We record sound using a microphone. A microphone takes discreet measurements of air pressure. The number of recordings it takes per second called the sample rate. A tipical sample rate is 44,100 meansurements per second.
If we plot the array of numbers from our microphone it looks like this.
[Waveform of Recording]
If we play the same recording back at a larger sample rate it is faster. It also makes the pitch of the audio higher.
Different sounds produce different wave forms. Here is a plot of me saying “aaah” and another off me saying “eee”.
[Plots]
We can use apply a Fourier transform to the audio stream to decompose it into different frequencies.
Record yourself saying something and see the plots below.
[Interactive component]
Tags: AI