February 16, 2006.
Music 209 : Calendar : Week 5Techniques for altering one sample to match another in preparation for a splice.
We began the lecture by defining what an ideal pitch shifting algorithm should do (see slides for details). We then showed some tutorial material on the mathematics of scales: this webpage aimed at high school students covers scale basics in a (very!) unintimidating way.
We then covered the basics of the psychoacoustics of pitch perception. This interactive tutorial is a good way to learn more about the ideas presented in class. For more advanced material, see this review article by William Hartmann.
This part of the talk also included a slide on the pitch properties of bell sounds. This article is a good introduction to synthesizing bell sounds, and includes the spectral plot I showed in class.
We then discussed signal processing algorithms for pitch shifting. A good tutorial on these algorithms (the time-domain algorithms presented in class, and also spectral algorithms which we will cover later in the course) is available here ( [1] [2] [3]).
The primary references for this work are:
Lent K. (1989). An efficient method for pitch shifting digitally sampled sounds. Computer Music Journal Vol. 13 No. 4.
Bristow-Johnson, Robert (1995). A Detailed Analysis of a Time-Domain Formant-Corrected Pitch-Shifting Algorithm. The Journal of the Audio Engineering Society. Volume 43 Number 5 pp. 340-352.
This part of the talk also included slides on the spectral properties of instruments. This article is a good conceptual introduction to formants in sound synthesis. These articles [1] [2] apply formant concepts to violin synthesis.
We ended the first part of the lecture by reviewing the vocal project ideas described here.