Microphone and speaker feedback - can I get rid of it?

I

iago2012-05-29 09:22:30

Android

iago, 2012-05-29 09:22:30

Good morning community!
Feedback problem. Do not take it for advertising, but you have to say 2 words about the application to explain the essence of the issue - this is a 3D football funnel, i.e. the user blows into the microphone of the device, and the speaker plays the sound of a football turbo. The sound of a noisy stadium is also played in the background. So far it 's ready for iPhone , but the same problem has been noticed in the android version, which will be uploaded to Google Play very soon, so I really, really ask for help!
If you put the phone on your hand in such a way that the palm “covers” both the microphone and the speaker, positive feedback appears, and although I am an automation engineer, I cannot figure out how to get rid of it without physical negative feedback. The processor cannot be heavily overloaded, because. it should simultaneously show the camera, draw the 3D model, read the microphone signal 100 times per second, and play the sound.
Does anyone know of any C/C++ libraries that can be used to solve this problem? There is no time to do research in this area from scratch. Maybe someone had a similar problem?

Reply

Answer the question

In order to leave comments, you need to log in

4 answer(s)

R

rPman, 2012-05-29
@rPman

If I understand noise reduction technologies correctly, then with the equipment that is provided in any cell phone, this is impossible (in the general case), a series of microphones (matrix) is needed there.
Filter by frequency, but it's even better not to output the sound from the microphone directly, but to detect by 'patterns' and produce a clean recording, etc.

K

kamlay, 2012-05-29
@kamlay

Shift the output signal relative to the input by a few tens of hertz. I don’t know how to do it mathematically, how difficult it is. But devices on this principle work in concert practice.

B

bagyr, 2012-05-29
@bagyr

In practice, this is done by adaptive filters. LMS is simple and works well on voice.
With digital filters, also darken something, google any implementation of the Chebyshev filter and play around, there are no such delays there.

K

kostyl, 2012-08-29
@kostyl

Invert what you reproduce and add to what you write, the excess will be subtracted ...