I download the handbrake installer with this URL - “http://ports.ubuntu.com/pool/universe/h/handbrake/handbrake_1.3.1+ds1-1build1_arm64.deb”, but there is error during install. Can anyone face the issues and how to resolve? please help and thank you.
Not sure but handbrake looks to be video converter. If you are looking for video transcoding functions on Jetson platforms, we would suggest use gstreamer for using hardware encoding/decoding engines. Please refer to gstreamer user guide
We don’t see much information about handbrake installer on Jetson platforms. This would need other users to share experience.
Follow Handbrake guideline and I able to resolve my issues now, thank you.
By the way, do know have experience to use Gstreamer to collect HW performance? Which to measure the hardware performance on (CPU - Latency, GPU/DLA - Throughput) during encode/decode the video clip using Gstreamer pipeline? Else, is there others source I can research for? As we know we can use Handbrake, but version supported to Jetson unit only is handbrake v1.1.2 which that unable to help to turn “NVENC” or “NVDEC” when running handbrake benchmark. It’s higher v1.3.3 supported when I run on Ubuntu 20.04 with Nvidia GPU card (T4). So, current looking the way how to do media benchmark on Jetson platform.
Below is my Jetson NX hardware info.
- NVIDIA Jetson Xavier NX (Developer Kit Version)
- Jetpack 4.6 [L4T 32.6.1]
- NV Power Mode: MODE_20W_6CORE - Type: 8
- jetson_stats.service: active
- Board info:
- Type: Xavier NX (Developer Kit Version)
- SOC Family: tegra194 - ID:25
- Module: P3668 - Board: P3509-000
- Code Name: jakku
- CUDA GPU architecture (ARCH_BIN): 7.2
- Serial Number: 1424620042014
- CUDA: 10.2.300
- cuDNN: 126.96.36.199
- TensorRT: 188.8.131.52
- Visionworks: 184.108.40.2061
- OpenCV: 4.1.1 compiled CUDA: NO
- VPI: ii libnvvpi1 1.1.11 arm64 NVIDIA Vision Programming Interface library
- Vulkan: 1.2.70
- Version 3.1.1
- Works on Python 3.6.9
On Jetson platforms, we use tegrastats to get system status. Please take a look at
And the predefined power modes:
I’m sorry I’m beginner of AI Benchmark member here, If I understand correctly, are you means I need to have 2 session execute as same time. one is running Gstreamer to encode/decode the video and another session to have tergastats to read and calculate the output of performance? correct me if I’m wrong. or do you have sample or example how to do that?