You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Hello, I am using the mic for smart-speaker application.
Microphone is connected to a single board computer via USB. Sound system is connected to 3.5mm audio jack on the microphone for AEC to work. Below is the demonstration of setup:
I have watched your video showing the AEC performance, and it seems very good. I hoped that I will have similar results, but unfortunately microphone recording gets distorted by the signal emitted through speaker more than on your video. When I compared recording of single-channel firmware vs the first-channel of 6-channel firmware(processed for AEC), the quality of the latter is much better. But still it is not sufficient. I need to understand what is wrong with my setup, and why it doesn't work as in your video.
For my application I am using Linux, and this is what I do for audio settings:
Through this, I make sure that only the first channel is used for my voice detection application(clear of noise), and that audio output is streamed into the microphone for the filtering purpose.
Below are various logs from pulse-audio regarding sampling rate and etc. Could you please check if it they are ok for good AEC performance?
What can I do to improve the noise cancellation?
May it be that amplifier induces a delay which worsens the filtering? If yes, what would you suggest?
Hello, I am using the mic for smart-speaker application.
Microphone is connected to a single board computer via USB. Sound system is connected to 3.5mm audio jack on the microphone for AEC to work. Below is the demonstration of setup:
I have watched your video showing the AEC performance, and it seems very good. I hoped that I will have similar results, but unfortunately microphone recording gets distorted by the signal emitted through speaker more than on your video. When I compared recording of single-channel firmware vs the first-channel of 6-channel firmware(processed for AEC), the quality of the latter is much better. But still it is not sufficient. I need to understand what is wrong with my setup, and why it doesn't work as in your video.
For my application I am using Linux, and this is what I do for audio settings:
Through this, I make sure that only the first channel is used for my voice detection application(clear of noise), and that audio output is streamed into the microphone for the filtering purpose.
from pacmd stat:
from pacmd list-sources:
from pacmd list-sinks:
P.S. If you want I can make a video which shows performance on audacity.
The text was updated successfully, but these errors were encountered: