Introduction

Executing the Signal Analysis

Analysis Method : Fast Fourier Transform (FFT)

Several types of analysis can be executed to analyze the structure of a sound with AudioSculpt. We will use the Fast Fourier Transform method, which allows to represent the temporal evolution of the components – partials, or harmonics – of a signal.

Fourier Transform Principles and Other Analysis Methods MODIFIER LA REFERENCE

Setting the Analysis Type and Parameters

Select the Analysis/Sonogram Analysis menu item, or press Cmd + D. You can remember this shortcut, for you might resort to sonogram analysis frequently. A dialogue window opens.

We only need to focus on two parameters : the analysis method, and the window size[1].

  • In the dialogue window, choose the FFT – Fast Fourier Transform.

  • The default value of the window size is 4096 samples, which is quite relevant to harmonic sounds. The window we're talking about represents a segment of signal to analyze. This means that at each step of the analysis, a calculus is performed upon segments of 4096 samples.

Window Step / Size MODIFIER LA REFERENCE
Window Size

The value of the window size should usually be equal to powers of 2: (2, 4, 8, 16, 32, 64, 128, 512, 1024, 4096, 8192...) and superior to 512.

  1. Window Size / Window Step
A propos...IRCAMRéalisé avec Scenari