A Comparative Study of the Precision of Carstens and Northern Digital Instruments Electromagnetic Articulographs Purpose This study compares the precision of the electromagnetic articulographs used in speech research: Northern Digital Instruments' Wave and Carstens' AG200, AG500, and AG501 systems. Method The fluctuation of distances between 3 pairs of sensors attached to a manually rotated device that can position them inside the measurement ... Research Article
Research Article  |   February 01, 2017
A Comparative Study of the Precision of Carstens and Northern Digital Instruments Electromagnetic Articulographs
 
Author Affiliations & Notes
  • Christophe Savariaux
    Univ. Grenoble Alpes, CNRS, GIPSA-Lab, Grenoble, France
  • Pierre Badin
    Univ. Grenoble Alpes, CNRS, GIPSA-Lab, Grenoble, France
  • Adeline Samson
    Univ. Grenoble Alpes, CNRS, Laboratoire Jean Kuntzmann, Grenoble, France
  • Silvain Gerber
    Univ. Grenoble Alpes, CNRS, GIPSA-Lab, Grenoble, France
  • Disclosure: The authors have declared that no competing interests existed at the time of publication.
    Disclosure: The authors have declared that no competing interests existed at the time of publication. ×
  • Correspondence to Christophe Savariaux: christophe.savariaux@gipsa-lab.grenoble-inp.fr
  • Editor: Julie Liss
    Editor: Julie Liss×
  • Associate Editor: Kate Bunton
    Associate Editor: Kate Bunton×
Article Information
Research Issues, Methods & Evidence-Based Practice / Speech / Research Articles
Research Article   |   February 01, 2017
A Comparative Study of the Precision of Carstens and Northern Digital Instruments Electromagnetic Articulographs
Journal of Speech, Language, and Hearing Research, February 2017, Vol. 60, 322-340. doi:10.1044/2016_JSLHR-S-15-0223
History: Received June 24, 2015 , Revised March 1, 2016 , Accepted June 23, 2016
 
Journal of Speech, Language, and Hearing Research, February 2017, Vol. 60, 322-340. doi:10.1044/2016_JSLHR-S-15-0223
History: Received June 24, 2015; Revised March 1, 2016; Accepted June 23, 2016
Web of Science® Times Cited: 3

Purpose This study compares the precision of the electromagnetic articulographs used in speech research: Northern Digital Instruments' Wave and Carstens' AG200, AG500, and AG501 systems.

Method The fluctuation of distances between 3 pairs of sensors attached to a manually rotated device that can position them inside the measurement volumes was determined. For each device, 2 precision estimates made on the basis of the 95% quantile range of these distances (QR95) were defined: The local QR95 was computed for bins around specific rotation angles, and the global QR95 was computed for all angles pooled.

Results For all devices, although the local precision lies around 0.1 cm, the global precision is much more worrisome, ranging from 0.03 cm to 2.18 cm, and displays large variations as a function of the position of the sensors in the measurement volume. No influence of the rotational speed was found. The AG501 produced—by far—the lowest errors, in particular concerning the global precision.

Conclusions The local precision can be considered suitable for speech articulatory measurements, but the variations of the global precision need to be taken into account by the knowledge of the spatial distribution of errors. A guideline for good practice in EMA recording is proposed for each system.

Acknowledgments
This work has been partially funded by the project ANR-13-TECS-0011-06 “e-SwallHome – Swallowing & Respiration: Modelling & e-Health at Home,” by the ANR-11-LABX-0025-01 LabEx Persyval-LAB, funded by the French ANR agency, and by the PEPS program funded by the Univ. Grenoble Alpes COMUE and the French CNRS. We are very thankful to all colleagues who helped us in collecting data at various places: Slim Ouni and Yves Laprie (LORIA/PAROLE, Nancy, France), Brigitte Wrobel-Dautcourt (LORIA/MAGRIT, Nancy, France), Thierry Legou and Noël Nguyen (Laboratoire Parole & Langage, Aix-en-Provence, France), and Ulrich Szagun (Carstens, Germany). We thank also the companies NDI (Germany) and Carstens (Germany) for their help in this study. We are also much indebted to Jean-Luc Schwartz (GIPSA-lab/PCMD, Grenoble, France) and Rémy Drouilhet (Laboratoire Jean Kuntzman, Grenoble, France) for useful advice on statistical analysis.
Order a Subscription
Pay Per View
Entire Journal of Speech, Language, and Hearing Research content & archive
24-hour access
This Article
24-hour access