[FFmpeg-devel,v1,5/6] swscale: Add unscaled XRGB->YUV420P functions

Message ID	20230820151022.2204421-6-jc@kynesim.co.uk
State	New
Headers	show Delivered-To: ffmpegpatchwork2@gmail.com Received-SPF: pass (google.com: domain of ffmpeg-devel-bounces@ffmpeg.org designates 79.124.17.100 as permitted sender) client-ip=79.124.17.100; From: John Cox <jc@kynesim.co.uk> To: ffmpeg-devel@ffmpeg.org Date: Sun, 20 Aug 2023 15:10:21 +0000 Message-Id: <20230820151022.2204421-6-jc@kynesim.co.uk> In-Reply-To: <20230820151022.2204421-1-jc@kynesim.co.uk> References: <20230820151022.2204421-1-jc@kynesim.co.uk> MIME-Version: 1.0 Subject: [FFmpeg-devel] [PATCH v1 5/6] swscale: Add unscaled XRGB->YUV420P functions Precedence: list Reply-To: FFmpeg development discussions and patches <ffmpeg-devel@ffmpeg.org> Cc: John Cox <jc@kynesim.co.uk> Content-Type: text/plain; charset="us-ascii" Content-Transfer-Encoding: 7bit Errors-To: ffmpeg-devel-bounces@ffmpeg.org Sender: "ffmpeg-devel" <ffmpeg-devel-bounces@ffmpeg.org>
Series	swscale: Add dedicated RGB->YUV unscaled functions & aarch64 asm \| expand [FFmpeg-devel,v1,0/6] swscale: Add dedicated RGB->YUV unscaled functions & aarch64 asm [FFmpeg-devel,v1,1/6] fate-filter-fps: Set swscale bitexact for tests that do conversions [FFmpeg-devel,v1,2/6] swscale: Rename BGR24->YUV conversion functions as bgr... [FFmpeg-devel,v1,3/6] swscale: Add explicit rgb24->yv12 conversion [FFmpeg-devel,v1,4/6] swscale: RGB24->YUV allow odd widths & improve C rounding [FFmpeg-devel,v1,5/6] swscale: Add unscaled XRGB->YUV420P functions [FFmpeg-devel,v1,6/6] swscale: Add aarch64 functions for RGB24->YUV420P

Context	Check	Description
yinshiyou/make_loongarch64	success	Make finished
yinshiyou/make_fate_loongarch64	success	Make fate finished
andriy/make_x86	success	Make finished
andriy/make_fate_x86	success	Make fate finished

diff --git a/libswscale/rgb2rgb.c b/libswscale/rgb2rgb.c index de90e5193f..b976341e70 100644 --- a/libswscale/rgb2rgb.c +++ b/libswscale/rgb2rgb.c @@ -88,6 +88,26 @@ void (*ff_rgb24toyv12)(const uint8_t *src, uint8_t *ydst, int width, int height, int lumStride, int chromStride, int srcStride, int32_t *rgb2yuv); +void (*ff_rgbxtoyv12)(const uint8_t *src, uint8_t *ydst, + uint8_t *udst, uint8_t *vdst, + int width, int height, + int lumStride, int chromStride, int srcStride, + int32_t *rgb2yuv); +void (*ff_bgrxtoyv12)(const uint8_t *src, uint8_t *ydst, + uint8_t *udst, uint8_t *vdst, + int width, int height, + int lumStride, int chromStride, int srcStride, + int32_t *rgb2yuv); +void (*ff_xrgbtoyv12)(const uint8_t *src, uint8_t *ydst, + uint8_t *udst, uint8_t *vdst, + int width, int height, + int lumStride, int chromStride, int srcStride, + int32_t *rgb2yuv); +void (*ff_xbgrtoyv12)(const uint8_t *src, uint8_t *ydst, + uint8_t *udst, uint8_t *vdst, + int width, int height, + int lumStride, int chromStride, int srcStride, + int32_t *rgb2yuv); void (*planar2x)(const uint8_t *src, uint8_t *dst, int width, int height, int srcStride, int dstStride); void (*interleaveBytes)(const uint8_t *src1, const uint8_t *src2, uint8_t *dst, diff --git a/libswscale/rgb2rgb.h b/libswscale/rgb2rgb.h index f7a76a92ba..0015b1568a 100644 --- a/libswscale/rgb2rgb.h +++ b/libswscale/rgb2rgb.h @@ -135,6 +135,22 @@ extern void (*ff_rgb24toyv12)(const uint8_t *src, uint8_t *ydst, uint8_t *udst, int width, int height, int lumStride, int chromStride, int srcStride, int32_t *rgb2yuv); +extern void (*ff_rgbxtoyv12)(const uint8_t *src, uint8_t *ydst, uint8_t *udst, uint8_t *vdst, + int width, int height, + int lumStride, int chromStride, int srcStride, + int32_t *rgb2yuv); +extern void (*ff_bgrxtoyv12)(const uint8_t *src, uint8_t *ydst, uint8_t *udst, uint8_t *vdst, + int width, int height, + int lumStride, int chromStride, int srcStride, + int32_t *rgb2yuv); +extern void (*ff_xrgbtoyv12)(const uint8_t *src, uint8_t *ydst, uint8_t *udst, uint8_t *vdst, + int width, int height, + int lumStride, int chromStride, int srcStride, + int32_t *rgb2yuv); +extern void (*ff_xbgrtoyv12)(const uint8_t *src, uint8_t *ydst, uint8_t *udst, uint8_t *vdst, + int width, int height, + int lumStride, int chromStride, int srcStride, + int32_t *rgb2yuv); extern void (*planar2x)(const uint8_t *src, uint8_t *dst, int width, int height, int srcStride, int dstStride); diff --git a/libswscale/rgb2rgb_template.c b/libswscale/rgb2rgb_template.c index 5503e58a29..22326807c5 100644 --- a/libswscale/rgb2rgb_template.c +++ b/libswscale/rgb2rgb_template.c @@ -742,6 +742,108 @@ void ff_rgb24toyv12_c(const uint8_t *src, uint8_t *ydst, uint8_t *udst, rgb24toyv12_x(src, ydst, udst, vdst, width, height, lumStride, chromStride, srcStride, rgb2yuv, x_rgb); } +static void rgbxtoyv12_x(const uint8_t *src, uint8_t *ydst, uint8_t *udst, + uint8_t *vdst, int width, int height, int lumStride, + int chromStride, int srcStride, int32_t *rgb2yuv, + const uint8_t x[9]) +{ + int32_t ry = rgb2yuv[x[0]], gy = rgb2yuv[x[1]], by = rgb2yuv[x[2]]; + int32_t ru = rgb2yuv[x[3]], gu = rgb2yuv[x[4]], bu = rgb2yuv[x[5]]; + int32_t rv = rgb2yuv[x[6]], gv = rgb2yuv[x[7]], bv = rgb2yuv[x[8]]; + int y; + const int chromWidth = width >> 1; + // Constants with both rounding and offset + const int32_t ky = ((16 << 1) + 1) << (RGB2YUV_SHIFT - 1); + const int32_t kc = ((128 << 1) + 1) << (RGB2YUV_SHIFT - 1); + + for (y = 0; y < height; y += 2) { + int i; + for (i = 0; i < chromWidth; i++) { + unsigned int b = src[8 * i + 0]; + unsigned int g = src[8 * i + 1]; + unsigned int r = src[8 * i + 2]; + + unsigned int Y = (ry * r + gy * g + by * b + ky) >> RGB2YUV_SHIFT; + unsigned int V = (rv * r + gv * g + bv * b + kc) >> RGB2YUV_SHIFT; + unsigned int U = (ru * r + gu * g + bu * b + kc) >> RGB2YUV_SHIFT; + + udst[i] = U; + vdst[i] = V; + ydst[2 * i] = Y; + + b = src[8 * i + 4]; + g = src[8 * i + 5]; + r = src[8 * i + 6]; + + Y = ((ry * r + gy * g + by * b) >> RGB2YUV_SHIFT) + 16; + ydst[2 * i + 1] = Y; + } + if ((width & 1) != 0) { + unsigned int b = src[8 * i + 0]; + unsigned int g = src[8 * i + 1]; + unsigned int r = src[8 * i + 2]; + + unsigned int Y = (ry * r + gy * g + by * b + ky) >> RGB2YUV_SHIFT; + unsigned int V = (rv * r + gv * g + bv * b + kc) >> RGB2YUV_SHIFT; + unsigned int U = (ru * r + gu * g + bu * b + kc) >> RGB2YUV_SHIFT; + + udst[i] = U; + vdst[i] = V; + ydst[2 * i] = Y; + } + ydst += lumStride; + src += srcStride; + + if (y+1 == height) + break; + + for (i = 0; i < width; i++) { + unsigned int b = src[4 * i + 0]; + unsigned int g = src[4 * i + 1]; + unsigned int r = src[4 * i + 2]; + + unsigned int Y = (ry * r + gy * g + by * b + ky) >> RGB2YUV_SHIFT; + + ydst[i] = Y; + } + udst += chromStride; + vdst += chromStride; + ydst += lumStride; + src += srcStride; + } +} + +static void ff_rgbxtoyv12_c(const uint8_t *src, uint8_t *ydst, uint8_t *udst, + uint8_t *vdst, int width, int height, int lumStride, + int chromStride, int srcStride, int32_t *rgb2yuv) +{ + rgbxtoyv12_x(src, ydst, udst, vdst, width, height, lumStride, chromStride, srcStride, rgb2yuv, x_rgb); +} + +static void ff_bgrxtoyv12_c(const uint8_t *src, uint8_t *ydst, uint8_t *udst, + uint8_t *vdst, int width, int height, int lumStride, + int chromStride, int srcStride, int32_t *rgb2yuv) +{ + rgbxtoyv12_x(src, ydst, udst, vdst, width, height, lumStride, chromStride, srcStride, rgb2yuv, x_bgr); +} + +// As the general code does no SIMD-like ops simply adding 1 to the src address +// will fix the ignored alpha position +static void ff_xrgbtoyv12_c(const uint8_t *src, uint8_t *ydst, uint8_t *udst, + uint8_t *vdst, int width, int height, int lumStride, + int chromStride, int srcStride, int32_t *rgb2yuv) +{ + rgbxtoyv12_x(src + 1, ydst, udst, vdst, width, height, lumStride, chromStride, srcStride, rgb2yuv, x_rgb); +} + +static void ff_xbgrtoyv12_c(const uint8_t *src, uint8_t *ydst, uint8_t *udst, + uint8_t *vdst, int width, int height, int lumStride, + int chromStride, int srcStride, int32_t *rgb2yuv) +{ + rgbxtoyv12_x(src + 1, ydst, udst, vdst, width, height, lumStride, chromStride, srcStride, rgb2yuv, x_bgr); +} + + static void interleaveBytes_c(const uint8_t *src1, const uint8_t *src2, uint8_t *dest, int width, int height, int src1Stride, int src2Stride, int dstStride) @@ -1016,6 +1118,10 @@ static av_cold void rgb2rgb_init_c(void) planar2x = planar2x_c; ff_rgb24toyv12 = ff_rgb24toyv12_c; ff_bgr24toyv12 = ff_bgr24toyv12_c; + ff_rgbxtoyv12 = ff_rgbxtoyv12_c; + ff_bgrxtoyv12 = ff_bgrxtoyv12_c; + ff_xrgbtoyv12 = ff_xrgbtoyv12_c; + ff_xbgrtoyv12 = ff_xbgrtoyv12_c; interleaveBytes = interleaveBytes_c; deinterleaveBytes = deinterleaveBytes_c; vu9_to_vu12 = vu9_to_vu12_c; diff --git a/libswscale/swscale_unscaled.c b/libswscale/swscale_unscaled.c index e10f967755..ff682d367c 100644 --- a/libswscale/swscale_unscaled.c +++ b/libswscale/swscale_unscaled.c @@ -1671,6 +1671,74 @@ static int rgb24ToYv12Wrapper(SwsContext *c, const uint8_t *src[], return srcSliceH; } +static int bgrxToYv12Wrapper(SwsContext *c, const uint8_t *src[], + int srcStride[], int srcSliceY, int srcSliceH, + uint8_t *dst[], int dstStride[]) +{ + ff_bgrxtoyv12( + src[0], + dst[0] + srcSliceY * dstStride[0], + dst[1] + (srcSliceY >> 1) * dstStride[1], + dst[2] + (srcSliceY >> 1) * dstStride[2], + c->srcW, srcSliceH, + dstStride[0], dstStride[1], srcStride[0], + c->input_rgb2yuv_table); + if (dst[3]) + fillPlane(dst[3], dstStride[3], c->srcW, srcSliceH, srcSliceY, 255); + return srcSliceH; +} + +static int rgbxToYv12Wrapper(SwsContext *c, const uint8_t *src[], + int srcStride[], int srcSliceY, int srcSliceH, + uint8_t *dst[], int dstStride[]) +{ + ff_rgbxtoyv12( + src[0], + dst[0] + srcSliceY * dstStride[0], + dst[1] + (srcSliceY >> 1) * dstStride[1], + dst[2] + (srcSliceY >> 1) * dstStride[2], + c->srcW, srcSliceH, + dstStride[0], dstStride[1], srcStride[0], + c->input_rgb2yuv_table); + if (dst[3]) + fillPlane(dst[3], dstStride[3], c->srcW, srcSliceH, srcSliceY, 255); + return srcSliceH; +} + +static int xbgrToYv12Wrapper(SwsContext *c, const uint8_t *src[], + int srcStride[], int srcSliceY, int srcSliceH, + uint8_t *dst[], int dstStride[]) +{ + ff_xbgrtoyv12( + src[0], + dst[0] + srcSliceY * dstStride[0], + dst[1] + (srcSliceY >> 1) * dstStride[1], + dst[2] + (srcSliceY >> 1) * dstStride[2], + c->srcW, srcSliceH, + dstStride[0], dstStride[1], srcStride[0], + c->input_rgb2yuv_table); + if (dst[3]) + fillPlane(dst[3], dstStride[3], c->srcW, srcSliceH, srcSliceY, 255); + return srcSliceH; +} + +static int xrgbToYv12Wrapper(SwsContext *c, const uint8_t *src[], + int srcStride[], int srcSliceY, int srcSliceH, + uint8_t *dst[], int dstStride[]) +{ + ff_xrgbtoyv12( + src[0], + dst[0] + srcSliceY * dstStride[0], + dst[1] + (srcSliceY >> 1) * dstStride[1], + dst[2] + (srcSliceY >> 1) * dstStride[2], + c->srcW, srcSliceH, + dstStride[0], dstStride[1], srcStride[0], + c->input_rgb2yuv_table); + if (dst[3]) + fillPlane(dst[3], dstStride[3], c->srcW, srcSliceH, srcSliceY, 255); + return srcSliceH; +} + static int yvu9ToYv12Wrapper(SwsContext *c, const uint8_t *src[], int srcStride[], int srcSliceY, int srcSliceH, uint8_t *dst[], int dstStride[]) @@ -2059,6 +2127,27 @@ void ff_get_unscaled_swscale(SwsContext *c) !(flags & (SWS_ACCURATE_RND | SWS_BITEXACT))) c->convert_unscaled = rgb24ToYv12Wrapper; + /* bgrxtoYV12 */ + if (((srcFormat == AV_PIX_FMT_BGRA && dstFormat == AV_PIX_FMT_YUV420P) || + (srcFormat == AV_PIX_FMT_BGR0 && (dstFormat == AV_PIX_FMT_YUV420P || dstFormat == AV_PIX_FMT_YUVA420P))) && + !(flags & (SWS_ACCURATE_RND | SWS_BITEXACT))) + c->convert_unscaled = bgrxToYv12Wrapper; + /* rgbx24toYV12 */ + if (((srcFormat == AV_PIX_FMT_RGBA && dstFormat == AV_PIX_FMT_YUV420P) || + (srcFormat == AV_PIX_FMT_RGB0 && (dstFormat == AV_PIX_FMT_YUV420P || dstFormat == AV_PIX_FMT_YUVA420P))) && + !(flags & (SWS_ACCURATE_RND | SWS_BITEXACT))) + c->convert_unscaled = rgbxToYv12Wrapper; + /* xbgrtoYV12 */ + if (((srcFormat == AV_PIX_FMT_ABGR && dstFormat == AV_PIX_FMT_YUV420P) || + (srcFormat == AV_PIX_FMT_0BGR && (dstFormat == AV_PIX_FMT_YUV420P || dstFormat == AV_PIX_FMT_YUVA420P))) && + !(flags & (SWS_ACCURATE_RND | SWS_BITEXACT))) + c->convert_unscaled = xbgrToYv12Wrapper; + /* xrgb24toYV12 */ + if (((srcFormat == AV_PIX_FMT_ARGB && dstFormat == AV_PIX_FMT_YUV420P) || + (srcFormat == AV_PIX_FMT_0RGB && (dstFormat == AV_PIX_FMT_YUV420P || dstFormat == AV_PIX_FMT_YUVA420P))) && + !(flags & (SWS_ACCURATE_RND | SWS_BITEXACT))) + c->convert_unscaled = xrgbToYv12Wrapper; + /* RGB/BGR -> RGB/BGR (no dither needed forms) */ if (isAnyRGB(srcFormat) && isAnyRGB(dstFormat) && findRgbConvFn(c) && (!needsDither || (c->flags&(SWS_FAST_BILINEAR|SWS_POINT))))

[FFmpeg-devel,v1,5/6] swscale: Add unscaled XRGB->YUV420P functions

Checks

Commit Message

Patch