Audio signal beamformer algo

Hi

I am looking for 3D audio signal information.

For example

[Signal,fs] = audioread('*.wav')
x_dim = Signal.x
y_dim = Signal.y
z_dim = Signal.z

Is there a way to record/capture audio signal from microphone in x,y,z axis/dimension ?

plot(x_dim,y_dim,z_dim) if so , how to plot x,y,z coordinates signal.

Derive Azimuth,elevation and desired signal for x,y,z axis co ordinates

Thank you!

[Adam Danz]

Adam Danz's answer moved here as a comment

==========================================

> Is there a way to record/capture audio signal from microphone...

Yes: https://www.mathworks.com/help/matlab/import_export/record-and-play-audio.html

> ...in x,y,z axis/dimension?

Is z the depth of the sound source? audiorecorder only supports 1 or 2 channels (mono|stereo).

> Derive Azimuth,elevation and desired signal for x,y,z axis co ordinates

That's a hard problem if you're expecting to derive the depth of the sound source. Maybe the Audio Toolbox has something to offce.

[Adam Danz]

Jogger's comment to Adam's Answer

===============================

Thanks Adam,

Probably I did not explain problem well enogh , sorry about that.

I am looking for audio beamformer for a microphone

In coming data to a microphone in x,y,z coordinate is in matrix form and microphone dsp to use smart beamforming algo to detect SNR for each coordinate . Estimate the Azimuth,elevation and propose desired location.

Can you please help me here with a sample code

Thank you!

[Adam Danz]

@Jogger, thanks for the clarification. I moved my answer and your response to the comment section so that your question does not appear as answered.

I don't have a background in audio beamformer signal processing so if I were you I'd start by googleing "matlab audio beamformer for a microphone" which lists a lot of potentially relevant material such as this link.

[Life is Wonderful]

Thank you!

I will follow shared link and learn .

Answer 1

[Gabriele Bunkheila]

audioread returns an N-by-M numeric matrix, where N is the number of samples and M is the number of recorded channels:

[signal,fs] = audioread("recording.wav");
t = (0:size(signal,1)-1)/fs;
plot(t,signal)
xlabel("Time (s)")
ylabel("Amplitude")

A conventional microphone measures scalar sound pressure versus time, not x, y, and z coordinates. Therefore, fields such as signal.x do not exist. Recorded audio in .wav files and functions reading them like audioread can't guess the topological placement of microphones used to record the audio samples stored in the file.

The same holds true for synchronized multichannel audio acquired from hardware (single-shot, but this could also be used for continuous live acquisition; this code requires Audio Toolbox):

as = audiostreamer("recorder");
fs = audiostreamer.SampleRate;
signal = record(as, 3*fs); % record three seconds of audio

in this case you need to ensure that the recording hardware exposes synchronized microphone channels.

On estimating the position of a sound source in open space - first of all a recording from single microphone doesn't allow to derive any spatial information about the recorded sound source.

Direction-of-arrival estimation algorithms require several synchronized microphones with known 3-D positions. The time or phase differences between their recordings can then be processed using methods such as GCC-PHAT, MUSIC, or delay-and-sum beamforming. The microphone positions are the spatial coordinates; the recorded channels remain pressure waveforms.

For azimuth and elevation estimation, the microphone geometry, spacing, number of microphones, and expected signal bandwidth must also be specified. With than information in hand, go through the following examples to learn more about available direction-of-arrival and beamforming techniques:

• Source Localization Using Generalized Cross Correlation
• Acoustic Beamforming Using a Microphone Array