FFT of /a/ as in pot shows power spectral density and formant peaks

The one on left is the actual recorded signal, on right is the approximation via LPC

 

FFT of /a/ as in pot shows power spectral density and formant peaks

The one on left is the actual recorded signal, on right is the approximation via LPC

Spectrogram of /s/ as in send shows frequency content

The one on left is the actual recorded signal, on right is the approximation via LPC

As can be seen from comparing the left and right panes, the signal synthesized using LPC coefficients provide a somewhat close approximation to actual speech.