Unsupported Pixel Format Error When Using Logitech BRIO 4K with MATLAB Simulink on Jetson Orin NX

[kaymazyakuperkan]

Hi,

I’m working with a Logitech BRIO 4K webcam on a Jetson Orin NX (JetPack 6.2.1) and attempting to run the official MathWorks Simulink example for deploying a deep learning model that classifies webcam images directly on the Jetson.

Environment:

• Jetson Orin NX (JetPack 6.2.1, V4L2 v1.22.1, GStreamer 1.20.3)

• MATLAB & Simulink R2025b

• MathWorks NVIDIA Jetson Support Package

• Logitech BRIO 4K (USB 3.0)

I can verify that the camera works with v4l2-ctl and also through tools like cheese and OpenCV. The problem arises only when Simulink attempts to launch the model and access the camera.

During model execution, I consistently get the following runtime error:

MW_camera.c:365 ERROR Unsupported pixel format

I have confirmed from v4l2-ctl --list-formats-ext that the camera supports:

• YUYV (YUV422)

• MJPG

• NV12

All tested at various resolutions including 640x480, 1280x720, 1920x1080.

I suspect this might be due to:

1. Simulink accessing the wrong camera device (IR sensor instead of RGB)

2. Logitech BRIO defaulting to MJPG, which is unsupported

3. GStreamer + V4L2 pipeline limitations in supported FourCC formats

My questions:

• Is there a way to force YUYV or NV12 format at a specific resolution (e.g., 1280x720) on the Jetson Orin using V4L2 or GStreamer configs?

• Have others successfully used Logitech BRIO 4K in uncompressed mode with V4L2/GStreamer on Jetson?

• Is this a known issue when cameras expose multiple video devices (e.g., IR and RGB sensors)?

Any guidance on how to configure the Logitech BRIO or Jetson to ensure compatible pixel formats for Simulink/V4L2 access would be greatly appreciated.

Thanks in advance.

[JerryChang]

hello kaymazyakuperkan,

you may use video converter, nvvidconv to convert the formats.
for instance,
here’s capture pipeline captures video from the /dev/video0 device at a resolution of 1280x720 and converts it to NV12 format.
$ gst-launch-1.0 nvv4l2camerasrc device=/dev/video0 ! 'video/x-raw(memory:NVMM), width=(int)1280, height=(int)720' ! nvvidconv ! 'video/x-raw(memory:NVMM), format=(string)NV12' ! nv3dsink -e

@kaymazyakuperkan

The problem arises only when Simulink attempts to launch the model and access the camera.

the root cause should be on the application side, you may dig into this for confirmation.

[kaymazyakuperkan]

@JerryChang

gst-launch-1.0 nvv4l2camerasrc device=/dev/video0 ! ‘video/x-raw(memory:NVMM), width=(int)1280, height=(int)720’ ! nvvidconv ! ‘video/x-raw(memory:NVMM), format=(string)NV12’ ! nv3dsink -e

Hello,

I am trying to deploy my Simulink model on the Jetson Orin NX using the Logitech BRIO 4K camera. However, I am still encountering the following error:

MW_camera.c:365 ERROR Unsupported pixel format

Current Situation:

• The Simulink model deploys successfully.

• When SDL video display opens, I can see colored blocks, but the image is not clear. The FPS value shows 6.01, indicating that data is being received.

Log Output & Issues:

1. Image is displayed like this (as shown in the screenshot above).

2. Log Output:

   • Video stream starts successfully, but I get the “Unsupported pixel format” error.

   • Camera format: I checked with v4l2-ctl --list-formats-ext and confirmed that YUYV, MJPG, and NV12 formats are supported.

Steps Taken:

1. GStreamer pipeline for format conversion:

   gst-launch-1.0 v4l2src device=/dev/video0 ! 'video/x-raw,width=1280,height=720' ! nvvidconv ! 'video/x-raw(memory:NVMM),format=NV12' ! nv3dsink
   
   

   However, this did not solve the issue.

2. For the Simulink Camera Block:

   • I ensured the correct device (/dev/video0) was selected in the Camera block.

   • I added a Color Space Conversion block in Simulink for YCbCr → RGB conversion.

My Expectation:

• To get RGB format video from the Logitech BRIO 4K camera and display it correctly in Simulink for image classification and real-time streaming.

The Error Still Occurs:

• The “Unsupported pixel format” error persists.

How can I solve this issue? Do I need to make additional settings or try another approach?

Thank you!

In additon,example link is here : https://www.mathworks.com/help/coder/nvidia/ug/deploy-classify-webcam-images-jetsonTX2-from-Simulink.html

[JerryChang]

hello kaymazyakuperkan,

@kaymazyakuperkan

1. GStreamer pipeline for format conversion:

this is the pipeline for rendering camera frame to display. it’s demonstration for format conversation.