odas_ros and audio_utils output

[rcodddow]

I have setup odas_ros using <alsa_input.usb-SEEED_ReSpeaker_4_Mic_Array__UAC1.0_-00.multichannel-input> and the sound source localization from the topic /odas/ssl seems to be correct. Is there anyway to configure odas_ros to output [audio_utils/AudioFrame] messages from one or more of the input channels ? I would like to run sound classification alongside the sound source location from odas_ros.

[kanimaru4]

I use "matrix creator" as a microphone array. So I don't know how to use it for "ReSpeaker". If you want to use multiple microphone inputs, you can use PulseAudio to combine multiple microphone inputs and treat them as a single virtual microphone. I used the following link as a reference for my setup. https://wisteriahill.sakura.ne.jp/CMS/WordPress/2022/03/06/raspberry-pi-odas_ros-respeaker-4-mic-array-for-raspberry-pi/

The text is written using a translation application and may not be correct.

[philippewarren]

You can run a audio_utils capture_node to capture the sound input, then send it both to ODAS and your sound classification setup using the capture_node output topic. The odas_ros input topic called raw is where you would send that raw input.

There is a similar setup in the T-Top robot repo:

• capture_node
• odas_ros with the input on the raw topic