Our design (Dm368 hardware based on leopardboard, Appro IPNC MT5 SDK 4.x) captures BT.1120 input from an LVDS converter. The input will be one of a small number of video modes.
As the video capture task / avserver will fail to start if it doesn't see the correct signals on its interface, it would be convenient if we could auto-detect the video mode and set the capture front end accordingly.
For example, we may only want to use four or five modes:
1080p30, 1080p25
720p60, 720p30, 720p25
Which means (I think) that from the HD/VD/PCLK timings it would be possible to decide what's there. In a bare-metal embedded system I can see this would be quite trivial, but I'm not sure how to proceed when Linux is between me & the hardware?
