Encode cropped images with RTX 3090 encoder

Accelerated Computing Intelligent Video Analytics DeepStream SDK
1

• Hardware Platform (Jetson / GPU): rtx 3090
• DeepStream Version: 5.1
• TensorRT Version: 7.2.3
• NVIDIA GPU Driver Version (valid for GPU only): 460.73.01
• Issue Type( questions, new requirements, bugs): questions

To crope and encode images from the detection I found nvds_obj_enc_process function (A nice example is in deepstream-image-meta-test example from sample_apps). However, nvds_obj_enc_process function uses JPEG encoding and my RTX 3090 encoder does not support JPEG encoding.

How to encode cropped images with RTX 3090 HW encoder?

2

I mean, JPEG encoding works on RTX 3090 but it consumes a lot of GPU performance, because RTX 3090 HW encoder does not support JPEG encoding.

4

What format do you need? JPEG?

1 Like
5

Any format for which the RTX 3090 will use its HW encoder.
Thank you!

6

nvds_obj_enc_process only support JPEG format encoding. But RTX3090 does not support JPEG HW encoding. Video Encode and Decode GPU Support Matrix [NEW] | NVIDIA Developer

7

@Fiona.Chen yes. Is there any other way how to encode cropped images with RTX 3090 HW encoder? Do you have any hint? Thank you!

8

There is only HW accelerated H264 and H265 video encoding, no HW picture encoding.

1 Like
9

Ok, thank you. Is there any source to find which GPUs support JPEG HW encoding? In the Video Encode and Decode GPU Support Matrix [NEW] | NVIDIA Developer there is no information about JPEG HW encoding.

10

HW JPEG encoder is only supported by embedded platform.

1 Like
11

Thank you. Besides the embedded platform (such as jetson nano?), do A100, V100, T4, RTX 8000 and RTX 4000 also support JPEG HW encoding according this link ?

12

This topic was automatically closed 14 days after the last reply. New replies are no longer allowed.