The effects of frequency lowering and spectral degradation on low pass filtered speech

Author(s):  
Mark Alexander Leung
2019 ◽  
Vol 40 (1) ◽  
pp. 3-17 ◽  
Author(s):  
Carina Pals ◽  
Anastasios Sarampalis ◽  
Mart van Dijk ◽  
Deniz Başkent

2020 ◽  
Vol 24 ◽  
pp. 233121652097563
Author(s):  
Christopher F. Hauth ◽  
Simon C. Berning ◽  
Birger Kollmeier ◽  
Thomas Brand

The equalization cancellation model is often used to predict the binaural masking level difference. Previously its application to speech in noise has required separate knowledge about the speech and noise signals to maximize the signal-to-noise ratio (SNR). Here, a novel, blind equalization cancellation model is introduced that can use the mixed signals. This approach does not require any assumptions about particular sound source directions. It uses different strategies for positive and negative SNRs, with the switching between the two steered by a blind decision stage utilizing modulation cues. The output of the model is a single-channel signal with enhanced SNR, which we analyzed using the speech intelligibility index to compare speech intelligibility predictions. In a first experiment, the model was tested on experimental data obtained in a scenario with spatially separated target and masker signals. Predicted speech recognition thresholds were in good agreement with measured speech recognition thresholds with a root mean square error less than 1 dB. A second experiment investigated signals at positive SNRs, which was achieved using time compressed and low-pass filtered speech. The results demonstrated that binaural unmasking of speech occurs at positive SNRs and that the modulation-based switching strategy can predict the experimental results.


1994 ◽  
Vol 78 (1) ◽  
pp. 348-350 ◽  
Author(s):  
Donald Fucci ◽  
Steve Domyan ◽  
Lee Ellis ◽  
Daniel Harris

17 subjects provided magnitude estimations in the form of quality judgments of a filtered speech stimulus which was a nonsense sentence containing all of the consonants of English from Fairbanks. It was presented to subjects at 8 high-pass and 8 low-pass filtering conditions. Consistent magnitude estimations to the filtered stimulus were similar for both conditions. Also, for both conditions, subjects' numerical responses consistently increased in value as stimulus quality was judged to be poorer.


1981 ◽  
Vol 90 (6) ◽  
pp. 543-545 ◽  
Author(s):  
Kimberly Hoffman-Lawless ◽  
Robert W. Keith ◽  
Robin T. Cotton

The present study was designed to determine whether auditory processing disorders are present in children with documented middle ear effusion (MEE) that required surgical treatment. Children with previous MEE and control subjects, in two age groups, were studied using five tests of auditory processing abilities: low-pass filtered speech, staggered spondaic word test, speech in noise, auditory sequential memory, and sound blending. Results found differences in groups at mean age 7 on the filtered speech test, but no statistical differences were found on any other test at age 7 or on any test at mean age 9. The results indicate that well-managed MEE appears to have no long-term effects on children acquiring this disease.


2015 ◽  
Vol 58 (3) ◽  
pp. 590-600 ◽  
Author(s):  
Maria V. Kondaurova ◽  
Tonya R. Bergeson ◽  
Huiping Xu ◽  
Christine Kitamura

Purpose The affective properties of infant-directed speech influence the attention of infants with normal hearing to speech sounds. This study explored the affective quality of maternal speech to infants with hearing impairment (HI) during the 1st year after cochlear implantation as compared to speech to infants with normal hearing. Method Mothers of infants with HI and mothers of infants with normal hearing matched by age (NH-AM) or hearing experience (NH-EM) were recorded playing with their infants during 3 sessions over a 12-month period. Speech samples of 25 s were low-pass filtered, leaving intonation but not speech information intact. Sixty adults rated the stimuli along 5 scales: positive/negative affect and intention to express affection, to encourage attention, to comfort/soothe, and to direct behavior. Results Low-pass filtered speech to HI and NH-EM groups was rated as more positive, affective, and comforting compared with the such speech to the NH-AM group. Speech to infants with HI and with NH-AM was rated as more directive than speech to the NH-EM group. Mothers decreased affective qualities in speech to all infants but increased directive qualities in speech to infants with NH-EM over time. Conclusions Mothers fine-tune communicative intent in speech to their infant's developmental stage. They adjust affective qualities to infants' hearing experience rather than to chronological age but adjust directive qualities of speech to the chronological age of their infants.


1987 ◽  
Vol 96 (1_suppl) ◽  
pp. 58-62 ◽  
Author(s):  
A. Boothroyd

The ability to perceive four suprasegmental and eight segmental speech pattern contrasts was measured in several subject groups. Mean performance of a group of House 3M cochlear implantees was similar to the mean performance of a group of Nucleus implantees with a trend toward better perception of suprasegmentals in the former and better perception of segmentals in the latter. Intersubject variability was higher in the Nucleus implantees, the best subject achieving an open set phoneme recognition score of 40%. Data from a single subject showed considerably superior performance in an aided ear with a 95-dB hearing loss than in an ear implanted with the House 3M device. Data from normal subjects listening to low pass filtered speech provide a basis for evaluating the results obtained by implantees. The results, though limited, illustrate the potential value of analytical tests such as the speech pattern contrast test used in these studies.


2012 ◽  
Vol 23 (03) ◽  
pp. 192-205
Author(s):  
Jeffrey S. Martin ◽  
Keiko Y. Gibson ◽  
Lisa C. Huston

Background: Assessment of interaural asymmetry (IA) on dichotic listening tests becomes challenging when accuracy on one or both ears are at ceiling. Cognitive task demands are often increased to avoid this situation, raising the possibility that non-auditory-specific processes will further influence IA estimation. Purpose: To investigate the utility of low-pass filtered dichotic speech stimuli (dichotic filtered words [DFWs]) as a potential auditory-specific technique for increasing task difficulty in the dichotic listening paradigm. Research Design: A prospective experimental study investigating differences in IA observed for a group of young adults on three dichotic listening tasks differentiated on the basis of stimulus type (unfiltered words, DFWs) and/or mode of test administration (divided attention [DIV]; directed attention [DIR]). Study Sample: Thirty-two adults between 18 and 34 yr of age participated in the study. Data Collection and Analysis: Accuracy scores on each ear and normalized values of IA (via laterality index) served as dependent measures evaluated in a repeated-measures ANOVA design. Correlation and regression analyses were carried out to investigate potential relationships between participants’ behavioral accuracy to experimental stimuli and the magnitude of observed IA. Behavioral profiles of IA were constructed for each participant. Results: Using a traditional DIV test mode, DFW stimuli produced a significantly larger right-ear advantage (REA) as compared to unfiltered speech. The magnitude of the average REA to DFW stimuli was not statistically different between DIV and DIR test modes for the group. Inspection of individual performances, however, revealed that out of the participants who demonstrated meaningful IA during DIV, approximately half produced the same or greater amount of IA during DIR whereas half showed a relative decrease. Participants’ accuracy to left-sided stimuli was most related to the magnitude of observed IA across tasks. Conclusions: The utility of low-pass filtered speech may be helpful in the evaluation of IA obtained during dichotic testing.


Sign in / Sign up

Export Citation Format

Share Document