I’ve read reports that say that capturing video from a CSI-connected camera will only provide bayer data, and that only using libargus enables the ISP.
I’d prefer to use v4l2, mostly because it has a C (vs C++) interface, which is easier to access via FFI. And, it does seem that when I use libv4l2, I get lots of libargus output.
Before I go all the way down the road of wiring up the video encoder, etc, I’d like to know whether this is the case – can I get YUV420M or NV12M or whatever out of the camera via v4l2, or do I need to use libargus?
(In case it matters, I’m using a partner-supplied camera with driver support already.)