This is an automatic generated email to let you know that the following patch were queued at the http://git.linuxtv.org/cgit.cgi/v4l-utils.git tree:
Subject: v4l2grab: rework conversion routines to add more YUV formats Author: Mauro Carvalho Chehab <[email protected]> Date: Sun Oct 31 14:24:58 2021 +0000 Rewrite the conversion code in order to simplify the addition of other YUV formats. While here, also use full-range by default, preparing ground to support other colorspace ranges. Signed-off-by: Mauro Carvalho Chehab <[email protected]> contrib/test/v4l2grab.c | 144 +++++++++++++++++++++++++++--------------------- 1 file changed, 81 insertions(+), 63 deletions(-) --- http://git.linuxtv.org/cgit.cgi/v4l-utils.git/commit/?id=9de3fcbd4a365d3a4fb9e14f6d2392c3a1b36734 diff --git a/contrib/test/v4l2grab.c b/contrib/test/v4l2grab.c index 1fa15b75d811..aae9c0a6a461 100644 --- a/contrib/test/v4l2grab.c +++ b/contrib/test/v4l2grab.c @@ -187,83 +187,107 @@ static void querycap(char *fname, int fd, enum io_method method) #define CLAMP(a) ((a) < 0 ? 0 : (a) > 255 ? 255 : (a)) -static void copy_two_pixels(uint32_t fourcc, enum v4l2_ycbcr_encoding enc, - unsigned char *src[2], unsigned char **dst, - int ypos) +static void convert_yuv(enum v4l2_ycbcr_encoding enc, + int32_t y, int32_t u, int32_t v, + unsigned char **dst) { + int full_scale = 1; // FIXME: add support for non-full_scale + + if (full_scale) + y *= 65536; + else + y = (y - 16) * 76284; + + u -= 128; + v -= 128; + + /* + * TODO: add BT2020 and SMPTE240M and better handle + * other differences + */ + switch (enc) { + case V4L2_YCBCR_ENC_601: + case V4L2_YCBCR_ENC_XV601: + case V4L2_YCBCR_ENC_SYCC: + /* + * ITU-R BT.601 matrix: + * R = 1.164 * y + 0.0 * u + 1.596 * v + * G = 1.164 * y + -0.392 * u + -0.813 * v + * B = 1.164 * y + 2.017 * u + 0.0 * v + */ + *(*dst)++ = CLAMP((y + 104595 * v) >> 16); + *(*dst)++ = CLAMP((y - 25690 * u - 53281 * v) >> 16); + *(*dst)++ = CLAMP((y + 132186 * u ) >> 16); + break; + default: + /* + * ITU-R BT.709 matrix: + * R = 1.164 * y + 0.0 * u + 1.793 * v + * G = 1.164 * y + -0.213 * u + -0.533 * v + * B = 1.164 * y + 2.112 * u + 0.0 * v + */ + *(*dst)++ = CLAMP((y + 117506 * v) >> 16); + *(*dst)++ = CLAMP((y - 13959 * u - 34931 * v) >> 16); + *(*dst)++ = CLAMP((y + 138412 * u ) >> 16); + break; + } +} + +static void copy_two_pixels(struct v4l2_format *fmt, + enum v4l2_ycbcr_encoding enc, + unsigned char *src, + unsigned char **dst) +{ + uint32_t fourcc = fmt->fmt.pix.pixelformat; int32_t y_off, u_off, u, v; + uint16_t pix; int i; - switch (fourcc) { + switch (fmt->fmt.pix.pixelformat) { case V4L2_PIX_FMT_RGB565: /* rrrrrggg gggbbbbb */ for (i = 0; i < 2; i++) { - uint16_t pix = (src[i][0] << 8) + src[i][1]; + pix = (src[0] << 8) + src[1]; *(*dst)++ = (unsigned char)(((pix & 0xf800) >> 11) << 3) | 0x07; *(*dst)++ = (unsigned char)((((pix & 0x07e0) >> 5)) << 2) | 0x03; *(*dst)++ = (unsigned char)((pix & 0x1f) << 3) | 0x07; + + src += 2; } break; case V4L2_PIX_FMT_RGB565X: /* gggbbbbb rrrrrggg */ for (i = 0; i < 2; i++) { - uint16_t pix = (src[i][1] << 8) + src[i][0]; + pix = (src[1] << 8) + src[0]; *(*dst)++ = (unsigned char)(((pix & 0xf800) >> 11) << 3) | 0x07; *(*dst)++ = (unsigned char)((((pix & 0x07e0) >> 5)) << 2) | 0x03; *(*dst)++ = (unsigned char)((pix & 0x1f) << 3) | 0x07; + + src += 2; } break; - case V4L2_PIX_FMT_YUYV: - case V4L2_PIX_FMT_UYVY: - case V4L2_PIX_FMT_YVYU: - case V4L2_PIX_FMT_VYUY: + case V4L2_PIX_FMT_YUYV: + case V4L2_PIX_FMT_UYVY: + case V4L2_PIX_FMT_YVYU: + case V4L2_PIX_FMT_VYUY: y_off = (fourcc == V4L2_PIX_FMT_YUYV || fourcc == V4L2_PIX_FMT_YVYU) ? 0 : 1; - u_off = (fourcc == V4L2_PIX_FMT_YUYV || fourcc == V4L2_PIX_FMT_UYVY) ? 0 : 1; + u_off = (fourcc == V4L2_PIX_FMT_YUYV || fourcc == V4L2_PIX_FMT_UYVY) ? 0 : 2; - u = src[1 - y_off][u_off] - 128; - v = src[1 - y_off][1 - u_off] - 128; + u = src[(1 - y_off) + u_off]; + v = src[(1 - y_off) + (2 - u_off)]; + + for (i = 0; i < 2; i++) + convert_yuv(enc, src[y_off + (i << 1)], u, v, dst); - for (i = 0; i < 2; i++) { - int32_t y = src[i][y_off] - 16; - - /* - * TODO: add BT2020 and SMPTE240M and better handle - * other differences - */ - switch (enc) { - case V4L2_YCBCR_ENC_601: - case V4L2_YCBCR_ENC_XV601: - case V4L2_YCBCR_ENC_SYCC: - /* - * ITU-R BT.601 matrix: - * R = 1.164 * y + 0.0 * u + 1.596 * v - * G = 1.164 * y + -0.392 * u + -0.813 * v - * B = 1.164 * y + 2.017 * u + 0.0 * v - */ - *(*dst)++ = CLAMP((76284 * y + 104595 * v + 32768) >> 16); - *(*dst)++ = CLAMP((76284 * y - 25690 * u - 53281 * v + 32768) >> 16); - *(*dst)++ = CLAMP((76284 * y + 132186 * u + 32768) >> 16); - break; - default: - /* - * ITU-R BT.709 matrix: - * R = 1.164 * y + 0.0 * u + 1.793 * v - * G = 1.164 * y + -0.213 * u + -0.533 * v - * B = 1.164 * y + 2.112 * u + 0.0 * v - */ - *(*dst)++ = CLAMP((76284 * y + 117506 * v + 32768) >> 16); - *(*dst)++ = CLAMP((76284 * y - 13959 * u - 34931 * v + 32768) >> 16); - *(*dst)++ = CLAMP((76284 * y + 138412 * u + 32768) >> 16); - break; - } - } break; default: case V4L2_PIX_FMT_BGR24: for (i = 0; i < 2; i++) { - *(*dst)++ = src[i][2]; - *(*dst)++ = src[i][1]; - *(*dst)++ = src[i][0]; + *(*dst)++ = src[2]; + *(*dst)++ = src[1]; + *(*dst)++ = src[0]; + + src += 3; } break; } @@ -272,10 +296,8 @@ static void copy_two_pixels(uint32_t fourcc, enum v4l2_ycbcr_encoding enc, static unsigned int convert_to_rgb24(struct v4l2_format *fmt, unsigned char *p_in, unsigned char *p_out) { - unsigned char *p_start, *p_in_x[2]; - unsigned int bytes_per_pixel; + unsigned char *p_start; unsigned int x, y, depth; - uint32_t fourcc = fmt->fmt.pix.pixelformat; uint32_t width = fmt->fmt.pix.width; uint32_t height = fmt->fmt.pix.height; enum v4l2_ycbcr_encoding enc; @@ -285,7 +307,8 @@ static unsigned int convert_to_rgb24(struct v4l2_format *fmt, unsigned char *p_i else enc = fmt->fmt.pix.ycbcr_enc; - switch (fourcc) { + /* Depth should be multiple of 4 */ + switch (fmt->fmt.pix.pixelformat) { case V4L2_PIX_FMT_BGR24: depth = 24; break; @@ -301,17 +324,12 @@ static unsigned int convert_to_rgb24(struct v4l2_format *fmt, unsigned char *p_i } p_start = p_out; - bytes_per_pixel = depth >> 3; for (y = 0; y < height; y++) { - for (x = 0; x < width; x++) { - p_in_x[0] = p_in; - x++; - p_in += bytes_per_pixel; - p_in_x[1] = p_in; - p_in += bytes_per_pixel; - - copy_two_pixels(fourcc, enc, p_in_x, &p_out, y); + for (x = 0; x < width >> 1; x++) { + copy_two_pixels(fmt, enc, p_in, &p_out); + + p_in += depth >> 2; } } _______________________________________________ linuxtv-commits mailing list [email protected] https://www.linuxtv.org/cgi-bin/mailman/listinfo/linuxtv-commits
