Display Layout
When nvsipl_camera
runs with the –d
option, the output
images for one camera group display using the layout below. If an output is not enabled,
the corresponding portion of the display is dark. You can use interactive commands to
switch between the outputs of different camera groups.
If only one output is enabled on a camera group, it displays using the full screen.
Note:
- ISP Formats Usage: when -d option is used, all enabled ISP outputs are passed to
VIC and composition occurs to obtain a RGB888 image that is pushed to display. VIC
does not support composition for floating point or 16-bit images. The default ISP2
output format (RGBA PACKED FP16 PL as documented in the NVIDIA DRIVE OS SDK API
Reference Guide for
INvSIPLCamera::GetImageAttributes
, is overridden to be YUV 420 SEMI-PLANAR UINT8 BL when ISP2 output is requested. - RGB-IR Sensor output: for RGB-IR sensors such as OV2312, ISP1 output cannot be displayed because the LUMA format is 16 bit and not supported by VIC.
- Display Interface: nvsipl_camera app uses OpenWFD display interfaces. This is not compatible with EGLDevice/X11/WinScreen/Wayland.
- VIC Bandwidth: when more than one output displays, while using 8MP camera, to preserve the limited VIC bandwidth, ISP downscales the images from 4k to 108op before passing the images to VIC for format conversion.
RAW output Link 0 |
ISP0 output Link 0 |
ISP1 output Link 0 |
ISP2 output Link 0 |
RAW output Link 1 |
ISP0 output Link 1 |
ISP1 output Link 1 |
ISP2 output Link 1 |
RAW output Link 2 |
ISP0 output Link 2 |
ISP1 output Link 2 |
ISP2 output Link 2 |
RAW output Link 3 |
ISP0 output Link 3 |
ISP1 output Link 3 |
ISP2 output Link 3 |