Frequency problem between reality and analysis

Bonjour ; je sais que l'oreil humain est capable d'entendre les frequences entre 20 et 20k (hz) , mais quand j'analyse un signal sous Matlab [Data,Fs] il donne toujours que Fs=42000 hz . J'arrive pas a comprendre comment ca marche

[Paul]

Hi Sofiane,

You are more likely to get a response if you post an example with data and code that shows what needs to be clarified.

[sofiane benhamza]

Hello , [D,fs]=audioread('example.mp3'); %D : data matrice (2columns) %fs = 42000 to 46000 and i can't get it why we still able to hear music if it's frequence is greater than 20e+3hz .

[Bjorn Gustavsson]

The sampling-frequency doesn't determine the frequency of the signal, only the maximum bandwidth of the signal that can be determined.

[sofiane benhamza]

So , we have 2 frequencies , the sampling (which is depend of speed of playing ) and the sound frequency which is still mystery , Thanks for your help <3

[Bjorn Gustavsson]

To look at the frequencies of the sound you could look at its spectrogram. How to calculate and display that have a look at the help and documentation of the spectrogram function.

Answer 1

[Jonas]

the nyquist theorem tells that you need to sample a wave at least two times to reproduce it without errors. Or in other words, if you want to catch frequencies up to 20kHz, you need a sampling (recording) frequency of at least 2*20kHz=40kHz.

If there are really high amounts of data in the high range it is recommended to either use an analog lowpass filter (anti aliasing) filter before it is sampled or to signfificantly incerease the sampling frequency (e.g. 100k) to avoid back folding (aliasing) of high frequency contents

[sofiane benhamza]

I don't know how to thank you . I got your words , Thaanks <3 men.