[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <94f0b9bc-542b-6e19-3ca1-332632c135f7@xs4all.nl>
Date: Sat, 16 Jul 2016 17:28:08 +0200
From: Hans Verkuil <hverkuil@...all.nl>
To: Ricardo Ribalda Delgado <ricardo.ribalda@...il.com>,
Laurent Pinchart <laurent.pinchart@...asonboard.com>
Cc: Mauro Carvalho Chehab <mchehab@....samsung.com>,
Laurent Pinchart <laurent.pinchart+renesas@...asonboard.com>,
Sakari Ailus <sakari.ailus@...ux.intel.com>,
Antti Palosaari <crope@....fi>,
Guennadi Liakhovetski <guennadi.liakhovetski@...el.com>,
Helen Mae Koike Fornazier <helen.koike@...labora.co.uk>,
Philipp Zabel <p.zabel@...gutronix.de>,
Shuah Khan <shuahkh@....samsung.com>,
linux-media <linux-media@...r.kernel.org>,
LKML <linux-kernel@...r.kernel.org>
Subject: Re: [PATCH v2 2/6] [media] Documentation: Add HSV format
On 07/16/2016 04:32 PM, Ricardo Ribalda Delgado wrote:
> Hi
>
> On Sat, Jul 16, 2016 at 4:12 PM, Laurent Pinchart
> <laurent.pinchart@...asonboard.com> wrote:
>
>> I'd still like to know about it for my personal information :-)
>
> Maybe it is just a very cheap gamma.
>
>>
>>> Anyway, I am inclined to use ycbcr_enc as well.
>>
>> I'm glad we agree.
>>
>
> Are you thinking about something like this:
>
>
> diff --git a/include/uapi/linux/videodev2.h b/include/uapi/linux/videodev2.h
> index c7fb760386cf..3e613fba1b20 100644
> --- a/include/uapi/linux/videodev2.h
> +++ b/include/uapi/linux/videodev2.h
> @@ -330,6 +330,16 @@ enum v4l2_ycbcr_encoding {
> V4L2_YCBCR_ENC_SMPTE240M = 8,
> };
>
> +enum v4l2_hsv_encoding {
> + V4L2_HSV_ENC_180 = 16,
> + V4L2_HSV_ENC_256 = 17,
> +};
Yes.
> +
> +enum v4l2_rgb_encoding {
> + V4L2_RGB_ENC_FULL = 32,
> + V4L2_HSV_ENC_16_235 = 33,
> +};
No.
The starting point is RGB, so there is nothing to encode. YCbCr and HSV
are transformations from RGB (or R'G'B' to be precise).
The basic chain is:
XYZ -> colorspace conversion -> RGB -> transfer function -> R'G'B'
and after this there is an optional step that converts R'G'B' to
Y'CbCr or HSV (ycbcr_enc or hsv_enc). The final step is the quantization
step where you end up with the actual color component values.
There doesn't seem to be something like limited range HSV, so I'd say it's
full range only.
Regards,
Hans
> +
> /*
> * Determine how YCBCR_ENC_DEFAULT should map to a proper Y'CbCr encoding.
> * This depends on the colorspace.
> @@ -455,7 +465,11 @@ struct v4l2_pix_format {
> __u32 colorspace; /* enum v4l2_colorspace */
> __u32 priv; /* private data,
> depends on pixelformat */
> __u32 flags; /* format flags
> (V4L2_PIX_FMT_FLAG_*) */
> - __u32 ycbcr_enc; /* enum v4l2_ycbcr_encoding */
> + union {
> + __u32 ycbcr_enc; /* enum
> v4l2_ycbcr_encoding */
> + __u32 hsv_enc; /* enum
> v4l2_hsv_encoding */
> + __u32 rgb_enc; /* enum
> v4l2_rgb_encoding */
> + };
> __u32 quantization; /* enum v4l2_quantization */
> __u32 xfer_func; /* enum v4l2_xfer_func */
> };
> @@ -1988,7 +2002,11 @@ struct v4l2_pix_format_mplane {
> struct v4l2_plane_pix_format plane_fmt[VIDEO_MAX_PLANES];
> __u8 num_planes;
> __u8 flags;
> - __u8 ycbcr_enc;
> + union {
> + __u8 ycbcr_enc; /* enum
> v4l2_ycbcr_encoding */
> + __u8 hsv_enc; /* enum
> v4l2_hsv_encoding */
> + __u8 rgb_enc; /* enum
> v4l2_rgb_encoding */
> + };
> __u8 quantization;
> __u8 xfer_func;
> __u8 reserved[7];
>
>> --
>
>
> Best regards!
>
Powered by blists - more mailing lists