In an automobile, it is difficult to reliably estimate intelligibility of annotation speech of an automobile navigation system using various loud speakers due to variations in interior noise and reverberation. In this paper, we recorded speech and noise separately in actual automobile interior using a binaural mannequin. Actual environment inside an automobile was simulated by mixing pre-recorded test samples, and presented via headphones. Articulation evaluations in an actual automobile and simulated environment were carried out and compared. The results were shown to be dependent on the speaker type; syllable articulations agree very well between simulated and actual environment when dynamic speakers are used. With piezoelectric speakers, the articulation scores in the actual environment were generally constantly low when noise was present, while in the simulated environment, the scores decreased gradually with noise level.
Towards reliable articulation evaluation inside automobiles using binaural mannequins
Zur Bewertung einer zuverlässigen Sprachverständlichkeit in Automobilen mit Hilfe binauraler Dummies
2009
7 Seiten, 7 Bilder, 8 Quellen
Conference paper
English
Evaluation comparative des performances des mannequins
Automotive engineering | 1983
|Optimization and Minimization of Boundary Mannequins
British Library Conference Proceedings | 2005
|Optimization and Minimization of Boundary Mannequins
SAE Technical Papers | 2005
|Fitting Research: Mannequins for improved garment fit
British Library Online Contents | 1994
Simplified models of the vibration of mannequins in car seats
Online Contents | 2003
|