Visual Representations of Acoustic Data: A Survey and Suggestions

ABSTRACT

Visual representations of acoustic data are becoming more common in Conversation Analysis (CA) and Interactional Linguistics (IL) research. This article provides a survey of visual representations of acoustic data in the journal Research on Language and Social Interaction. Shortcomings in their preparation and use are identified and discussed. Comparisons are made with visual representations prepared by expert phoneticians. Suggestions are made as to how visual representations could be prepared and used more effectively to support CA/IL researchers’ claims and to allow readers to independently verify them.

Notes

¹ Some search terms returned no hits: loudness, duration, rhythm, creak, tempo.

² The average number of articles per year since 2001 is 20 (sd = 4).

³ Estimates from visual representations have been arrived at by taking measurements from the PDF versions of the source using Preview on Mac OS(X). Estimates are identified by ≈ in the text.

⁴ Nonmodal phonation can produce very low frequencies (e.g., creak) and very high ones (e.g., falsetto). A decision therefore needs to be made as to whether these measures are included in estimates of range. Their inclusion could radically alter the visual representation of the speaker’s range. Since speakers usually only produce talk in those registers infrequently, a practical solution is to exclude such stretches from the estimation of the speakers’ pitch range (hence locutions such as “normal speaking range”) and then adjust individual pitch traces when necessary to show nonmodal portions outside the normal pitch range (see, for example, G. Walker, 2016, Figure 5a, Figure 6).