int x, y, z, r, g, b; + + if (desc->flags & AV_PIX_FMT_FLAG_BE) { + r = AV_RB16(src + xp + 0); + g = AV_RB16(src + xp + 1); + b = AV_RB16(src + xp + 2); + } else { + r = AV_RL16(src + xp + 0); + g = AV_RL16(src + xp + 1); + b = AV_RL16(src + xp + 2); + } + + r = c->rgbgammainv[r>>4]; + g = c->rgbgammainv[g>>4]; + b = c->rgbgammainv[b>>4]; + + // convert from sRGBlinear to XYZlinear + x = c->rgb2xyz_matrix[0][0] * r + + c->rgb2xyz_matrix[0][1] * g + + c->rgb2xyz_matrix[0][2] * b >> 12; + y = c->rgb2xyz_matrix[1][0] * r + + c->rgb2xyz_matrix[1][1] * g + + c->rgb2xyz_matrix[1][2] * b >> 12; + z = c->rgb2xyz_matrix[2][0] * r + + c->rgb2xyz_matrix[2][1] * g + + c->rgb2xyz_matrix[2][2] * b >> 12; + + // limit values to 12-bit depth + x = av_clip_uintp2(x, 12); + y = av_clip_uintp2(y, 12); + z = av_clip_uintp2(z, 12); + + // convert from XYZlinear to X'Y'Z' and scale from 12bit to 16bit + if (desc->flags & AV_PIX_FMT_FLAG_BE) { + AV_WB16(dst + xp + 0, c->xyzgammainv[x] << 4); + AV_WB16(dst + xp + 1, c->xyzgammainv[y] << 4); + AV_WB16(dst + xp + 2, c->xyzgammainv[z] << 4); + } else { + AV_WL16(dst + xp + 0, c->xyzgammainv[x] << 4); + AV_WL16(dst + xp + 1, c->xyzgammainv[y] << 4); + AV_WL16(dst + xp + 2, c->xyzgammainv[z] << 4); + } + } + src += stride; + dst += stride; + } +} + +/** + * swscale wrapper, so we don't need to export the SwsContext. + * Assumes planar YUV to be in YUV order instead of YVU. + */ +int attribute_align_arg sws_scale(struct SwsContext *c, + const uint8_t * const srcSlice[], + const int srcStride[], int srcSliceY, + int srcSliceH, uint8_t *const dst[], + const int dstStride[]) +{ + int i, ret; + const uint8_t *src2[4]; + uint8_t *dst2[4]; + uint8_t *rgb0_tmp = NULL; + + if (!srcStride || !dstStride || !dst || !srcSlice) { + av_log(c, AV_LOG_ERROR, "One of the input parameters to sws_scale() is NULL, please check the calling code\n"); + return 0; + } + if (c->cascaded_context[0] && srcSliceY == 0 && srcSliceH == c->cascaded_context[0]->srcH) { + ret = sws_scale(c->cascaded_context[0], + srcSlice, srcStride, srcSliceY, srcSliceH, + c->cascaded_tmp, c->cascaded_tmpStride); + if (ret < 0) + return ret; + ret = sws_scale(c->cascaded_context[1], + (const uint8_t * const * )c->cascaded_tmp, c->cascaded_tmpStride, 0, c->cascaded_context[0]->dstH, + dst, dstStride); + return ret; + } + + memcpy(src2, srcSlice, sizeof(src2)); + memcpy(dst2, dst, sizeof(dst2)); + + // do not mess up sliceDir if we have a "trailing" 0-size slice + if (srcSliceH == 0) + return 0; + + if (!check_image_pointers(srcSlice, c->srcFormat, srcStride)) { + av_log(c, AV_LOG_ERROR, "bad src image pointers\n"); + return 0; + } + if (!check_image_pointers((const uint8_t* const*)dst, c->dstFormat, dstStride)) { + av_log(c, AV_LOG_ERROR, "bad dst image pointers\n"); + return 0; + } + + if (c->sliceDir == 0 && srcSliceY != 0 && srcSliceY + srcSliceH != c->srcH) { + av_log(c, AV_LOG_ERROR, "Slices start in the middle!\n"); + return 0; + } + if (c->sliceDir == 0) { + if (srcSliceY == 0) c->sliceDir = 1; else c->sliceDir = -1; + } + + if (usePal(c->srcFormat)) { + for (i = 0; i < 256; i++) { + int r, g, b, y, u, v, a = 0xff; + if (c->srcFormat == AV_PIX_FMT_PAL8) { + uint32_t p = ((const uint32_t *)(srcSlice[1]))[i]; + a = (p >> 24) & 0xFF; + r = (p >> 16) & 0xFF; + g = (p >> 8) & 0xFF; + b = p & 0xFF; + } else if (c->srcFormat == AV_PIX_FMT_RGB8) { + r = ( i >> 5 ) * 36; + g = ((i >> 2) & 7) * 36; + b = ( i & 3) * 85; + } else if (c->srcFormat == AV_PIX_FMT_BGR8) { + b = ( i >> 6 ) * 85; + g = ((i >> 3) & 7) * 36; + r = ( i & 7) * 36; + } else if (c->srcFormat == AV_PIX_FMT_RGB4_BYTE) { + r = ( i >> 3 ) * 255; + g = ((i >> 1) & 3) * 85; + b = ( i & 1) * 255; + } else if (c->srcFormat == AV_PIX_FMT_GRAY8 || c->srcFormat == AV_PIX_FMT_GRAY8A) { + r = g = b = i; + } else { + av_assert1(c->srcFormat == AV_PIX_FMT_BGR4_BYTE); + b = ( i >> 3 ) * 255; + g = ((i >> 1) & 3) * 85; + r = ( i & 1) * 255; + } +#define RGB2YUV_SHIFT 15 +#define BY ( (int) (0.114 * 219 / 255 * (1 << RGB2YUV_SHIFT) + 0.5)) +#define BV (-(int) (0.081 * 224 / 255 * (1 << RGB2YUV_SHIFT) + 0.5)) +#define BU ( (int) (0.500 * 224 / 255 * (1 << RGB2YUV_SHIFT) + 0.5)) +#define GY ( (int) (0.587 * 219 / 255 * (1 << RGB2YUV_SHIFT) + 0.5)) +#define GV (-(int) (0.419 * 224 / 255 * (1 << RGB2YUV_SHIFT) + 0.5)) +#define GU (-(int) (0.331 * 224 / 255 * (1 << RGB2YUV_SHIFT) + 0.5)) +#define RY ( (int) (0.299 * 219 / 255 * (1 << RGB2YUV_SHIFT) + 0.5)) +#define RV ( (int) (0.500 * 224 / 255 * (1 << RGB2YUV_SHIFT) + 0.5)) +#define RU (-(int) (0.169 * 224 / 255 * (1 << RGB2YUV_SHIFT) + 0.5)) + + y = av_clip_uint8((RY * r + GY * g + BY * b + ( 33 << (RGB2YUV_SHIFT - 1))) >> RGB2YUV_SHIFT); + u = av_clip_uint8((RU * r + GU * g + BU * b + (257 << (RGB2YUV_SHIFT - 1))) >> RGB2YUV_SHIFT); + v = av_clip_uint8((RV * r + GV * g + BV * b + (257 << (RGB2YUV_SHIFT - 1))) >> RGB2YUV_SHIFT); + c->pal_yuv[i]= y + (u<<8) + (v<<16) + ((unsigned)a<<24); + + switch (c->dstFormat) { + case AV_PIX_FMT_BGR32: +#if !HAVE_BIGENDIAN + case AV_PIX_FMT_RGB24: +#endif + c->pal_rgb[i]= r + (g<<8) + (b<<16) + ((unsigned)a<<24); + break; + case AV_PIX_FMT_BGR32_1: +#if HAVE_BIGENDIAN + case AV_PIX_FMT_BGR24: +#endif + c->pal_rgb[i]= a + (r<<8) + (g<<16) + ((unsigned)b<<24); + break; + case AV_PIX_FMT_RGB32_1: +#if HAVE_BIGENDIAN + case AV_PIX_FMT_RGB24: +#endif + c->pal_rgb[i]= a + (b<<8) + (g<<16) + ((unsigned)r<<24); + break; + case AV_PIX_FMT_RGB32: +#if !HAVE_BIGENDIAN + case AV_PIX_FMT_BGR24: +#endif + default: + c->pal_rgb[i]= b + (g<<8) + (r<<16) + ((unsigned)a<<24); + } + } + } + + if (c->src0Alpha && !c->dst0Alpha && isALPHA(c->dstFormat)) { + uint8_t *base; + int x,y; + rgb0_tmp = av_malloc(FFABS(srcStride[0]) * srcSliceH + 32); + if (!rgb0_tmp) + return AVERROR(ENOMEM); + + base = srcStride[0] < 0 ? rgb0_tmp - srcStride[0] * (srcSliceH-1) : rgb0_tmp; + for (y=0; y<srcSliceH; y++){ + memcpy(base + srcStride[0]*y, src2[0] + srcStride[0]*y, 4*c->srcW); + for (x=c->src0Alpha-1; x<4*c->srcW; x+=4) { + base[ srcStride[0]*y + x] = 0xFF; + } + } + src2[0] = base; + } + + if (c->srcXYZ && !(c->dstXYZ && c->srcW==c->dstW && c->srcH==c->dstH)) { + uint8_t *base; + rgb0_tmp = av_malloc(FFABS(srcStride[0]) * srcSliceH + 32); + if (!rgb0_tmp) + return AVERROR(ENOMEM); + + base = srcStride[0] < 0 ? rgb0_tmp - srcStride[0] * (srcSliceH-1) : rgb0_tmp; + + xyz12Torgb48(c, (uint16_t*)base, (const uint16_t*)src2[0], srcStride[0]/2, srcSliceH); + src2[0] = base; + } + + if (!srcSliceY && (c->flags & SWS_BITEXACT) && c->dither == SWS_DITHER_ED && c->dither_error[0]) + for (i = 0; i < 4; i++) + memset(c->dither_error[i], 0, sizeof(c->dither_error[0][0]) * (c->dstW+2)); + + + // copy strides, so they can safely be modified + if (c->sliceDir == 1) { + // slices go from top to bottom + int srcStride2[4] = { srcStride[0], srcStride[1], srcStride[2], + srcStride[3] }; + int dstStride2[4] = { dstStride[0], dstStride[1], dstStride[2], + dstStride[3] }; + + reset_ptr(src2, c->srcFormat); + reset_ptr((void*)dst2, c->dstFormat); + + /* reset slice direction at end of frame */ + if (srcSliceY + srcSliceH == c->srcH) + c->sliceDir = 0; + + ret = c->swscale(c, src2, srcStride2, srcSliceY, srcSliceH, dst2, + dstStride2); + } else { + // slices go from bottom to top => we flip the image internally + int srcStride2[4] = { -srcStride[0], -srcStride[1], -srcStride[2], + -srcStride[3] }; + int dstStride2[4] = { -dstStride[0], -dstStride[1], -dstStride[2], + -dstStride[3] }; + + src2[0] += (srcSliceH - 1) * srcStride[0]; + if (!usePal(c->srcFormat)) + src2[1] += ((srcSliceH >> c->chrSrcVSubSample) - 1) * srcStride[1]; + src2[2] += ((srcSliceH >> c->chrSrcVSubSample) - 1) * srcStride[2]; + src2[3] += (srcSliceH - 1) * srcStride[3]; + dst2[0] += ( c->dstH - 1) * dstStride[0]; + dst2[1] += ((c->dstH >> c->chrDstVSubSample) - 1) * dstStride[1]; + dst2[2] += ((c->dstH >> c->chrDstVSubSample) - 1) * dstStride[2]; + dst2[3] += ( c->dstH - 1) * dstStride[3]; + + reset_ptr(src2, c->srcFormat); + reset_ptr((void*)dst2, c->dstFormat); + + /* reset slice direction at end of frame */ + if (!srcSliceY) + c->sliceDir = 0; + + ret = c->swscale(c, src2, srcStride2, c->srcH-srcSliceY-srcSliceH, + srcSliceH, dst2, dstStride2); + } + + + if (c->dstXYZ && !(c->srcXYZ && c->srcW==c->dstW && c->srcH==c->dstH)) { + /* replace on the same data */ + rgb48Toxyz12(c, (uint16_t*)dst2[0], (const uint16_t*)dst2[0], dstStride[0]/2, ret); + } + + av_free(rgb0_tmp); + return ret; +} + |