What should be done for the human voice?

Hi, I'm a bit of a beginner. My goal is to filter the human voice. I want the human voice to be taken, that is, to be shot, among the constantly changing noises. So far, I have played in the fdatool window. I guess I didn't write the band gap and Db range correctly. I'm trying to start from scratch now. What should the band gap be for the human voice? What should the db range be? and I am using the data obtained from the FDATOOL window with the y=filter(b,a,x) function, do you think I am doing it wrong? Is a Band-Pass filter alone enough? Is there another way?

Answer 1

[Meet]

Hi Mustafa,

Let me go through each query one by one:

1. The typical frequency range for human speech is about 300 Hz to 3000 Hz. So, when designing a band-pass filter, you should set the passband to cover this range.
2. For speech, you might aim for a passband ripple of around 1 dB or less and a stopband attenuation of at least 40 dB to ensure clarity and effective noise reduction.
3. You are doing it right by designing the filter in "FDATool" and then applying it using the "filter" function. This function applies the filter coefficients b and a to your signal x.
4. A band-pass filter is a good start, but depending on the noise characteristics, you might need additional processing. If the signal is recorded using just one microphone, you can use methods such as spectral subtraction. This method is more suitable for constant noise, like the noise from a fan or an idle engine. Other methods rely on statistics and perceptual models of speech. If the signal is recorded with several microphones, you can use blind source separation for separating the (speech) signals. As it stands today, you would not get perfect results. The typical end-result is always a trade off between "noise" and clarity of the speech signal of interest. More "noise" suppression --> more degradation of the signal of interest.

Happy filtering!