[Bug 755937] v4l2object: probe colorspace supported by device

[GStreamer (GNOME Bugzilla)]

https://bugzilla.gnome.org/show_bug.cgi?id=755937

--- Comment #9 from Aurélien Zanelli <aurelien.zanelli at parrot.com> ---
(In reply to Nicolas Dufresne (stormer) from comment #6)
> Btw, is the opposite properly working, does:
> 
>   v4lsrc ! video/x-raw,colorimetry=bt2020 ! ..
> 
> Will negotiate ?

With Vivid, no. Accordin to the v4l2 spec colorspace is set by the driver for a
capture device and vivid seems to not allow variation. But this definition
could be annoying for colorspace converter.

(In reply to Nicolas Dufresne (stormer) from comment #8)
> gst-launch-1.0 v4l2src device=/dev/video0 num-buffers=1 ! fakesink
> 
> Before:
>   real	0m1.186s
> After
>   real	1m28.960s
> 
> That's more then "slightly" I must say.

Agreed. Well I think ioctl flood on USB interface is not really efficient :-P.
For now I think of two solutions: first we can only probe output devices since,
as I said above, there is no room in the spec for the application to set the
colorspace for capture.
A second one is to just probe the colorspace and not all the
colorspace/matrix/transfer/range variants.

These two changes can also be done together. I will update the patch tomorrow.

-- 
You are receiving this mail because:
You are the QA Contact for the bug.
You are the assignee for the bug.