PubMed · 5941761
The source did not provide an abstract. Follow the original record for more information.
Explore connections, maps & timelines
D E Broadbent. 1966. Two-state threshold model and rating-scale experiments.. https://doi.org/10.1121/1.1910047
Cite the original work for its findings. Save a collection to share your selection of sources.
KEEP EXPLORING
Listeners with sensorineural hearing loss often have difficulty discriminating stop consonants even when the speech signals are presented at high levels. One possible explanation for this deficit is that hearing-impaired listeners cannot use the information contained in the rapid formant transitions as well as normal-hearing listeners. If this is the case, then perhaps slowing the rate of frequency change in formant transitions might assist their ability to perceive these speech sounds. In the present study, sets of consonant plus vowel (CV) syllables were synthesized corresponding to /ba, da, ga/ with formant transitions for each set ranging from 5 to 160 ms in duration. The listener's task was to identify the consonant in a three-alternative, closed-set response task. The results for normal-hearing listeners showed nearly perfect performance for transitions of 20 ms and longer, whereas the shortest transitions yielded poorer performance. A group of eight hearing-impaired listeners pure-tone averages (PTAs) ranging from 30 to 62 dB HL) was also tested. The hearing-impaired listeners tended to show poorer performance than the normals for transitions of all durations; however, the performance of a few hearing-impaired subjects was equal to that of normals for the shortest-duration transitions. A strong inverse relation was observed between degree of hearing loss and improvement in score as a function of transition duration. These results suggest that increasing the duration of formant transitions for listeners with more severe hearing losses may not provide a helpful solution to their speech recognition difficulties.
Aided and unaided binaural versus monaural recognition of speech has been studied with more or less laborious methods when speech and noise sources are spatially separated. The efficacy of the new Finnish speech in noise test to show binaural advantage was evaluated in free field conditions. Thirteen subjects with normal hearing were tested measuring speech recognition thresholds in noise (SRTN) in anechoic chambre with different situations of spatial separation of speech and noise sources monaurally and binaurally. The noise source was fixed above the head (treating both ears similarly) and equidistant speech was delivered at different azimuths. SRTNs with mixed speech and noise without spatial separation served as normalizing reference. Differencies in SRTNs consistently followed the changing head and pinna shadow effects on the speech-to-noise ratio with a smaller binaural advantage. This new test seems to give quite economically the same results as those of previous more sophisticated investigations on speech recognition in free field situations.
The effect of presentation level on word recognition performance-intensity functions in continuous and interrupted broadband noise and in quiet was explored. Normal-hearing participants were tested at 30 and 50 dB sensation levels (SLs). Performance-intensity functions in both noises were determined at signal-to-noise ratios (S/Ns) of 10, 5, 0, -5, -10, -15, and -20 dB. There was no effect of SL presentation on word recognition performance in quiet (p = 0.136). A significant main effect was observed for S/N in both continuous and interrupted broadband noise conditions (p < 0.0001). Performance increased with increases in S/N regardless of the competing noise condition. A significant main effect for SL presentation was only observed in the interrupted noise condition (p = 0.0019). That is, performance was higher for the 50 SL for the interrupted noise condition only. It is suggested that the observed difference in performance in interrupted noise at different SLs offers additional evidence for level-dependent, temporal masking phenomena.