PHONANIUM plug-ins with scripts for acoustic voice/speech analysis in the program Praat are developed to help all clinicians and students (especially in communication sciences and disorders) understanding the process of speech production and undertaking objective voice/speech assessment, regardless of their background, resources and/or location.
Why … acoustic analysis?
Acoustic assessment of voice (and speech) signals has some general advantages. First, it concerns non-invasive methods and can therefore easily be implemented in clinical as well as research protocols. Second, it is considered to be objective, as long as a standardized protocol is followed in terms of recording and signal processing. Of course the clinician and the subject/patient can influence the output of an acoustic analysis, but items like hardware and signal processing are quite neutral and unbiased. Acoustic analysis is therefore ideal for both quantitative and qualitative tracking of voice/speech treatment outcomes and specific speech-related aspects across time. Third, many aspects of the voice/speech signal are associated with specific aspects of speech physiology and/or speech perception, for example: pitch is predominantly related to fundamental frequency, loudness is primarily related to intensity level, vowel type is mainly determined by the first and the second formants, etc. It is therefore logical to use acoustic methods to learn about speech production and to show students how speech is produced and differs between persons. An additional advantage of acoustic methods is that it has a relatively low cost in comparison to other assessment techniques and therefore is quite accessible. However, commercially available tools often still require substantial financial resources, which may inhibit clinicians to purchase such systems and to apply acoustic methods.
It thus makes a lot of sense to apply acoustic analyses in the clinical assessment of voice and speech disorders as well as in the scientific study of (disordered) voice and speech signals.
Why … program Praat?
There are many tools, programs and solutions for acoustic analysis of voice and speech signals. Some are commercially available, whereas others are freely available/downloadable. The program Praat (Paul Boersma and David Weenink, Intsitute for Phonetic Sciences, University of Amsterdam, The Netherlands), however, outperforms many of the other programs because of its combination of the following assets (Maryn et al., 2017). First, Praat itself can be downloaded for free at www.praat.org. The GNU General Public License under which this program runs, can be downloaded at www.fon.hum.uva.nl/praat/GNU_General_Public_License.txt. Second, it is available for the most popular and commonly installed computer operating systems (i.e., Windows, Macintosh, and Linux). Praat can therefore be applied regardless the operating platform used by the voice and speech clinician/scientist. Third, the operator/clinician controls many of the parameters of the acoustic analyses and therefore isn’t completely reliant on unmodifiable settings as pre-defined by the manufacturer. This increases the clinician’s freedom and control, and enables experimenting with parameter arguments. Fourth, many clinically relevant markers (i.e., acoustic voice measures related to fundamental frequency, intensity level, formant, perturbation, spectral configuration, cepstral configuration, etc.) and all their relevant statistics are readily obtainable in Praat. Fifth, clinicians with questions and doubts can consult Praat’s extensive manual and help function as well as its online discussion forum (https://uk.groups.yahoo.com/neo/groups/praat-users/info). Sixth, Praat can be applied to prepare (i.e., select, extract, zoom in and out, reverse, modify intensity and/or fundamental frequency, resample, annotate, filter, concatenate, etc.) voice and speech signals and thus editing them for whatever clinical or scientific purpose needed. Seventh, Praat can be employed to draw various graphs (e.g., oscillogram, spectrum, spectrogram, cepstrum and cepstrogram) and statistics relevant to voice and speech evaluation and research. Eighth, there are numerous Praat scripts for automated analysis of the acoustic voice signal. When coupled to (i.e., installed behind) a button in Praat, scripts provide single-button prompts to sometimes particularly complex procedures that otherwise would be way too time- and labor-consuming to apply in the clinic. Nineth, the program Praat has already proven to be exceptionally interesting for voice/speech analysis in a multitude of research papers across study areas (i.e., speech-language pathology, phonetic sciences, linguistic sciences, medicine, etc.).
Based on these items, it is fair to appreciate Praat as a highly valuable program in the study of voice and speech signals and disorders.
Why … scripts?
A script for the program Praat refers to an executable text mainly consisting of menu and action commands. Running such a script, automatically prompts all these operations. As such, each script can be regarded as a separate subprogram telling Praat what actions to do and in what order to do them. As mentioned above, a script can be coupled to a button in the fixed or the dynamic menu of the program Praat. It offers a single-button trigger to sometimes very complex procedures that otherwise would be way too time-consuming and labor-intensive in the realm of clinical practice. Implementing scripts dramatically increases the user-friendliness and the efficiency of the program Praat. Furthermore, it induces standardization and consistency in analysis methods, because the algorithms, the analyses and the drawings always run in the same way.
A PHONANIUM plug-in for the program Praat typically contains one analysis script together with a setup.praat script. Saving the plug-in in the Praat preferences directory, will automatically couple the analysis script to a button when the program Praat starts up.
Why … Phonanium?
There are many scripts available for the program Praat. However, PHONANIUM’s scripts are written especially for voice/speech clinicians and scientists … by voice/speech clinicians and scientists. This implies that they are ready for use in a clinical assessment and treatment outcomes protocol. To learn about voice and speech of subjects/patients, these scripts actually accompany text and video with background information on the different ‘elements of phonatory sound‘. They aim at operationalizing relevant acoustic theory in clinical practice. Most of the scripts from PHONANIUM have been used extensively and routinely in the clinic, and are therefore clinically validated. Some scripts have even been validated scientifically (e.g., the Acoustic Voice Quality Index).
PHONANIUM offers two kinds of scripts. First, there is a report-/dossier-/file-based script. This script writes the personal information of the subject to be assessed to a table, in order to retrieve this information later from this table whenever it is asked for by the clinician in the course of another script. Working with such a non-measurement in-between script additionally increases the efficiency of working with Praat. Second, there are measurement scripts. These scripts intend to measure specific elements of phonatory sound, such as: vocal fundamental frequency (i.e., vocal fo), vocal intensity level (i.e., vocal IL), vocal range estimation, and voice quality. They not only provide a numerical summary (e.g., median vocal fo), but they also bring forth a set of relevant statistics (e.g., standard deviation and interquartile range in vocal fo) and several applicable graphs (e.g., a fo-time-plot aligned with an oscillogram and a fo-histogram). Providing information in a combination of both numerical and graphical outputs, and supported by video material, PHONANIUM offers an as complete as possible perspective on the voice/speech element under investigation and is hypothesized to strengthen the interpretation of this element.
References
Maryn Y (2017). Practical acoustics in clinical voice assessment: a Praat primer. Perspectives of the ASHA Special Interest Groups SIG 3, 2(Part 1), 14-32.
Are you interested in following or organizing a course on acoustic voice analyses?
Are you interested in the development of a script for acoustic voice/speech analysis?
Is there something you like to ask us? Or do you have an idea to share with us?
