This weekend-length workshop is specifically designed for engineers or developers working with audio who are interested in deepening their background in digital audio theory. The workshop covers the implementation of the Fast Fourier Transform (FFT) in digital signal processing, focusing on practical spectrum analysis, sound synthesis with spectral models, and signal processing using the FFT.
The course begins with a review of Fourier theory, which includes a discussion of important Fourier Transform properties. The Short-Time Fourier Transform (STFT) of time-varying audio signals is presented next. Various analysis windows are discussed along with their application in audio spectrum analysis and filter design. Both interpretations of the STFT, as either an overlap-add, or a filterbank summation procedure, will be addressed. Applications of the STFT for time/frequency analysis, spectral modification, spectrum analysis trade-offs as well as time-varying and nonlinear modifications will be studied.
The next section of the class is concerned with sinusoidal modeling
and additive synthesis of audio signals. Specific topics will
include various analysis and synthesis techniques such as the
channel and phase vocoders, and Spectral Modeling Synthesis (SMS).
We will focus on the time compression and expansion applications
as well as band-limited interpolation. Depending on the interest
of the participants, potential topics of the course may include
wavelets and MPEG data compression.
