What's the difference of v4l2 and tegradrivers?

@ShaneCCC
What’s the difference of v4l2 and tegradrivers?
I have a camera module(1080P60) and the driver is v4l2. I just do a test like that capture and display on the screen. I found that there is about 60ms delay.
I see the camera driver guide. There are two methods to drive the camera(v4l2 and tegradrivers).
I want to know whether tegradrivers is faster than v4l2?

hello AlfredNG,

please refer to Camera Architecture Stack, for the camera pipeline.

are you talking about capture-to-display latency?
if yes, it should be a reasonable value. may I know what’s the expectation you would like to have?
thanks

Yes! It’s about capture-to-display latency.
Can I reduce the time if I use tegradrivers or any suggestions?

Does it YUV sensor or Bayer sensor?
If it’s YUV sensor that cause current VI driver need to fill 3 ring buffers than return it back to user space.

Try to reduce the QUEUED_BUFFERS in …/drivers/media/platform/tegra/camera/mc_common.h try 3, and 2

hello AlfredNG,

FYI,
here’s evaluation from our side, please expect there’re 2-3 frames for the capture-to-display latency.