I was wondering if you guys have the 6ch firmware working properly? My board is not doing the voice extraction, it is only summing the raw data onto channel0, and ch 5 has nothing.
Please let me know how to fix this. And if you have any information about getting access to the source code of the built-in algorithms, would be great!
Thanks,
Do you feed the output sound through the device?
I always thought that the board would output the extracted voice automatically. Do you happen to have the code that feeds this output? And which channels exactly should be fed?
Thanks for your reply!
Well, as far as I know, the board performs echo cancelation and voice extraction by subtracting the sound it is playing from the sound sensed by mics. (I.e. subtract output from input).
I use a different sound card for sound output, but I also copy it to the board with pulseaudio module-combine-think so it could have something to subtract
Looks like acoustic echo cancelation is far from perfect in my case as well. It would be nice to have a guide on how to achieve best results, detailed parameter descriptions, and which parameters were used to produce the demo on the device promo page.
I will also try echo cancelation with pulseaudio and post the results
Hello Shulyaka!
I’ve been working with the raw data from the microphones only because my system does not output sound. Therefore, the AEC processing does not help me too much. But I was hoping to get better voice amplification / noise suppression from the processed channel.
I am using the chip recently under windows system. I just cant figure it out how to get the chip output in 6 channels,. Whether you have a method or any way to get the 6 channel do its job , please inform me.
@Tragic All you have to do is set the number of recording channels to 6. There should be a dropdown menu on the device toolbar for this. It’s next to the dropdown used for selecting the recording device (which should be set to “ReSpeaker 4 Mic Array (UAC 1.0)”).