Jetson 硬件解码 .H264 为 YUV 文件出现丢失

Robotics & Edge Computing Jetson & Embedded Systems Jetson Orin NX
1

a587cc19c75b2a1d8a404fa22858291a
a587cc19c75b2a1d8a404fa22858291a1045×629 32.7 KB

用 Jetson 硬件解码 .H264 文件解码出来的 YUV 文件很小只有前面几帧，但是用软件解码同一个 .H264 文件出来的 YUV 是正常的。这个问题要怎么排查呢？

Module: NVIDIA Jetson Orin NX (16GB ram)
Jetpack：6.2.1

硬解命令：
ffmpeg -i swim_720p.h264 -s 1280x720 -c:v h264_nvmpi -f rawvideo -pix_fmt yuv420p swim_720p.yuv

3

*** Please note that this reply is generated by LLM automatically ***

Based on the provided context, it appears that you are using the NVIDIA Jetson Orin NX 16GB module with Jetpack 6.2.1. You are attempting to use FFmpeg to transcode a video file named “swim720p.h264” to a raw video file named “swim720p.yuv” with a resolution of 1280x720 and a pixel format of yuv420p.

Here is a breakdown of the command you provided:

  • ffmpeg: The command-line tool for FFmpeg.
  • -i swim720p.h264: Specifies the input file name.
  • -s 1280x720: Sets the resolution of the output video to 1280x720.
  • -c:v h264_nvmpi: Specifies the video codec to use for the output video. In this case, it is the NVIDIA H.264 encoder.
  • -f rawvideo: Specifies the format of the output video as raw video.
  • -pix_fmt yuv420p: Sets the pixel format of the output video to yuv420p.
  • swim720p.yuv: Specifies the output file name.

The NVIDIA Jetson Orin NX 16GB module is a powerful AI computing platform that is capable of handling complex video processing tasks. The module features a quad-core ARM Cortex-A78 CPU, 16GB of LPDDR5 RAM, and a 256-core NVIDIA Ampere GPU.

In comparison to other NVIDIA Jetson modules, the Orin NX 16GB module has a similar architecture to the Jetson Orin NX 8GB module, but with more RAM and a slightly higher maximum operating frequency (1229 MHz vs 610 MHz). The Orin NX 16GB module also has a higher maximum sparse INT8 operations per second (80 TOPS vs 20 TOPS) compared to the Orin NX 8GB module.

In terms of video processing performance, the NVIDIA Jetson Orin NX 16GB module is capable of handling 4K video encoding and decoding at 60 FPS, as well as 8K video encoding and decoding at 30 FPS. The module also supports hardware-accelerated video codecs such as H.264, H.265, and VP9.

Overall, the NVIDIA Jetson Orin NX 16GB module is a powerful platform for video processing and AI applications, and is well-suited for tasks such as video transcoding, object detection, and image classification.

*** Please give the thumbs up if you get answers you like. Or provide feedback to help us improve the answer. ***

5

Hi,
Please refer to FAQ to install ffmpeg with hardware decoding:
Jetson AGX Orin FAQ

Or you can try jetson_multimedia_api sample:

/usr/src/jetson_multimedia_api/samples/00_video_decode/

7

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