MUST 1220 Lecture Notes - Lecture 12: Fast Fourier Transform, Frequency Domain

• School: Northeastern University
• Department: Music Technology
• Course: MUST 1220
• Professor: Frengel Mike

Document Summary

Each block of data = frame, which has 2 things: a magnitude spectrum that depicts amplitude of every analyzed frequency component and, a phase spectrum that shows the initial phase value for every frequency component. Algorithm to transform a signal from time domain into frequency domain. Frequency bin = collects energy/amplitude from small frequency range. Time/frequency uncertainty if we want high resolution in time domain, we sacrifice frequency resolution. We can tell that an event occurred at a precise time, but we cannot say exactly what frequencies it contained and vice versa. Ca(cid:374)"t fi(cid:454) (cid:272)lippi(cid:374)g (cid:271)(cid:455) lo(cid:449)eri(cid:374)g (cid:448)olu(cid:373)e. keeps flat li(cid:374)e fro(cid:373) (cid:272)lippi(cid:374)g = (cid:272)ha(cid:374)ges ti(cid:373)(cid:271)re a sound. Finds loudest point and difference between that and max. Raises loudest point to max, adjusts rest accordingly. Last stage of production process = normalize to make it sound as loud as possible!