Tone superimposition technique in Speech Sciences: a tutorial

Abstract

In the literature, we encounter papers reporting manipulating pitch contours in speech tokens for a specific problem to be addressed in experiments (e.g., learning pitch patterns superimposed onto a pseudo-syllable), usually in the field of Speech Perception and Spoken Word Recognition. This type of research often tests listeners’ perceptual and processing skills in tonal languages (e.g., Mandarin, Thai, etc.), and requires superimposing a pitch contour onto a spoken syllable. However, very few studies reported in detail how this critical manipulation was done to meet specific experimental needs. In addition, there was neither specific guideline or description of the techniques being used, nor how ‘natural’ these manipulated tokens sounded in a particular language upon speech synthesis. Because this technique is crucial in establishing the conclusions in various studies, here, we will demonstrate our method of establishing this technique of tone superimposition (i.e., lexical tones in Mandarin) onto English syllables. In line with the open science model, we will also show our stimuli and procedures via OSF for readers to evaluate the validity of this technique. Manipulating the pitch contour in a spoken syllable can be complicated and change the perception of the spoken syllable in a significant way. Thus, we will also show the important factors to be considered in this process for doing research in Speech Sciences.