Pitchtracker

Pitch Tracker reads an audio file of arbitrary length and produces a PostScript document with precise graphs of the fundamental tone frequency. The program can only analyze monophonic audio sources.

In musicological context it is often interesting to find fundamental tone frequency with high accuracy, and to produce publishable graphs. Analysis of intonation, microtonality, use of vibrato and other ornamentation require data with a higher resolution than you typically get from commercial pitchtrackere. With our program you can easily see rapid variations in pitch down to a few cents.

The figure below shows a typical graph produced by Pitch Tracker – an analysis of 15 seconds woman singing. Each horizontal dotted line corresponds to one quarter tone. Note the use of vibrato, and how individual notes start with glissando upwards.

We have developed a special algorithm for accurate pitch tracking. Although this algorithm is simple and relatively primitive, leaving it to be both robust, fast and precise. Especially works well for singing.

The algorithm is based on a fasevokoder that using FFT estimate accurately the frequency of the partials that are located within each analysis channel, ca. 40 times per second (fasevokoderen fail if multiple partials fall within the same channel). The problem is then to determine which partial which constitutes the fundamental. First we seek for the most powerful partial. Often this is the fundamental tone, but it can also be the first or second overtone. Therefore, we check if it is concentrated energy around half, third party or 2/3 of the frequency of the strongest partials, and otionalle choose an lower partial as keynote.

Pitchtracker is designed and programmed by Øyvind Hammer and later maintained by Kjetil S. Matheussen.