diff options
Diffstat (limited to 'libswscale/utils.c')
-rw-r--r-- | libswscale/utils.c | 525 |
1 files changed, 418 insertions, 107 deletions
diff --git a/libswscale/utils.c b/libswscale/utils.c index 27819854e4..53687f46c2 100644 --- a/libswscale/utils.c +++ b/libswscale/utils.c @@ -1,27 +1,27 @@ /* * Copyright (C) 2001-2003 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 "config.h" #define _SVID_SOURCE // needed for MAP_ANONYMOUS -#include <assert.h> +#define _DARWIN_C_SOURCE // needed for MAP_ANON #include <inttypes.h> #include <math.h> #include <stdio.h> @@ -38,6 +38,7 @@ #endif #include "libavutil/attributes.h" +#include "libavutil/avassert.h" #include "libavutil/avutil.h" #include "libavutil/bswap.h" #include "libavutil/cpu.h" @@ -52,20 +53,23 @@ #include "swscale.h" #include "swscale_internal.h" +static void handle_formats(SwsContext *c); + unsigned swscale_version(void) { + av_assert0(LIBSWSCALE_VERSION_MICRO >= 100); return LIBSWSCALE_VERSION_INT; } const char *swscale_configuration(void) { - return LIBAV_CONFIGURATION; + return FFMPEG_CONFIGURATION; } const char *swscale_license(void) { #define LICENSE_PREFIX "libswscale license: " - return LICENSE_PREFIX LIBAV_LICENSE + sizeof(LICENSE_PREFIX) - 1; + return LICENSE_PREFIX FFMPEG_LICENSE + sizeof(LICENSE_PREFIX) - 1; } #define RET 0xC3 // near return opcode for x86 @@ -90,6 +94,7 @@ static const FormatEntry format_entries[AV_PIX_FMT_NB] = { [AV_PIX_FMT_MONOBLACK] = { 1, 1 }, [AV_PIX_FMT_PAL8] = { 1, 0 }, [AV_PIX_FMT_YUVJ420P] = { 1, 1 }, + [AV_PIX_FMT_YUVJ411P] = { 1, 1 }, [AV_PIX_FMT_YUVJ422P] = { 1, 1 }, [AV_PIX_FMT_YUVJ444P] = { 1, 1 }, [AV_PIX_FMT_UYVY422] = { 1, 1 }, @@ -106,6 +111,10 @@ static const FormatEntry format_entries[AV_PIX_FMT_NB] = { [AV_PIX_FMT_RGBA] = { 1, 1 }, [AV_PIX_FMT_ABGR] = { 1, 1 }, [AV_PIX_FMT_BGRA] = { 1, 1 }, + [AV_PIX_FMT_0RGB] = { 1, 1 }, + [AV_PIX_FMT_RGB0] = { 1, 1 }, + [AV_PIX_FMT_0BGR] = { 1, 1 }, + [AV_PIX_FMT_BGR0] = { 1, 1 }, [AV_PIX_FMT_GRAY16BE] = { 1, 1 }, [AV_PIX_FMT_GRAY16LE] = { 1, 1 }, [AV_PIX_FMT_YUV440P] = { 1, 1 }, @@ -133,6 +142,8 @@ static const FormatEntry format_entries[AV_PIX_FMT_NB] = { [AV_PIX_FMT_YUVA444P16LE]= { 1, 1 }, [AV_PIX_FMT_RGB48BE] = { 1, 1 }, [AV_PIX_FMT_RGB48LE] = { 1, 1 }, + [AV_PIX_FMT_RGBA64BE] = { 1, 1 }, + [AV_PIX_FMT_RGBA64LE] = { 1, 1 }, [AV_PIX_FMT_RGB565BE] = { 1, 1 }, [AV_PIX_FMT_RGB565LE] = { 1, 1 }, [AV_PIX_FMT_RGB555BE] = { 1, 1 }, @@ -154,27 +165,48 @@ static const FormatEntry format_entries[AV_PIX_FMT_NB] = { [AV_PIX_FMT_Y400A] = { 1, 0 }, [AV_PIX_FMT_BGR48BE] = { 1, 1 }, [AV_PIX_FMT_BGR48LE] = { 1, 1 }, + [AV_PIX_FMT_BGRA64BE] = { 0, 0 }, + [AV_PIX_FMT_BGRA64LE] = { 0, 0 }, [AV_PIX_FMT_YUV420P9BE] = { 1, 1 }, [AV_PIX_FMT_YUV420P9LE] = { 1, 1 }, [AV_PIX_FMT_YUV420P10BE] = { 1, 1 }, [AV_PIX_FMT_YUV420P10LE] = { 1, 1 }, + [AV_PIX_FMT_YUV420P12BE] = { 1, 1 }, + [AV_PIX_FMT_YUV420P12LE] = { 1, 1 }, + [AV_PIX_FMT_YUV420P14BE] = { 1, 1 }, + [AV_PIX_FMT_YUV420P14LE] = { 1, 1 }, [AV_PIX_FMT_YUV422P9BE] = { 1, 1 }, [AV_PIX_FMT_YUV422P9LE] = { 1, 1 }, [AV_PIX_FMT_YUV422P10BE] = { 1, 1 }, [AV_PIX_FMT_YUV422P10LE] = { 1, 1 }, + [AV_PIX_FMT_YUV422P12BE] = { 1, 1 }, + [AV_PIX_FMT_YUV422P12LE] = { 1, 1 }, + [AV_PIX_FMT_YUV422P14BE] = { 1, 1 }, + [AV_PIX_FMT_YUV422P14LE] = { 1, 1 }, [AV_PIX_FMT_YUV444P9BE] = { 1, 1 }, [AV_PIX_FMT_YUV444P9LE] = { 1, 1 }, [AV_PIX_FMT_YUV444P10BE] = { 1, 1 }, [AV_PIX_FMT_YUV444P10LE] = { 1, 1 }, + [AV_PIX_FMT_YUV444P12BE] = { 1, 1 }, + [AV_PIX_FMT_YUV444P12LE] = { 1, 1 }, + [AV_PIX_FMT_YUV444P14BE] = { 1, 1 }, + [AV_PIX_FMT_YUV444P14LE] = { 1, 1 }, [AV_PIX_FMT_GBRP] = { 1, 1 }, [AV_PIX_FMT_GBRP9LE] = { 1, 1 }, [AV_PIX_FMT_GBRP9BE] = { 1, 1 }, [AV_PIX_FMT_GBRP10LE] = { 1, 1 }, [AV_PIX_FMT_GBRP10BE] = { 1, 1 }, + [AV_PIX_FMT_GBRP12LE] = { 1, 1 }, + [AV_PIX_FMT_GBRP12BE] = { 1, 1 }, + [AV_PIX_FMT_GBRP14LE] = { 1, 1 }, + [AV_PIX_FMT_GBRP14BE] = { 1, 1 }, [AV_PIX_FMT_GBRP16LE] = { 1, 0 }, [AV_PIX_FMT_GBRP16BE] = { 1, 0 }, - [AV_PIX_FMT_XYZ12BE] = { 0, 0, 1 }, - [AV_PIX_FMT_XYZ12LE] = { 0, 0, 1 }, + [AV_PIX_FMT_XYZ12BE] = { 1, 1, 1 }, + [AV_PIX_FMT_XYZ12LE] = { 1, 1, 1 }, + [AV_PIX_FMT_GBRAP] = { 1, 1 }, + [AV_PIX_FMT_GBRAP16LE] = { 1, 0 }, + [AV_PIX_FMT_GBRAP16BE] = { 1, 0 }, }; int sws_isSupportedInput(enum AVPixelFormat pix_fmt) @@ -195,6 +227,7 @@ int sws_isSupportedEndiannessConversion(enum AVPixelFormat pix_fmt) format_entries[pix_fmt].is_supported_endianness : 0; } +#if FF_API_SWS_FORMAT_NAME const char *sws_format_name(enum AVPixelFormat format) { const AVPixFmtDescriptor *desc = av_pix_fmt_desc_get(format); @@ -203,6 +236,7 @@ const char *sws_format_name(enum AVPixelFormat format) else return "Unknown format"; } +#endif static double getSplineCoeff(double a, double b, double c, double d, double dist) @@ -217,12 +251,21 @@ static double getSplineCoeff(double a, double b, double c, double d, dist - 1.0); } +static av_cold int get_local_pos(SwsContext *s, int chr_subsample, int pos, int dir) +{ + if (pos < 0) { + pos = (128 << chr_subsample) - 128; + } + pos += 128; // relative to ideal left edge + return pos >> chr_subsample; +} + static av_cold int initFilter(int16_t **outFilter, int32_t **filterPos, int *outFilterSize, int xInc, int srcW, int dstW, int filterAlign, int one, int flags, int cpu_flags, SwsVector *srcFilter, SwsVector *dstFilter, - double param[2], int is_horizontal) + double param[2], int srcPos, int dstPos) { int i; int filterSize; @@ -230,7 +273,7 @@ static av_cold int initFilter(int16_t **outFilter, int32_t **filterPos, int minFilterSize; int64_t *filter = NULL; int64_t *filter2 = NULL; - const int64_t fone = 1LL << 54; + const int64_t fone = 1LL << (54 - FFMIN(av_log2(srcW/dstW), 8)); int ret = -1; emms_c(); // FIXME should not be required but IS (even for non-MMX versions) @@ -238,7 +281,7 @@ static av_cold int initFilter(int16_t **outFilter, int32_t **filterPos, // NOTE: the +3 is for the MMX(+1) / SSE(+3) scaler which reads over the end FF_ALLOC_OR_GOTO(NULL, *filterPos, (dstW + 3) * sizeof(**filterPos), fail); - if (FFABS(xInc - 0x10000) < 10) { // unscaled + if (FFABS(xInc - 0x10000) < 10 && srcPos == dstPos) { // unscaled int i; filterSize = 1; FF_ALLOCZ_OR_GOTO(NULL, filter, @@ -250,12 +293,12 @@ static av_cold int initFilter(int16_t **outFilter, int32_t **filterPos, } } else if (flags & SWS_POINT) { // lame looking point sampling mode int i; - int xDstInSrc; + int64_t xDstInSrc; filterSize = 1; FF_ALLOC_OR_GOTO(NULL, filter, dstW * sizeof(*filter) * filterSize, fail); - xDstInSrc = xInc / 2 - 0x8000; + xDstInSrc = ((dstPos*(int64_t)xInc)>>8) - ((srcPos*0x8000LL)>>7); for (i = 0; i < dstW; i++) { int xx = (xDstInSrc - ((filterSize - 1) << 15) + (1 << 15)) >> 16; @@ -266,12 +309,12 @@ static av_cold int initFilter(int16_t **outFilter, int32_t **filterPos, } else if ((xInc <= (1 << 16) && (flags & SWS_AREA)) || (flags & SWS_FAST_BILINEAR)) { // bilinear upscale int i; - int xDstInSrc; + int64_t xDstInSrc; filterSize = 2; FF_ALLOC_OR_GOTO(NULL, filter, dstW * sizeof(*filter) * filterSize, fail); - xDstInSrc = xInc / 2 - 0x8000; + xDstInSrc = ((dstPos*(int64_t)xInc)>>8) - ((srcPos*0x8000LL)>>7); for (i = 0; i < dstW; i++) { int xx = (xDstInSrc - ((filterSize - 1) << 15) + (1 << 15)) >> 16; int j; @@ -279,8 +322,7 @@ static av_cold int initFilter(int16_t **outFilter, int32_t **filterPos, (*filterPos)[i] = xx; // bilinear upscale / linear interpolate / area averaging for (j = 0; j < filterSize; j++) { - int64_t coeff = fone - FFABS((xx << 16) - xDstInSrc) * - (fone >> 16); + int64_t coeff= fone - FFABS(((int64_t)xx<<16) - xDstInSrc)*(fone>>16); if (coeff < 0) coeff = 0; filter[i * filterSize + j] = coeff; @@ -309,8 +351,7 @@ static av_cold int initFilter(int16_t **outFilter, int32_t **filterPos, else if (flags & SWS_BILINEAR) sizeFactor = 2; else { - sizeFactor = 0; // GCC warning killer - assert(0); + av_assert0(0); } if (xInc <= 1 << 16) @@ -324,7 +365,7 @@ static av_cold int initFilter(int16_t **outFilter, int32_t **filterPos, FF_ALLOC_OR_GOTO(NULL, filter, dstW * sizeof(*filter) * filterSize, fail); - xDstInSrc = xInc - 0x10000; + xDstInSrc = ((dstPos*(int64_t)xInc)>>7) - ((srcPos*0x10000LL)>>7); for (i = 0; i < dstW; i++) { int xx = (xDstInSrc - ((filterSize - 2) << 16)) / (1 << 17); int j; @@ -358,7 +399,7 @@ static av_cold int initFilter(int16_t **outFilter, int32_t **filterPos, (-12 * B - 48 * C) * d + (8 * B + 24 * C) * (1 << 30); } - coeff *= fone >> (30 + 24); + coeff /= (1LL<<54)/fone; } #if 0 else if (flags & SWS_X) { @@ -409,8 +450,7 @@ static av_cold int initFilter(int16_t **outFilter, int32_t **filterPos, double p = -2.196152422706632; coeff = getSplineCoeff(1.0, 0.0, p, -p - 1.0, floatd) * fone; } else { - coeff = 0.0; // GCC warning killer - assert(0); + av_assert0(0); } filter[i * filterSize + j] = coeff; @@ -423,13 +463,13 @@ static av_cold int initFilter(int16_t **outFilter, int32_t **filterPos, /* apply src & dst Filter to filter -> filter2 * av_free(filter); */ - assert(filterSize > 0); + av_assert0(filterSize > 0); filter2Size = filterSize; if (srcFilter) filter2Size += srcFilter->length - 1; if (dstFilter) filter2Size += dstFilter->length - 1; - assert(filter2Size > 0); + av_assert0(filter2Size > 0); FF_ALLOCZ_OR_GOTO(NULL, filter2, filter2Size * dstW * sizeof(*filter2), fail); for (i = 0; i < dstW; i++) { @@ -505,19 +545,21 @@ static av_cold int initFilter(int16_t **outFilter, int32_t **filterPos, filterAlign = 1; } - if (INLINE_MMX(cpu_flags)) { + if (HAVE_MMX && cpu_flags & AV_CPU_FLAG_MMX) { // special case for unscaled vertical filtering if (minFilterSize == 1 && filterAlign == 2) filterAlign = 1; } - assert(minFilterSize > 0); + av_assert0(minFilterSize > 0); filterSize = (minFilterSize + (filterAlign - 1)) & (~(filterAlign - 1)); - assert(filterSize > 0); + av_assert0(filterSize > 0); filter = av_malloc(filterSize * dstW * sizeof(*filter)); if (filterSize >= MAX_FILTER_SIZE * 16 / - ((flags & SWS_ACCURATE_RND) ? APCK_SIZE : 16) || !filter) + ((flags & SWS_ACCURATE_RND) ? APCK_SIZE : 16) || !filter) { + av_log(NULL, AV_LOG_ERROR, "sws: filterSize %d is too large, try less extreem scaling or increase MAX_FILTER_SIZE and recompile\n", filterSize); goto fail; + } *outFilterSize = filterSize; if (flags & SWS_PRINT_INFO) @@ -541,29 +583,27 @@ static av_cold int initFilter(int16_t **outFilter, int32_t **filterPos, // FIXME try to align filterPos if possible // fix borders - if (is_horizontal) { - for (i = 0; i < dstW; i++) { - int j; - if ((*filterPos)[i] < 0) { - // move filter coefficients left to compensate for filterPos - for (j = 1; j < filterSize; j++) { - int left = FFMAX(j + (*filterPos)[i], 0); - filter[i * filterSize + left] += filter[i * filterSize + j]; - filter[i * filterSize + j] = 0; - } - (*filterPos)[i] = 0; + for (i = 0; i < dstW; i++) { + int j; + if ((*filterPos)[i] < 0) { + // move filter coefficients left to compensate for filterPos + for (j = 1; j < filterSize; j++) { + int left = FFMAX(j + (*filterPos)[i], 0); + filter[i * filterSize + left] += filter[i * filterSize + j]; + filter[i * filterSize + j] = 0; } + (*filterPos)[i]= 0; + } - if ((*filterPos)[i] + filterSize > srcW) { - int shift = (*filterPos)[i] + filterSize - srcW; - // move filter coefficients right to compensate for filterPos - for (j = filterSize - 2; j >= 0; j--) { - int right = FFMIN(j + shift, filterSize - 1); - filter[i * filterSize + right] += filter[i * filterSize + j]; - filter[i * filterSize + j] = 0; - } - (*filterPos)[i] = srcW - filterSize; + if ((*filterPos)[i] + filterSize > srcW) { + int shift = (*filterPos)[i] + filterSize - srcW; + // move filter coefficients right to compensate for filterPos + for (j = filterSize - 2; j >= 0; j--) { + int right = FFMIN(j + shift, filterSize - 1); + filter[i * filterSize + right] += filter[i * filterSize + j]; + filter[i * filterSize + j] = 0; } + (*filterPos)[i]= srcW - filterSize; } } @@ -604,6 +644,8 @@ static av_cold int initFilter(int16_t **outFilter, int32_t **filterPos, ret = 0; fail: + if(ret < 0) + av_log(NULL, AV_LOG_ERROR, "sws: initFilter failed\n"); av_free(filter); av_free(filter2); return ret; @@ -782,26 +824,169 @@ static void getSubSampleFactors(int *h, int *v, enum AVPixelFormat format) *v = desc->log2_chroma_h; } +static void fill_rgb2yuv_table(SwsContext *c, const int table[4], int dstRange) +{ + int64_t W, V, Z, Cy, Cu, Cv; + int64_t vr = table[0]; + int64_t ub = table[1]; + int64_t ug = -table[2]; + int64_t vg = -table[3]; + int64_t ONE = 65536; + int64_t cy = ONE; + uint8_t *p = (uint8_t*)c->input_rgb2yuv_table; + int i; + static const int8_t map[] = { + BY_IDX, GY_IDX, -1 , BY_IDX, BY_IDX, GY_IDX, -1 , BY_IDX, + RY_IDX, -1 , GY_IDX, RY_IDX, RY_IDX, -1 , GY_IDX, RY_IDX, + RY_IDX, GY_IDX, -1 , RY_IDX, RY_IDX, GY_IDX, -1 , RY_IDX, + BY_IDX, -1 , GY_IDX, BY_IDX, BY_IDX, -1 , GY_IDX, BY_IDX, + BU_IDX, GU_IDX, -1 , BU_IDX, BU_IDX, GU_IDX, -1 , BU_IDX, + RU_IDX, -1 , GU_IDX, RU_IDX, RU_IDX, -1 , GU_IDX, RU_IDX, + RU_IDX, GU_IDX, -1 , RU_IDX, RU_IDX, GU_IDX, -1 , RU_IDX, + BU_IDX, -1 , GU_IDX, BU_IDX, BU_IDX, -1 , GU_IDX, BU_IDX, + BV_IDX, GV_IDX, -1 , BV_IDX, BV_IDX, GV_IDX, -1 , BV_IDX, + RV_IDX, -1 , GV_IDX, RV_IDX, RV_IDX, -1 , GV_IDX, RV_IDX, + RV_IDX, GV_IDX, -1 , RV_IDX, RV_IDX, GV_IDX, -1 , RV_IDX, + BV_IDX, -1 , GV_IDX, BV_IDX, BV_IDX, -1 , GV_IDX, BV_IDX, + RY_IDX, BY_IDX, RY_IDX, BY_IDX, RY_IDX, BY_IDX, RY_IDX, BY_IDX, + BY_IDX, RY_IDX, BY_IDX, RY_IDX, BY_IDX, RY_IDX, BY_IDX, RY_IDX, + GY_IDX, -1 , GY_IDX, -1 , GY_IDX, -1 , GY_IDX, -1 , + -1 , GY_IDX, -1 , GY_IDX, -1 , GY_IDX, -1 , GY_IDX, + RU_IDX, BU_IDX, RU_IDX, BU_IDX, RU_IDX, BU_IDX, RU_IDX, BU_IDX, + BU_IDX, RU_IDX, BU_IDX, RU_IDX, BU_IDX, RU_IDX, BU_IDX, RU_IDX, + GU_IDX, -1 , GU_IDX, -1 , GU_IDX, -1 , GU_IDX, -1 , + -1 , GU_IDX, -1 , GU_IDX, -1 , GU_IDX, -1 , GU_IDX, + RV_IDX, BV_IDX, RV_IDX, BV_IDX, RV_IDX, BV_IDX, RV_IDX, BV_IDX, + BV_IDX, RV_IDX, BV_IDX, RV_IDX, BV_IDX, RV_IDX, BV_IDX, RV_IDX, + GV_IDX, -1 , GV_IDX, -1 , GV_IDX, -1 , GV_IDX, -1 , + -1 , GV_IDX, -1 , GV_IDX, -1 , GV_IDX, -1 , GV_IDX, //23 + -1 , -1 , -1 , -1 , -1 , -1 , -1 , -1 , //24 + -1 , -1 , -1 , -1 , -1 , -1 , -1 , -1 , //25 + -1 , -1 , -1 , -1 , -1 , -1 , -1 , -1 , //26 + -1 , -1 , -1 , -1 , -1 , -1 , -1 , -1 , //27 + -1 , -1 , -1 , -1 , -1 , -1 , -1 , -1 , //28 + -1 , -1 , -1 , -1 , -1 , -1 , -1 , -1 , //29 + -1 , -1 , -1 , -1 , -1 , -1 , -1 , -1 , //30 + -1 , -1 , -1 , -1 , -1 , -1 , -1 , -1 , //31 + BY_IDX, GY_IDX, RY_IDX, -1 , -1 , -1 , -1 , -1 , //32 + BU_IDX, GU_IDX, RU_IDX, -1 , -1 , -1 , -1 , -1 , //33 + BV_IDX, GV_IDX, RV_IDX, -1 , -1 , -1 , -1 , -1 , //34 + }; + + dstRange = 0; //FIXME range = 1 is handled elsewhere + + if (!dstRange) { + cy = cy * 255 / 219; + } else { + vr = vr * 224 / 255; + ub = ub * 224 / 255; + ug = ug * 224 / 255; + vg = vg * 224 / 255; + } + W = ROUNDED_DIV(ONE*ONE*ug, ub); + V = ROUNDED_DIV(ONE*ONE*vg, vr); + Z = ONE*ONE-W-V; + + Cy = ROUNDED_DIV(cy*Z, ONE); + Cu = ROUNDED_DIV(ub*Z, ONE); + Cv = ROUNDED_DIV(vr*Z, ONE); + + c->input_rgb2yuv_table[RY_IDX] = -ROUNDED_DIV((1 << RGB2YUV_SHIFT)*V , Cy); + c->input_rgb2yuv_table[GY_IDX] = ROUNDED_DIV((1 << RGB2YUV_SHIFT)*ONE*ONE , Cy); + c->input_rgb2yuv_table[BY_IDX] = -ROUNDED_DIV((1 << RGB2YUV_SHIFT)*W , Cy); + + c->input_rgb2yuv_table[RU_IDX] = ROUNDED_DIV((1 << RGB2YUV_SHIFT)*V , Cu); + c->input_rgb2yuv_table[GU_IDX] = -ROUNDED_DIV((1 << RGB2YUV_SHIFT)*ONE*ONE , Cu); + c->input_rgb2yuv_table[BU_IDX] = ROUNDED_DIV((1 << RGB2YUV_SHIFT)*(Z+W) , Cu); + + c->input_rgb2yuv_table[RV_IDX] = ROUNDED_DIV((1 << RGB2YUV_SHIFT)*(V+Z) , Cv); + c->input_rgb2yuv_table[GV_IDX] = -ROUNDED_DIV((1 << RGB2YUV_SHIFT)*ONE*ONE , Cv); + c->input_rgb2yuv_table[BV_IDX] = ROUNDED_DIV((1 << RGB2YUV_SHIFT)*W , Cv); + + if(/*!dstRange && */!memcmp(table, ff_yuv2rgb_coeffs[SWS_CS_DEFAULT], sizeof(ff_yuv2rgb_coeffs[SWS_CS_DEFAULT]))) { + c->input_rgb2yuv_table[BY_IDX] = ((int)(0.114 * 219 / 255 * (1 << RGB2YUV_SHIFT) + 0.5)); + c->input_rgb2yuv_table[BV_IDX] = (-(int)(0.081 * 224 / 255 * (1 << RGB2YUV_SHIFT) + 0.5)); + c->input_rgb2yuv_table[BU_IDX] = ((int)(0.500 * 224 / 255 * (1 << RGB2YUV_SHIFT) + 0.5)); + c->input_rgb2yuv_table[GY_IDX] = ((int)(0.587 * 219 / 255 * (1 << RGB2YUV_SHIFT) + 0.5)); + c->input_rgb2yuv_table[GV_IDX] = (-(int)(0.419 * 224 / 255 * (1 << RGB2YUV_SHIFT) + 0.5)); + c->input_rgb2yuv_table[GU_IDX] = (-(int)(0.331 * 224 / 255 * (1 << RGB2YUV_SHIFT) + 0.5)); + c->input_rgb2yuv_table[RY_IDX] = ((int)(0.299 * 219 / 255 * (1 << RGB2YUV_SHIFT) + 0.5)); + c->input_rgb2yuv_table[RV_IDX] = ((int)(0.500 * 224 / 255 * (1 << RGB2YUV_SHIFT) + 0.5)); + c->input_rgb2yuv_table[RU_IDX] = (-(int)(0.169 * 224 / 255 * (1 << RGB2YUV_SHIFT) + 0.5)); + } + for(i=0; i<FF_ARRAY_ELEMS(map); i++) + AV_WL16(p + 16*4 + 2*i, map[i] >= 0 ? c->input_rgb2yuv_table[map[i]] : 0); +} + +static void fill_xyztables(struct SwsContext *c) +{ + int i; + double xyzgamma = XYZ_GAMMA; + double rgbgamma = 1.0 / RGB_GAMMA; + double xyzgammainv = 1.0 / XYZ_GAMMA; + double rgbgammainv = RGB_GAMMA; + static const int16_t xyz2rgb_matrix[3][4] = { + {13270, -6295, -2041}, + {-3969, 7682, 170}, + { 228, -835, 4329} }; + static const int16_t rgb2xyz_matrix[3][4] = { + {1689, 1464, 739}, + { 871, 2929, 296}, + { 79, 488, 3891} }; + static int16_t xyzgamma_tab[4096], rgbgamma_tab[4096], xyzgammainv_tab[4096], rgbgammainv_tab[4096]; + + memcpy(c->xyz2rgb_matrix, xyz2rgb_matrix, sizeof(c->xyz2rgb_matrix)); + memcpy(c->rgb2xyz_matrix, rgb2xyz_matrix, sizeof(c->rgb2xyz_matrix)); + c->xyzgamma = xyzgamma_tab; + c->rgbgamma = rgbgamma_tab; + c->xyzgammainv = xyzgammainv_tab; + c->rgbgammainv = rgbgammainv_tab; + + if (rgbgamma_tab[4095]) + return; + + /* set gamma vectors */ + for (i = 0; i < 4096; i++) { + xyzgamma_tab[i] = lrint(pow(i / 4095.0, xyzgamma) * 4095.0); + rgbgamma_tab[i] = lrint(pow(i / 4095.0, rgbgamma) * 4095.0); + xyzgammainv_tab[i] = lrint(pow(i / 4095.0, xyzgammainv) * 4095.0); + rgbgammainv_tab[i] = lrint(pow(i / 4095.0, rgbgammainv) * 4095.0); + } +} + int sws_setColorspaceDetails(struct SwsContext *c, const int inv_table[4], int srcRange, const int table[4], int dstRange, int brightness, int contrast, int saturation) { - const AVPixFmtDescriptor *desc_dst = av_pix_fmt_desc_get(c->dstFormat); - const AVPixFmtDescriptor *desc_src = av_pix_fmt_desc_get(c->srcFormat); + const AVPixFmtDescriptor *desc_dst; + const AVPixFmtDescriptor *desc_src; memcpy(c->srcColorspaceTable, inv_table, sizeof(int) * 4); memcpy(c->dstColorspaceTable, table, sizeof(int) * 4); + handle_formats(c); + desc_dst = av_pix_fmt_desc_get(c->dstFormat); + desc_src = av_pix_fmt_desc_get(c->srcFormat); + + if(!isYUV(c->dstFormat) && !isGray(c->dstFormat)) + dstRange = 0; + if(!isYUV(c->srcFormat) && !isGray(c->srcFormat)) + srcRange = 0; + c->brightness = brightness; c->contrast = contrast; c->saturation = saturation; c->srcRange = srcRange; c->dstRange = dstRange; - if (isYUV(c->dstFormat) || isGray(c->dstFormat)) + + fill_xyztables(c); + + if ((isYUV(c->dstFormat) || isGray(c->dstFormat)) && (isYUV(c->srcFormat) || isGray(c->srcFormat))) return -1; c->dstFormatBpp = av_get_bits_per_pixel(desc_dst); c->srcFormatBpp = av_get_bits_per_pixel(desc_src); + if (!isYUV(c->dstFormat) && !isGray(c->dstFormat)) { ff_yuv2rgb_c_init_tables(c, inv_table, srcRange, brightness, contrast, saturation); // FIXME factorize @@ -809,6 +994,10 @@ int sws_setColorspaceDetails(struct SwsContext *c, const int inv_table[4], if (ARCH_PPC) ff_yuv2rgb_init_tables_ppc(c, inv_table, brightness, contrast, saturation); + } + + fill_rgb2yuv_table(c, table, dstRange); + return 0; } @@ -816,7 +1005,7 @@ int sws_getColorspaceDetails(struct SwsContext *c, int **inv_table, int *srcRange, int **table, int *dstRange, int *brightness, int *contrast, int *saturation) { - if (isYUV(c->dstFormat) || isGray(c->dstFormat)) + if (!c ) return -1; *inv_table = c->srcColorspaceTable; @@ -836,6 +1025,9 @@ static int handle_jpeg(enum AVPixelFormat *format) case AV_PIX_FMT_YUVJ420P: *format = AV_PIX_FMT_YUV420P; return 1; + case AV_PIX_FMT_YUVJ411P: + *format = AV_PIX_FMT_YUV411P; + return 1; case AV_PIX_FMT_YUVJ422P: *format = AV_PIX_FMT_YUV422P; return 1; @@ -845,11 +1037,41 @@ static int handle_jpeg(enum AVPixelFormat *format) case AV_PIX_FMT_YUVJ440P: *format = AV_PIX_FMT_YUV440P; return 1; + case AV_PIX_FMT_GRAY8: + return 1; default: return 0; } } +static int handle_0alpha(enum AVPixelFormat *format) +{ + switch (*format) { + case AV_PIX_FMT_0BGR : *format = AV_PIX_FMT_ABGR ; return 1; + case AV_PIX_FMT_BGR0 : *format = AV_PIX_FMT_BGRA ; return 4; + case AV_PIX_FMT_0RGB : *format = AV_PIX_FMT_ARGB ; return 1; + case AV_PIX_FMT_RGB0 : *format = AV_PIX_FMT_RGBA ; return 4; + default: return 0; + } +} + +static int handle_xyz(enum AVPixelFormat *format) +{ + switch (*format) { + case AV_PIX_FMT_XYZ12BE : *format = AV_PIX_FMT_RGB48BE; return 1; + case AV_PIX_FMT_XYZ12LE : *format = AV_PIX_FMT_RGB48LE; return 1; + default: return 0; + } +} + +static void handle_formats(SwsContext *c) +{ + c->src0Alpha |= handle_0alpha(&c->srcFormat); + c->dst0Alpha |= handle_0alpha(&c->dstFormat); + c->srcXYZ |= handle_xyz(&c->srcFormat); + c->dstXYZ |= handle_xyz(&c->dstFormat); +} + SwsContext *sws_alloc_context(void) { SwsContext *c = av_mallocz(sizeof(SwsContext)); @@ -865,7 +1087,7 @@ SwsContext *sws_alloc_context(void) av_cold int sws_init_context(SwsContext *c, SwsFilter *srcFilter, SwsFilter *dstFilter) { - int i; + int i, j; int usesVFilter, usesHFilter; int unscaled; SwsFilter dummyFilter = { NULL, NULL, NULL, NULL }; @@ -873,13 +1095,12 @@ av_cold int sws_init_context(SwsContext *c, SwsFilter *srcFilter, int srcH = c->srcH; int dstW = c->dstW; int dstH = c->dstH; - int dst_stride = FFALIGN(dstW * sizeof(int16_t) + 16, 16); - int dst_stride_px = dst_stride >> 1; + int dst_stride = FFALIGN(dstW * sizeof(int16_t) + 66, 16); int flags, cpu_flags; enum AVPixelFormat srcFormat = c->srcFormat; enum AVPixelFormat dstFormat = c->dstFormat; - const AVPixFmtDescriptor *desc_src = av_pix_fmt_desc_get(srcFormat); - const AVPixFmtDescriptor *desc_dst = av_pix_fmt_desc_get(dstFormat); + const AVPixFmtDescriptor *desc_src; + const AVPixFmtDescriptor *desc_dst; cpu_flags = av_get_cpu_flags(); flags = c->flags; @@ -889,16 +1110,32 @@ av_cold int sws_init_context(SwsContext *c, SwsFilter *srcFilter, unscaled = (srcW == dstW && srcH == dstH); + c->srcRange |= handle_jpeg(&c->srcFormat); + c->dstRange |= handle_jpeg(&c->dstFormat); + + if (!c->contrast && !c->saturation && !c->dstFormatBpp) + sws_setColorspaceDetails(c, ff_yuv2rgb_coeffs[SWS_CS_DEFAULT], c->srcRange, + ff_yuv2rgb_coeffs[SWS_CS_DEFAULT], + c->dstRange, 0, 1 << 16, 1 << 16); + + if(srcFormat!=c->srcFormat || dstFormat!=c->dstFormat) + av_log(c, AV_LOG_WARNING, "deprecated pixel format used, make sure you did set range correctly\n"); + handle_formats(c); + srcFormat = c->srcFormat; + dstFormat = c->dstFormat; + desc_src = av_pix_fmt_desc_get(srcFormat); + desc_dst = av_pix_fmt_desc_get(dstFormat); + if (!(unscaled && sws_isSupportedEndiannessConversion(srcFormat) && av_pix_fmt_swap_endianness(srcFormat) == dstFormat)) { if (!sws_isSupportedInput(srcFormat)) { av_log(c, AV_LOG_ERROR, "%s is not supported as input pixel format\n", - sws_format_name(srcFormat)); + av_get_pix_fmt_name(srcFormat)); return AVERROR(EINVAL); } if (!sws_isSupportedOutput(dstFormat)) { av_log(c, AV_LOG_ERROR, "%s is not supported as output pixel format\n", - sws_format_name(dstFormat)); + av_get_pix_fmt_name(dstFormat)); return AVERROR(EINVAL); } } @@ -915,12 +1152,11 @@ av_cold int sws_init_context(SwsContext *c, SwsFilter *srcFilter, SWS_SPLINE | SWS_BICUBLIN); if (!i || (i & (i - 1))) { - av_log(c, AV_LOG_ERROR, - "Exactly one scaler algorithm must be chosen\n"); + av_log(c, AV_LOG_ERROR, "Exactly one scaler algorithm must be chosen, got %X\n", i); return AVERROR(EINVAL); } /* sanity check */ - if (srcW < 4 || srcH < 1 || dstW < 8 || dstH < 1) { + if (srcW < 1 || srcH < 1 || dstW < 1 || dstH < 1) { /* FIXME check if these are enough and try to lower them after * fixing the relevant parts of the code */ av_log(c, AV_LOG_ERROR, "%dx%d -> %dx%d is invalid scaling dimension\n", @@ -951,6 +1187,43 @@ av_cold int sws_init_context(SwsContext *c, SwsFilter *srcFilter, getSubSampleFactors(&c->chrSrcHSubSample, &c->chrSrcVSubSample, srcFormat); getSubSampleFactors(&c->chrDstHSubSample, &c->chrDstVSubSample, dstFormat); + if (isAnyRGB(dstFormat) && !(flags&SWS_FULL_CHR_H_INT)) { + if (dstW&1) { + av_log(c, AV_LOG_DEBUG, "Forcing full internal H chroma due to odd output size\n"); + flags |= SWS_FULL_CHR_H_INT; + c->flags = flags; + } + } + + if (c->dither == SWS_DITHER_AUTO) { + if (flags & SWS_ERROR_DIFFUSION) + c->dither = SWS_DITHER_ED; + } + + if(dstFormat == AV_PIX_FMT_BGR4_BYTE || + dstFormat == AV_PIX_FMT_RGB4_BYTE || + dstFormat == AV_PIX_FMT_BGR8 || + dstFormat == AV_PIX_FMT_RGB8) { + if (c->dither == SWS_DITHER_AUTO) + c->dither = (flags & SWS_FULL_CHR_H_INT) ? SWS_DITHER_ED : SWS_DITHER_BAYER; + if (!(flags & SWS_FULL_CHR_H_INT)) { + if (c->dither == SWS_DITHER_ED) { + av_log(c, AV_LOG_DEBUG, + "Desired dithering only supported in full chroma interpolation for destination format '%s'\n", + av_get_pix_fmt_name(dstFormat)); + flags |= SWS_FULL_CHR_H_INT; + c->flags = flags; + } + } + if (flags & SWS_FULL_CHR_H_INT) { + if (c->dither == SWS_DITHER_BAYER) { + av_log(c, AV_LOG_DEBUG, + "Ordered dither is not supported in full chroma interpolation for destination format '%s'\n", + av_get_pix_fmt_name(dstFormat)); + c->dither = SWS_DITHER_ED; + } + } + } if (isPlanarRGB(dstFormat)) { if (!(flags & SWS_FULL_CHR_H_INT)) { av_log(c, AV_LOG_DEBUG, @@ -971,10 +1244,15 @@ av_cold int sws_init_context(SwsContext *c, SwsFilter *srcFilter, dstFormat != AV_PIX_FMT_BGRA && dstFormat != AV_PIX_FMT_ABGR && dstFormat != AV_PIX_FMT_RGB24 && - dstFormat != AV_PIX_FMT_BGR24) { - av_log(c, AV_LOG_ERROR, + dstFormat != AV_PIX_FMT_BGR24 && + dstFormat != AV_PIX_FMT_BGR4_BYTE && + dstFormat != AV_PIX_FMT_RGB4_BYTE && + dstFormat != AV_PIX_FMT_BGR8 && + dstFormat != AV_PIX_FMT_RGB8 + ) { + av_log(c, AV_LOG_WARNING, "full chroma interpolation for destination format '%s' not yet implemented\n", - sws_format_name(dstFormat)); + av_get_pix_fmt_name(dstFormat)); flags &= ~SWS_FULL_CHR_H_INT; c->flags = flags; } @@ -994,16 +1272,20 @@ av_cold int sws_init_context(SwsContext *c, SwsFilter *srcFilter, srcFormat != AV_PIX_FMT_RGB4_BYTE && srcFormat != AV_PIX_FMT_BGR4_BYTE && srcFormat != AV_PIX_FMT_GBRP9BE && srcFormat != AV_PIX_FMT_GBRP9LE && srcFormat != AV_PIX_FMT_GBRP10BE && srcFormat != AV_PIX_FMT_GBRP10LE && + srcFormat != AV_PIX_FMT_GBRP12BE && srcFormat != AV_PIX_FMT_GBRP12LE && + srcFormat != AV_PIX_FMT_GBRP14BE && srcFormat != AV_PIX_FMT_GBRP14LE && srcFormat != AV_PIX_FMT_GBRP16BE && srcFormat != AV_PIX_FMT_GBRP16LE && ((dstW >> c->chrDstHSubSample) <= (srcW >> 1) || (flags & SWS_FAST_BILINEAR))) c->chrSrcHSubSample = 1; - // Note the -((-x)>>y) is so that we always round toward +inf. - c->chrSrcW = -((-srcW) >> c->chrSrcHSubSample); - c->chrSrcH = -((-srcH) >> c->chrSrcVSubSample); - c->chrDstW = -((-dstW) >> c->chrDstHSubSample); - c->chrDstH = -((-dstH) >> c->chrDstVSubSample); + // Note the FF_CEIL_RSHIFT is so that we always round toward +inf. + c->chrSrcW = FF_CEIL_RSHIFT(srcW, c->chrSrcHSubSample); + c->chrSrcH = FF_CEIL_RSHIFT(srcH, c->chrSrcVSubSample); + c->chrDstW = FF_CEIL_RSHIFT(dstW, c->chrDstHSubSample); + c->chrDstH = FF_CEIL_RSHIFT(dstH, c->chrDstVSubSample); + + FF_ALLOC_OR_GOTO(c, c->formatConvBuffer, FFALIGN(srcW*2+78, 16) * 2, fail); /* unscaled special cases */ if (unscaled && !usesHFilter && !usesVFilter && @@ -1014,7 +1296,7 @@ av_cold int sws_init_context(SwsContext *c, SwsFilter *srcFilter, if (flags & SWS_PRINT_INFO) av_log(c, AV_LOG_INFO, "using unscaled %s -> %s special converter\n", - sws_format_name(srcFormat), sws_format_name(dstFormat)); + av_get_pix_fmt_name(srcFormat), av_get_pix_fmt_name(dstFormat)); return 0; } } @@ -1025,21 +1307,22 @@ av_cold int sws_init_context(SwsContext *c, SwsFilter *srcFilter, c->dstBpc = 1 + desc_dst->comp[0].depth_minus1; if (c->dstBpc < 8) c->dstBpc = 8; + if (isAnyRGB(srcFormat) || srcFormat == AV_PIX_FMT_PAL8) + c->srcBpc = 16; if (c->dstBpc == 16) dst_stride <<= 1; - FF_ALLOC_OR_GOTO(c, c->formatConvBuffer, - (FFALIGN(srcW, 16) * 2 * FFALIGN(c->srcBpc, 8) >> 3) + 16, - fail); - if (INLINE_MMXEXT(cpu_flags) && c->srcBpc == 8 && c->dstBpc <= 10) { + + if (INLINE_MMXEXT(cpu_flags) && c->srcBpc == 8 && c->dstBpc <= 14) { c->canMMXEXTBeUsed = (dstW >= srcW && (dstW & 31) == 0 && (srcW & 15) == 0) ? 1 : 0; if (!c->canMMXEXTBeUsed && dstW >= srcW && (srcW & 15) == 0 + && (flags & SWS_FAST_BILINEAR)) { if (flags & SWS_PRINT_INFO) av_log(c, AV_LOG_INFO, "output width is not a multiple of 32 -> no MMXEXT scaler\n"); } - if (usesHFilter) + if (usesHFilter || isNBPS(c->srcFormat) || is16BPS(c->srcFormat) || isAnyRGB(c->srcFormat)) c->canMMXEXTBeUsed = 0; } else c->canMMXEXTBeUsed = 0; @@ -1060,7 +1343,7 @@ av_cold int sws_init_context(SwsContext *c, SwsFilter *srcFilter, c->chrXInc += 20; } // we don't use the x86 asm scaler if MMX is available - else if (INLINE_MMX(cpu_flags)) { + else if (INLINE_MMX(cpu_flags) && c->dstBpc <= 14) { c->lumXInc = ((int64_t)(srcW - 2) << 16) / (dstW - 2) - 20; c->chrXInc = ((int64_t)(c->chrSrcW - 2) << 16) / (c->chrDstW - 2) - 20; } @@ -1101,17 +1384,25 @@ av_cold int sws_init_context(SwsContext *c, SwsFilter *srcFilter, c->chrMmxextFilterCode = av_malloc(c->chrMmxextFilterCodeSize); #endif +#ifdef MAP_ANONYMOUS + if (c->lumMmxextFilterCode == MAP_FAILED || c->chrMmxextFilterCode == MAP_FAILED) +#else if (!c->lumMmxextFilterCode || !c->chrMmxextFilterCode) +#endif + { + av_log(c, AV_LOG_ERROR, "Failed to allocate MMX2FilterCode\n"); return AVERROR(ENOMEM); + } + FF_ALLOCZ_OR_GOTO(c, c->hLumFilter, (dstW / 8 + 8) * sizeof(int16_t), fail); FF_ALLOCZ_OR_GOTO(c, c->hChrFilter, (c->chrDstW / 4 + 8) * sizeof(int16_t), fail); FF_ALLOCZ_OR_GOTO(c, c->hLumFilterPos, (dstW / 2 / 8 + 8) * sizeof(int32_t), fail); FF_ALLOCZ_OR_GOTO(c, c->hChrFilterPos, (c->chrDstW / 2 / 4 + 8) * sizeof(int32_t), fail); - init_hscaler_mmxext(dstW, c->lumXInc, c->lumMmxextFilterCode, - c->hLumFilter, c->hLumFilterPos, 8); + init_hscaler_mmxext( dstW, c->lumXInc, c->lumMmxextFilterCode, + c->hLumFilter, (uint32_t*)c->hLumFilterPos, 8); init_hscaler_mmxext(c->chrDstW, c->chrXInc, c->chrMmxextFilterCode, - c->hChrFilter, c->hChrFilterPos, 4); + c->hChrFilter, (uint32_t*)c->hChrFilterPos, 4); #if USE_MMAP mprotect(c->lumMmxextFilterCode, c->lumMmxextFilterCodeSize, PROT_EXEC | PROT_READ); @@ -1128,14 +1419,18 @@ av_cold int sws_init_context(SwsContext *c, SwsFilter *srcFilter, srcW, dstW, filterAlign, 1 << 14, (flags & SWS_BICUBLIN) ? (flags | SWS_BICUBIC) : flags, cpu_flags, srcFilter->lumH, dstFilter->lumH, - c->param, 1) < 0) + c->param, + get_local_pos(c, 0, 0, 0), + get_local_pos(c, 0, 0, 0)) < 0) goto fail; if (initFilter(&c->hChrFilter, &c->hChrFilterPos, &c->hChrFilterSize, c->chrXInc, c->chrSrcW, c->chrDstW, filterAlign, 1 << 14, (flags & SWS_BICUBLIN) ? (flags | SWS_BILINEAR) : flags, cpu_flags, srcFilter->chrH, dstFilter->chrH, - c->param, 1) < 0) + c->param, + get_local_pos(c, c->chrSrcHSubSample, c->src_h_chr_pos, 0), + get_local_pos(c, c->chrDstHSubSample, c->dst_h_chr_pos, 0)) < 0) goto fail; } } // initialize horizontal stuff @@ -1149,14 +1444,19 @@ av_cold int sws_init_context(SwsContext *c, SwsFilter *srcFilter, c->lumYInc, srcH, dstH, filterAlign, (1 << 12), (flags & SWS_BICUBLIN) ? (flags | SWS_BICUBIC) : flags, cpu_flags, srcFilter->lumV, dstFilter->lumV, - c->param, 0) < 0) + c->param, + get_local_pos(c, 0, 0, 1), + get_local_pos(c, 0, 0, 1)) < 0) goto fail; if (initFilter(&c->vChrFilter, &c->vChrFilterPos, &c->vChrFilterSize, c->chrYInc, c->chrSrcH, c->chrDstH, filterAlign, (1 << 12), (flags & SWS_BICUBLIN) ? (flags | SWS_BILINEAR) : flags, cpu_flags, srcFilter->chrV, dstFilter->chrV, - c->param, 0) < 0) + c->param, + get_local_pos(c, c->chrSrcVSubSample, c->src_v_chr_pos, 1), + get_local_pos(c, c->chrDstVSubSample, c->dst_v_chr_pos, 1)) < 0) + goto fail; #if HAVE_ALTIVEC @@ -1198,6 +1498,9 @@ av_cold int sws_init_context(SwsContext *c, SwsFilter *srcFilter, c->vChrFilterPos[chrI]; } + for (i = 0; i < 4; i++) + FF_ALLOCZ_OR_GOTO(c, c->dither_error[i], (c->dstW+2) * sizeof(int), fail); + /* Allocate pixbufs (we use dynamic allocation because otherwise we would * need to allocate several megabytes to handle all possible cases) */ FF_ALLOC_OR_GOTO(c, c->lumPixBuf, c->vLumBufSize * 3 * sizeof(int16_t *), fail); @@ -1213,9 +1516,9 @@ av_cold int sws_init_context(SwsContext *c, SwsFilter *srcFilter, dst_stride + 16, fail); c->lumPixBuf[i] = c->lumPixBuf[i + c->vLumBufSize]; } - // 64 / (c->dstBpc & ~7) is the same as 16 / sizeof(scaling_intermediate) - c->uv_off_px = dst_stride_px + 64 / (c->dstBpc & ~7); - c->uv_off_byte = dst_stride + 16; + // 64 / c->scalingBpp is the same as 16 / sizeof(scaling_intermediate) + c->uv_off = (dst_stride>>1) + 64 / (c->dstBpc &~ 7); + c->uv_offx2 = dst_stride + 16; for (i = 0; i < c->vChrBufSize; i++) { FF_ALLOC_OR_GOTO(c, c->chrUPixBuf[i + c->vChrBufSize], dst_stride * 2 + 32, fail); @@ -1232,9 +1535,15 @@ av_cold int sws_init_context(SwsContext *c, SwsFilter *srcFilter, // try to avoid drawing green stuff between the right end and the stride end for (i = 0; i < c->vChrBufSize; i++) - memset(c->chrUPixBuf[i], 64, dst_stride * 2 + 1); + if(desc_dst->comp[0].depth_minus1 == 15){ + av_assert0(c->dstBpc > 14); + for(j=0; j<dst_stride/2+1; j++) + ((int32_t*)(c->chrUPixBuf[i]))[j] = 1<<18; + } else + for(j=0; j<dst_stride+1; j++) + ((int16_t*)(c->chrUPixBuf[i]))[j] = 1<<14; - assert(c->chrDstH <= dstH); + av_assert0(c->chrDstH <= dstH); if (flags & SWS_PRINT_INFO) { if (flags & SWS_FAST_BILINEAR) @@ -1263,7 +1572,7 @@ av_cold int sws_init_context(SwsContext *c, SwsFilter *srcFilter, av_log(c, AV_LOG_INFO, "ehh flags invalid?! "); av_log(c, AV_LOG_INFO, "from %s to %s%s ", - sws_format_name(srcFormat), + av_get_pix_fmt_name(srcFormat), #ifdef DITHER1XBPP dstFormat == AV_PIX_FMT_BGR555 || dstFormat == AV_PIX_FMT_BGR565 || dstFormat == AV_PIX_FMT_RGB444BE || dstFormat == AV_PIX_FMT_RGB444LE || @@ -1272,7 +1581,7 @@ av_cold int sws_init_context(SwsContext *c, SwsFilter *srcFilter, #else "", #endif - sws_format_name(dstFormat)); + av_get_pix_fmt_name(dstFormat)); if (INLINE_MMXEXT(cpu_flags)) av_log(c, AV_LOG_INFO, "using MMXEXT\n"); @@ -1317,8 +1626,6 @@ SwsContext *sws_getContext(int srcW, int srcH, enum AVPixelFormat srcFormat, c->srcH = srcH; c->dstW = dstW; c->dstH = dstH; - c->srcRange = handle_jpeg(&srcFormat); - c->dstRange = handle_jpeg(&dstFormat); c->srcFormat = srcFormat; c->dstFormat = dstFormat; @@ -1326,9 +1633,6 @@ SwsContext *sws_getContext(int srcW, int srcH, enum AVPixelFormat srcFormat, c->param[0] = param[0]; c->param[1] = param[1]; } - sws_setColorspaceDetails(c, ff_yuv2rgb_coeffs[SWS_CS_DEFAULT], c->srcRange, - ff_yuv2rgb_coeffs[SWS_CS_DEFAULT] /* FIXME*/, - c->dstRange, 0, 1 << 16, 1 << 16); if (sws_init_context(c, srcFilter, dstFilter) < 0) { sws_freeContext(c); @@ -1403,7 +1707,12 @@ SwsFilter *sws_getDefaultFilter(float lumaGBlur, float chromaGBlur, SwsVector *sws_allocVec(int length) { - SwsVector *vec = av_malloc(sizeof(SwsVector)); + SwsVector *vec; + + if(length <= 0 || length > INT_MAX/ sizeof(double)) + return NULL; + + vec = av_malloc(sizeof(SwsVector)); if (!vec) return NULL; vec->length = length; @@ -1418,7 +1727,12 @@ SwsVector *sws_getGaussianVec(double variance, double quality) const int length = (int)(variance * quality + 0.5) | 1; int i; double middle = (length - 1) * 0.5; - SwsVector *vec = sws_allocVec(length); + SwsVector *vec; + + if(variance < 0 || quality < 0) + return NULL; + + vec = sws_allocVec(length); if (!vec) return NULL; @@ -1673,6 +1987,9 @@ void sws_freeContext(SwsContext *c) av_freep(&c->alpPixBuf); } + for (i = 0; i < 4; i++) + av_freep(&c->dither_error[i]); + av_freep(&c->vLumFilter); av_freep(&c->vChrFilter); av_freep(&c->hLumFilter); @@ -1707,7 +2024,7 @@ void sws_freeContext(SwsContext *c) #endif /* HAVE_MMX_INLINE */ av_freep(&c->yuvTable); - av_free(c->formatConvBuffer); + av_freep(&c->formatConvBuffer); av_free(c); } @@ -1745,19 +2062,13 @@ struct SwsContext *sws_getCachedContext(struct SwsContext *context, int srcW, return NULL; context->srcW = srcW; context->srcH = srcH; - context->srcRange = handle_jpeg(&srcFormat); context->srcFormat = srcFormat; context->dstW = dstW; context->dstH = dstH; - context->dstRange = handle_jpeg(&dstFormat); context->dstFormat = dstFormat; context->flags = flags; context->param[0] = param[0]; context->param[1] = param[1]; - sws_setColorspaceDetails(context, ff_yuv2rgb_coeffs[SWS_CS_DEFAULT], - context->srcRange, - ff_yuv2rgb_coeffs[SWS_CS_DEFAULT] /* FIXME*/, - context->dstRange, 0, 1 << 16, 1 << 16); if (sws_init_context(context, srcFilter, dstFilter) < 0) { sws_freeContext(context); return NULL; |