Jens Bo Nielsen1, Torsten Dau. 1. Centre for Applied Hearing Research, Department of Electrical Engineering, Technical University of Denmark, Lyngby, Denmark. jbn@elektro.dtu.dk
Abstract
OBJECTIVE: A Danish version of the hearing in noise test (HINT) has been developed and evaluated in normal-hearing (NH) and hearing-impaired (HI) listeners. The speech material originated from Nielsen & Dau (2009) where a sentence-based intelligibility equalization method was presented. DESIGN: In the present study, the speech material was evaluated for naturalness and a subset of sentences selected. The new sentence lists were validated, and after three weeks retested. An additional experiment investigated how recollection of sentences affected the listeners' performance. STUDY SAMPLE: 16 NH and 16 HI listeners participated in the validation and retest. Twelve HI listeners participated in the experiment on recollection. RESULTS: The average speech recognition threshold in noise (SRT(N)) for the NH listeners was -2.52 dB, with an overall standard deviation of 0.87 dB. The within-subject standard deviation was similar for the NH and the HI listeners. In the retest, the SRT(N) decreased by 0.4 dB in both groups. CONCLUSIONS: The Danish HINT consists of 10 test lists and three practice lists each containing 20 sentences. The validation results are comparable to those of other versions of HINT. The test seems equally reliable for NH and HI listeners. After three weeks, reliable results can be obtained when sentence lists are reused with the same listeners.
OBJECTIVE: A Danish version of the hearing in noise test (HINT) has been developed and evaluated in normal-hearing (NH) and hearing-impaired (HI) listeners. The speech material originated from Nielsen & Dau (2009) where a sentence-based intelligibility equalization method was presented. DESIGN: In the present study, the speech material was evaluated for naturalness and a subset of sentences selected. The new sentence lists were validated, and after three weeks retested. An additional experiment investigated how recollection of sentences affected the listeners' performance. STUDY SAMPLE: 16 NH and 16 HI listeners participated in the validation and retest. Twelve HI listeners participated in the experiment on recollection. RESULTS: The average speech recognition threshold in noise (SRT(N)) for the NH listeners was -2.52 dB, with an overall standard deviation of 0.87 dB. The within-subject standard deviation was similar for the NH and the HI listeners. In the retest, the SRT(N) decreased by 0.4 dB in both groups. CONCLUSIONS: The Danish HINT consists of 10 test lists and three practice lists each containing 20 sentences. The validation results are comparable to those of other versions of HINT. The test seems equally reliable for NH and HI listeners. After three weeks, reliable results can be obtained when sentence lists are reused with the same listeners.