Black isn't always just memset(ptr, 0, size). Limited YUV in particular requires relatively non-obvious values, and filling a frame with repeating 0 bytes is disallowed in some contexts. With component sizes larger than 8 or packed YUV, this can become relatively complicated. So having a generic function for this seems helpful.
In order to handle the complex cases in a generic way without destroying performance, this code attempts to compute a black pixel, and then uses that value to clear the image data quickly by using a function like memset. Common cases like yuv410p10 or rgba can't be handled with a simple memset, so there is some code to fill memory with 2 and 4 byte patterns. For the remaining cases, a generic slow fallback is used. This code should probably have a 8 byte case too, to deal with rgba64. --- Fixed: - simply reduction to memset code - fix monow/monob handling - change alpha range - add 8 bytes path (untested) - allow dst_data==NULL with special semantics TODO: APIchanges, version bumps, FATE test --- libavutil/imgutils.c | 167 +++++++++++++++++++++++++++++++++++++++++++++++++++ libavutil/imgutils.h | 27 +++++++++ 2 files changed, 194 insertions(+) diff --git a/libavutil/imgutils.c b/libavutil/imgutils.c index 84abb11656..c43185ccff 100644 --- a/libavutil/imgutils.c +++ b/libavutil/imgutils.c @@ -435,3 +435,170 @@ int av_image_copy_to_buffer(uint8_t *dst, int dst_size, return size; } + +// Fill dst[0..dst_size] with the bytes in clear[0..clear_size]. The clear +// bytes are repeated until dst_size is reached. If dst_size is unaligned (i.e. +// dst_size%clear_size!=0), the remaining data will be filled with the beginning +// of the clear data only. +static void memset_bytes(uint8_t *dst, size_t dst_size, uint8_t *clear, + size_t clear_size) +{ + size_t pos = 0; + int same = 1; + int i; + + if (!clear_size) + return; + + // Reduce to memset() if possible. + for (i = 0; i < clear_size; i++) { + if (clear[i] != clear[0]) { + same = 0; + break; + } + } + if (same) + clear_size = 1; + + if (clear_size == 1) { + memset(dst, clear[0], dst_size); + dst_size = 0; + } else if (clear_size == 2) { + uint16_t val = AV_RN16(clear); + for (; dst_size >= 2; dst_size -= 2) { + AV_WN16(dst, val); + dst += 2; + } + } else if (clear_size == 4) { + uint32_t val = AV_RN32(clear); + for (; dst_size >= 4; dst_size -= 4) { + AV_WN32(dst, val); + dst += 4; + } + } else if (clear_size == 8) { + uint32_t val = AV_RN64(clear); + for (; dst_size >= 8; dst_size -= 8) { + AV_WN64(dst, val); + dst += 8; + } + } + + for (; dst_size; dst_size--) + *dst++ = clear[pos++ % clear_size]; +} + +// Maximum size in bytes of a plane element (usually a pixel, or multiple pixels +// if it's a subsampled packed format). +#define MAX_PIXEL_SIZE 32 + +int av_image_fill_black(uint8_t *dst_data[4], const ptrdiff_t dst_linesize[4], + enum AVPixelFormat pix_fmt, enum AVColorRange range, + int width, int height) +{ + const AVPixFmtDescriptor *desc = av_pix_fmt_desc_get(pix_fmt); + int nb_planes = av_pix_fmt_count_planes(pix_fmt); + // A pixel on each plane, with a value that represents black. + // Consider e.g. AV_PIX_FMT_UYVY422 for non-trivial cases. + uint8_t pixel[4][MAX_PIXEL_SIZE] = {0}; // clear padding with 0 + int pixel_size[4] = {0}; + ptrdiff_t plane_line_bytes[4] = {0}; + int rgb, limited; + int plane, c; + + if (!desc || nb_planes < 1 || nb_planes > 4 || desc->flags & AV_PIX_FMT_FLAG_HWACCEL) + return AVERROR(EINVAL); + + rgb = !!(desc->flags & AV_PIX_FMT_FLAG_RGB); + limited = !rgb && !range; + + if (desc->flags & AV_PIX_FMT_FLAG_BITSTREAM) { + ptrdiff_t bytewidth = av_image_get_linesize(pix_fmt, width, 0); + uint8_t *data; + int mono = pix_fmt == AV_PIX_FMT_MONOWHITE || pix_fmt == AV_PIX_FMT_MONOBLACK; + int fill = pix_fmt == AV_PIX_FMT_MONOWHITE ? 0xFF : 0; + if (nb_planes != 1 || !(rgb || mono) || bytewidth < 1) + return AVERROR(EINVAL); + + if (!dst_data) + return 0; + + data = dst_data[0]; + + // (Bitstream + alpha will be handled incorrectly - it'll remain transparent.) + for (;height > 0; height--) { + memset(data, fill, bytewidth); + data += dst_linesize[0]; + } + return 0; + } + + for (c = 0; c < desc->nb_components; c++) { + const AVComponentDescriptor comp = desc->comp[c]; + + // We try to operate on entire non-subsampled pixel groups (for + // AV_PIX_FMT_UYVY422 this would mean two consecutive pixels). + pixel_size[comp.plane] = FFMAX(pixel_size[comp.plane], comp.step); + + if (pixel_size[comp.plane] > MAX_PIXEL_SIZE) + return AVERROR(EINVAL); + } + + // Create a byte array for clearing 1 pixel (sometimes several pixels). + for (c = 0; c < desc->nb_components; c++) { + const AVComponentDescriptor comp = desc->comp[c]; + // (Multiple pixels happen e.g. with AV_PIX_FMT_UYVY422.) + int w = pixel_size[comp.plane] / comp.step; + uint8_t *c_data[4]; + const int c_linesize[4] = {0}; + uint16_t src_array[MAX_PIXEL_SIZE]; + uint16_t src = 0; + int x; + + if (comp.depth > 16) + return AVERROR(EINVAL); + if (!rgb && comp.depth < 8) + return AVERROR(EINVAL); + if (w < 1) + return AVERROR(EINVAL); + + if (c == 0 && limited) { + src = 16 << (comp.depth - 8); + } else if ((c == 1 || c == 2) && !rgb) { + src = 128 << (comp.depth - 8); + } else if (c == 3) { + // (Assume even limited YUV uses full range alpha.) + src = (1 << comp.depth) - 1; + } + + for (x = 0; x < w; x++) + src_array[x] = src; + + for (x = 0; x < 4; x++) + c_data[x] = &pixel[x][0]; + + av_write_image_line(src_array, c_data, c_linesize, desc, 0, 0, c, w); + } + + for (plane = 0; plane < nb_planes; plane++) { + plane_line_bytes[plane] = av_image_get_linesize(pix_fmt, width, plane); + if (plane_line_bytes[plane] < 0) + return AVERROR(EINVAL); + } + + if (!dst_data) + return 0; + + for (plane = 0; plane < nb_planes; plane++) { + size_t bytewidth = plane_line_bytes[plane]; + uint8_t *data = dst_data[plane]; + int chroma_div = plane == 1 || plane == 2 ? desc->log2_chroma_h : 0; + int plane_h = ((height + ( 1 << chroma_div) - 1)) >> chroma_div; + + for (; plane_h > 0; plane_h--) { + memset_bytes(data, bytewidth, &pixel[plane][0], pixel_size[plane]); + data += dst_linesize[plane]; + } + } + + return 0; +} diff --git a/libavutil/imgutils.h b/libavutil/imgutils.h index fafabc48ca..ae2e6287eb 100644 --- a/libavutil/imgutils.h +++ b/libavutil/imgutils.h @@ -225,6 +225,33 @@ int av_image_check_size(unsigned int w, unsigned int h, int log_offset, void *lo int av_image_check_sar(unsigned int w, unsigned int h, AVRational sar); /** + * Overwrite the image data with black. This is suitable for filling a + * sub-rectangle of an image, meaning the padding between the right most pixel + * and the left most pixel on the next line will not be overwritten. For some + * formats, the image size might be rounded up due to inherent alignment. + * + * If the pixel format has alpha, the alpha is cleared to opaque. + * + * This can return an error if the pixel format is not supported. Normally, all + * non-hwaccel pixel formats should be supported. + * + * Passing NULL for dst_data is allowed. Then the function returns whether the + * operation would have succeeded. (It can return an error if the pix_fmt is + * not supported.) + * + * @param dst_data data pointers to destination image + * @param dst_linesize linesizes for the destination image + * @param pix_fmt the pixel format of the image + * @param range the color range of the image (important for colorspaces such as YUV) + * @param width the width of the image in pixels + * @param height the height of the image in pixels + * @return 0 if the image data was cleared, a negative AVERROR code otherwise + */ +int av_image_fill_black(uint8_t *dst_data[4], const ptrdiff_t dst_linesize[4], + enum AVPixelFormat pix_fmt, enum AVColorRange range, + int width, int height); + +/** * @} */ -- 2.11.0 _______________________________________________ libav-devel mailing list libav-devel@libav.org https://lists.libav.org/mailman/listinfo/libav-devel