Video Image Compositor(VIC) Vs Image Signal Proc(ISP)

Hello,

I am trying to understand functions of VIC and ISP. The CSI-2 camera sensor producing bayer can use ISP block for pixel format conversion, white balancing etc. But if I want to by-pass ISP block, then VIC block can be used for pixel format conversion - do I understand this right?

Where on L4T Camera Software Architecture(nvl4t_docs->Camera Development->Camera Architecture) the VIC is represented?

Is the samples/12_camera_v4l2_cuda example making using of VIC via NvVideoConverter?

Sorry, but I am confused…

Thanks

hello nkumarpsjes,

please refer to Release 28.1 Development Guide, [Camera Development]->[Sensor Driver Programming Guide]
there are two types of camera programming paths.
“Scalable Camera Framework” would coming through ISP block, “Direct V4L2 Interface” would bypass ISP.

Is the samples/12_camera_v4l2_cuda example making using of VIC via NvVideoConverter?
yes, you’re corrected.

Thank you Jerry.
One more question - if I choose CSI camera sensor with built-in ISP, am I forced to use DirectV4L2 interface or I can still use SCF (i.e can I still use Argus APIs)?

Also, is there libargus application example that uses USB camera?

hello nkumarpsjes,

argus application only supporting bayer sensor which coming through the ISP.
you can refer to the samples/12_camera_v4l2_cuda sample code which not go into argus.
thanks