Hello Forum!
I am using a respeaker core v2 and wanted to extract DoA data from the device. As show in this video (https://www.youtube.com/watch?v=XivM-6PRgX8) which respeaker made. In order to do this I took the example file “pulse_snowboy_1b_test.cc” from the libspeaker website here (https://respeaker.io/librespeaker_doc/index.html). I then disabled the RegisterHotwordDectectionNode() function and added the function SetAutoDOAUpdate() with “true” as the parameter. Next, I called the DoA function in a while loop and averaged the results. However, the output DoA is terribly inconsistent compared to this video.
Does anyone have any ideas as to how to make this data more clean and consistent?
Thank you,
Grant Barthelmes