Fourier transform- a general question

     Fourier Transform is used for representing non-periodic waves in terms of cosines and sines. Yesterday when I recorded my voice I saw sinusoidal waves in the plot of my voice. Of course I know that they are due to vibration in the microphone which is converted into electrical energy. Does this sinusoidal wave has anything to do with the Fourier transform? I am getting this doubt because I learnt that noises are at high frequencies and when I saw the plot of the noise (ie portion that does not contain my voice) they seem to be high frequency sinusoidal waves.

Just a doubt that popped into my mind. That's all

Thanks in advance.

Answer 1

[Wayne King]

Hi, you are correct. When you are referring to the noise parts of a speech waveform, you are most likely talking about sounds like 's' or 'sh' as in say or sheep. These are called fricatives in speech and the energy in these fricatives is isolated in higher frequencies than the enery in the voiced, periodic sounds (vowels).

There is a Fourier theory for random processes.

By the way, for analyzing speech, you'll want to learn about the spectrogram, so you can visualize how the spectrum changes over time.

[i Venky]

I was talking about the noise that we get when I don't say anything. That noise was like a sinusoidal wave oscillating at a very high frequency.

[Wayne King]

yes, that is correct. Actually the spectrum of white noise (like you get with randn()) contains equal power contributions at all frequencies

[Image Analyst]

I'm not sure how that's correct. First I'd like to know if "i" was looking in the time domain or the frequency domain. Equal power at all frequencies would give a flat spectrum not one that looks like a sine wave. And a sine wave in the time domain would look like a spike in the frequency domain.

[Wayne King]

@imageanalyst I think the OP was just noticing that noise (when there is no DC shift) tends to jump from positive to negative values and vice versa, that's what I assume he meant by "noise was like a sinusoidal wave oscillating at a very high frequency."