Re: [linux-uvc-devel] Support for Imaging Source cameras

[James F. <ja...@fi...>]

On 11/11/13 19:11, Paulo Assis wrote:

> You are probably right, but if it's there, then someone confirmed that
> a device with that exact VID/PID was working at some point with the
> uvc driver.
> I'm not sure if different revisions of that device/firmware would
> still be labeled with the same PID, but I guess that can happen. That
> at least would explain why a certain device works and another one
> don't.
> In any case if the device is actually UVC compliant, then simply
> mapping it's VID:PID in the driver should be enough to make it work.

This is really why it would be useful to contact someone who knows the
history.  I think there's more going on here than is immediately
obvious.  For example, all of the cameras on this webpage:

   http://www.firstlightoptics.com/imaging-source-cameras.html

are supported according to the ideasonboard.org supported devices list
(they all have VID/PID 199e:8101), but I don't believe any of them will
actually work.

It is possible that there was another firmware release for most of these
cameras to make them fully UVC-compliant, but I haven't got to the
bottom of that yet.  I would be tempted to take the view that "fully"
compliant means no deviation from the standard at all, so claiming to
have a vendor-specific interface type wouldn't count.  On the other
hand, perhaps the alternative firmware release changed the PID so it
would work.  And if different firmware is required to achieve full UVC
compliance, perhaps the default firmware is "close enough" to be able
to get images out of the camera given the right configuration.

It's all very confusing :(

However, I have a written capture application that works for V4L2
devices and some non-V4L2 USB cameras, so I shall push on making it work
where I can and see how far I can get.

Thanks for your help so far.

James