NVIDIA Jetson AGX Orin 64GB Setup

I am following the howto: https://developer.nvidia.com/embedded/learn/tutorials/vnc-setup

Flashing new firmware according to https://wiki.seeedstudio.com/ReSpeaker-USB-Mic-Array/

I am flashing the “1 channel firmware”:

sudo apt-get update
sudo pip install pyusb click
 
git clone https://github.com/respeaker/usb_4_mic_array.git
cd usb_4_mic_array
 
# if you want to use 1 channel,then the command should be like:
sudo python dfu.py --download 1_channel_firmware.bin
 
# Alternative, did not work for me (mic not showing up in GNOME):
# sudo python dfu.py --download 6_channels_firmware.bin  # The 6 channels version

Then disconnect ReSpeaker from USB and reconnect. The mircophone occurs in the GNOME audio settings menu.

• https://wiki.seeedstudio.com/ReSpeaker_Product_Guide/
• https://wiki.seeedstudio.com/ReSpeaker_Mic_Array_v2.0/

To test the microphone and to record audios for training the LLM and ASR.

Buggy or weird behavior of Audacity on Linux.

Don't do that: “sudo snap connect audacity:alsa”. In case you did you can “sudo snap disconnect audacity:alsa”.

Don't use the button “Access ot ALSA has been granted.”

Rather click on button “Start and do not ask again” when Audacity comes up.

I am following https://www.jetson-ai-lab.com/tutorial_whisper.html

git clone https://github.com/dusty-nv/jetson-containers
bash jetson-containers/install.sh

jetson-containers run $(autotag whisper)

Open …

git clone https://github.com/LJ-Hao/Deploy-Whisper-on-NVIDIA-Jetson-Orin-for-Real-time-Speech-to-Text.git
cd Deploy-Whisper-on-NVIDIA-Jetson-Orin-for-Real-time-Speech-to-Text
pip install -r requirements.txt
sudo apt update && sudo apt install ffmpeg
arecord -D hw:2,0 --dump-hw-params #set microphone rate to 16000

PyAudio missing. Requires trouble shooting … On hold …

I am following https://www.jetson-ai-lab.com/tutorial_llamaspeak.html