Drift was prototyped in 2016 by Robert Ochshorn and Max Hawkins with support from Marit MacArthur’s ACLS Digital Innovations Fellowship. It is a highly accurate pitch-tracker that also incorporates the forced alignment features of Gentle, visualizing a pitch trace over time and aligning it with a transcript. (Note: Gentle can also create rough transcripts of speech recordings.) Using an algorithm developed by Byung Suk Lee and Daniel P. W. Ellis at Columbia University to work with precise accuracy on the noisy, low-quality vocal recordings common in the audio archive, Drift measures what human listeners perceive as vocal pitch (the fundamental frequency, the vibration of the vocal cords, as measured in hertz) every 10 milliseconds in a given recording.
Running Drift on your own computer requires installing and opening Gentle first. Both work best on recordings of a single speaker; overlapping speakers will generate confusing pitch contours and erroneous Voxit prosodic measures.
About Voxit: Vocal Analysis Tools
Voxit is an open-source toolkit that performs automated analysis of vocal parameters in audio recordings of speech and song. It was developed by Lee M. Miller in collaboration with Marit J. MacArthur at the University of California, Davis, with additional input from Robert Ochshorn. It works best on recordings of a single speaker; overlapping speakers will generate erroneous Voxit prosodic measures. The standalone executable version of Voxit does not require a Matlab license. (Using the full editable codebase does.) Some of the Voxit prosodic measures in Drift4 differ slightly from the same prosodic measures in Voxit itself, because Voxit analyzes voiced speech and ignores unvoiced speech (e.g., fricatives and plosives, which do not involve the vocal cords), and thus is language-agnostic, while Drift analyzes words in English, based on Gentle transcripts. Click here for more information on Voxit, which is very useful for analyzing large archives of recordings and non-English language recordings.