Workshops on FFT & Max
FFT is a scary acronym for a scarier algorithm, the Fast Fourier Transform. Scary, but when it comes to sound processing, this is the door to go beyond the mirror of time domain. Sound is air in movement, atoms moving through space & time. When you push the door of the spectral domain (using the FFT algorithm), you access the frequencies that constitute the sound. In Shanghai Conservatory, on October 17th and 18th, I taught how to use the FFT in the software environment Max.
Free tools for spectral sound processing
Where the first day introduced the theory, the second day was a hands-on workshop. We spent quite some time on the distinction between time domain and spectral domain in Max/MSP (made easy by the [pfft~] object).
- October 17th: topic introduction
- waveforms (LP vinyl, tape)
- time domain sound processing: speed & pitch link
- analog granulation, digital granulation
- spectral domain (spectrum, sonogram)
- have fun with graphical sound transformations
- October 18th: hands-on workshop
- real-time (denoiser, "noiser", frequency shifting, pitch shifting)
- freeze a sound
- elastic time, automatic slicing
But most importantly, I wanted the students to be able to use spectral sound processing even if they would not become Max/MSP specialists. You can consider my Max MSP patches as free spectral processing tools: download the Max Runtime for free, and download the patches (link updated 01/12/2010). One student asked me how to record the output of the patches. Of course, if you own Max, you can modify the patches and add a sound file recorder. But before investing in Max, you can use a digital audio routing system, such as Soundflower, or Jack Audio, and any application that records audio (Quicktime Pro, Audacity, GarageBand, etc...)