Re: [PATCH] V4L: add Y12I, Y8I and Z16 pixel format documentation

[Guennadi Liakhovetski <g.liakhovetski@gmx.de>]

Hi Daniel,

On Mon, 11 Jan 2016, Daniel Johnson wrote:

> On Sat, Jan 9, 2016 at 2:27 AM, Guennadi Liakhovetski
> <g.liakhovetski@gmx.de> wrote:
> > Hi Mauro,
> >
> > Ping - what about this patch? If there are no comments - would you like me
> > to push it via my tree?
> 
> In testing the V4L2_PIX_FMT_Z16 ('Z16 ') format documentation seems to
> be incomplete.
> 
> uvc_xu_control_query unit=2 selector=4 seems to be a z scale factor.
> Changing the value of that control greatly changes the value of
> pixels. Millimeters seems to be correct for the default value of that
> control. This control is on the /dev node for the infrared camera
> rather than the node using the Z16 depth format.

Thanks for your comments. Let me point out though, that the purpose of 
those patches isn't a complete programmers guide of those RealSense 
cameras, rather giving minimum boot-up info. Providing more information 
would require significantly more work.

Thanks
Guennadi

> The one thing that every depth camera I've ever used has in common is
> factory calibration. Translating pixel values to Z is only 1/3 of what
> is needed to translate a depth image into a point cloud. The
> calibration is needed to translate X and Y pixel indexes into
> positions in 3d space. Documentation on fetching, and parsing the
> factory calibration would make the camera much more usable.
> 
> Documentation on the 21 UVC controls would be helpful, but less
> critical than the calibration data.
>