Dubravko Moravski
2014-06-18 18:03:18 UTC
Hello,
I'm assigned to develop software for a new camera for Linux. Since my
experience with Linux is limited, I would appreciate any advice and
comments.
Currently on Windows, we use UVC to send video and HID to adjust camera
settings. This works well for usage scenarios we encountered so for. Is this
architecture good for Linux too? Regarding UVC, do I really only need to add
VID & PID to the uvc_ids[] array in uvc_driver.c? (For adjustments over HID,
we know we need to port our settings-adjusting application.)
Would there be any benefits to alternative approach where we don't use UVC
and HID, but develop a driver for V4L2? We also write firmware for the
camera, so we can "tell" it to provide any interface we need.
What is the current status of still image capture feature, that our clients
wish to have? http://www.ideasonboard.org/uvc/ lists it as not supported -
is this correct, or is there at least any experimental support?
Thanks and best regards,
Dubravski
I'm assigned to develop software for a new camera for Linux. Since my
experience with Linux is limited, I would appreciate any advice and
comments.
Currently on Windows, we use UVC to send video and HID to adjust camera
settings. This works well for usage scenarios we encountered so for. Is this
architecture good for Linux too? Regarding UVC, do I really only need to add
VID & PID to the uvc_ids[] array in uvc_driver.c? (For adjustments over HID,
we know we need to port our settings-adjusting application.)
Would there be any benefits to alternative approach where we don't use UVC
and HID, but develop a driver for V4L2? We also write firmware for the
camera, so we can "tell" it to provide any interface we need.
What is the current status of still image capture feature, that our clients
wish to have? http://www.ideasonboard.org/uvc/ lists it as not supported -
is this correct, or is there at least any experimental support?
Thanks and best regards,
Dubravski