In the clinical speech world, the term speech intelligibility refers to the ability of a typical listener to understand a person’s speech (Yorkston et al., 1996). Many people with speech disorders find it difficult to be understood by others. Perceptual measures can also involve listeners making more objective judgements about what they heard, for example, being forced to choose between two similar sounds or being asked to write down what they heard a speaker say. These ratings often take the form of descriptors (e.g., poor-fair-good), numbers (e.g., 1 to 10), or a position on a line with anchors (e.g., low to high). That is, an individual is asked to rate some aspect of their own or another person’s speech. Perceptual measures can take many forms, but one method we often rely on is the use of rating scales. The last two are examples of perceptual measures. Listener perception: Do the people that listen to me speak notice differences? Do these differences impact how much of my message they are able to understand?.Self-perception: Do I, as a talker, notice differences in my own speech compared to those around me? Do those differences bother me? Do they impact my communication?.Acoustic analysis: What can the physical properties of the speech signal tell us about a person’s spoken communication? Are certain acoustic parameters deviant compared to otherwise healthy speakers? Can these acoustic differences help to explain the differences noticed by the speaker or by listeners?.Different aspects of spoken communication can be measured in multiple ways. That is, atypical speech production that arises due to abnormalities in the brain (for example, changes in the brain associated with Parkinson’s disease). In my research, I work with people who have neurological speech disorders. Come along for the joy of learning where the heck the backspace key lives in Praat, and other fun discoveries.
#PYTHON PRAAT SCRIPT SERIES#
Future posts in this series will introduce a script that allows for a listener to transcribe speech in Praat, rather than (or in addition to) rating it along a VAS.
An upcoming post will provide a step-by-step of how it was written in Praat (in large part as a reminder to future me, in the inevitable case that I eventually forget what I did).This post will walk through the motivation for automating listener tasks like this and a demo of how it works.
It was designed to be used for administering speech intelligibility tasks in our lab, whereby participants hear an audio clip of somebody speaking, and must rate how intelligible (i.e., understandable) it was to them along a scale, for example, from “low” to “high.” It is adaptable to any perceptual experiment designed to be delivered in VAS form.
0 kommentar(er)
