Inhibiting the Lombard effect.

The Lombard effect is the tendency to increase one's vocal intensity in noise. The present study reports three experiments that test the robustness of the Lombard effect when speakers are given instructions and training with visual feedback to help suppress it. The Lombard effect was found to be extremely stable and robust. Instructions alone had little influence on the response to the noise among untrained speakers. When visual feedback correlated with vocal intensity was presented, however, subjects could inhibit the Lombard response. Furthermore, the inhibition remained after the visual feedback was removed. The data are interpreted as indicating that the Lombard response is largely automatic and not ordinarily under volitional control. When subjects do learn to suppress the effect, they seem to do so by changing overall vocal level rather than their specific response to the noise.

The Lombard sign as a function of age and task.

The Lombard sign was investigated as a function of age and task. Twenty adults and 20 5-year-old children spoke in quiet and in the presence of a 90-dB SPL noise. Half the subjects labelled a series of pictures, half told stories about the pictures. All subjects increased vocal intensity in the presence of noise. Children increased intensity equally in both tasks. Adults increased intensity more during the story-telling task than during the labelling task.

Parameters of auditory feedback.

Adult speakers participated in three feedback conditions: Lombard, sidetone amplification, and delayed auditory feedback (DAF). All the procedures affected vocal intensity significantly, but only DAF caused changes in syllable rate. The three conditions did not provide equivalent data. Performance on any one of the feedback procedures was not predictive of performance on the others. Generalizations concerning auditory feedback must, therefore, take account of the particular method used and the parameter of speech studied.

Auditory system abnormalities in human albinos.

Auditory abnormalities were evaluated in human albinos using temporary threshold shift and dichotic listening tasks. Albinos exhibited a greater temporary threshold shift than did normally pigmented individuals after exposure to a fatiguing tone. In the dichotic listening task, albinos demonstrated a significantly larger right ear advantage than normally pigmented individuals. These results support earlier work indicating that hemispheric asymmetries are present in a albino auditory system and suggest that abnormalities may be present at the cochlear level as well.

The effects of different masking noises on children with /s/ or /r/ errors.

Children with /s/ or /r/ misarticulations were administered an articulation test for their target phoneme in three conditions. In the first condition, the children spoke in quiet. In the second condition, half of the children spoke while hearing an 80 dB SPL s-noise (high pass 2000 Hz) and the other half spoke under 80 dB of r-noise (low pass 1250 Hz). In the third condition, the noises were switched. It was assumed that a noise selected to overlap with the frequency range of the target phoneme would be most disruptive to articulation in the final stages of phoneme mastery. The hypothesis was borne out for the children with /s/ errors. These children had more errors under the s-noise than the r-noise. However, both noises caused a breakdown of articulation among children with /r/ errors. Though a feedback explanation can be adduced in both instances, there does seem to be a difference between these two speech defective populations.

Regulation of vocal intensity in the presence of feedback filtering and amplification.

Subjects read while their voices were fed back with 0 and 20 dB of amplification. In Experiment I, feedback was returned unfiltered or low-pass filtered with cut-off frequencies of 1000, 500, and 300 Hz. Subjects decreased vocal intensity with amplified feedback when the feedback was unfiltered or low-pass filtered with cut-off frequencies of 1000 or 500 Hz. The amplification effect disappeared when the cut-off frequency was 300 Hz. In Experiment II, feedback was low-pass filtered and amplified in the presence of a steady background noise. The presence of noise potentiated the amplification effect in the unfiltered condition. When filtering was introduced and/or the cut-off frequency was lowered, the amplification effect decreased. In Experiment III, subjects heard their voices unfiltered and high-pass filtered with cut-off frequencies of 500, 1000, and 2000 Hz. The amplification effect was the same across filtering conditions. It is suggested that subjects respond differently and high- and low-pass filtering because air- and bone-conducted feedback interact differentially in these conditions. It is hypothesized that changes may be mediated by self-perception of loudness or intelligibility.

The effects on speech of surgical premaxillary osteotomy.

Patients were tested before and after undergoing surgical premaxillary osteotomy for correction of skeletal and soft-tissue discrepancies. Both structural and speech measurements were made before surgery and for a year after surgery. The role of hearing and oral sensation in adaptation to surgery was evaluated. Immediately after surgery, speech was disrupted. Distortions of /s/ predominated. However, there were non long-term effects on speech. Hearing and oral sensation played little role in adaptation.

Delayed auditory feedback with children.

There is a controversy in the literature concerning the effects of delayed auditory feedback (DAF) on the speech of subjects of varying ages. In the current experiment the subjects were five-year-olds, eight-year-olds, and adult speakers who performed a sentence repetition task under: 0-delay, 250, 375, 500, and 625 msec of amplified delayed auditory feedback. All subjects performed the task under normal rate instructions and under instructions to speak as rapidly as possible. A developmental pattern emerged, with the youngest children significantly more affected by the DAF than the older children or the adults. There was only weak evidence for a critical delay interval that varied according to age of the subjects. Rate instructions had essentially no effect on the DAF or age patterns.

The effects of presentation of noise and dental appliances on speech.

The effects of dental appliances on speech were studied when subjects wore the appliances, both in quiet and in the presence of an intense noise. A group of 24 normal-speaking subjects read lists of syllables, words, and sentences and spoke spontaneously in each of six appliance and noise conditions. Several acoustic and perceptual measurements were made in each condition. In general, speech deteriorated when appliances were placed and when noise was presented. The type and amount of speech disruption varied as a function of speech task and aspect of speech. There was no evidence that the effects of appliances on speech differed in quiet and noise conditions. Inter-subject variability was large.

The effects of noise and palatal appliances on the speech of five-year-old children.

Forty children of age five were divided into two groups according to articulation skill. Each child named pictures while wearing a palatal dental appliance in quiet and in the presence of an intense noise. Articulation errors and vocal intensity were measured in all conditions. Errors increased when appliances were placed and when noise was presented, but there was no interaction between the effects of appliances and noise on articulation. Intensity increased when noise was presented and decreased when appliances were placed. There was no interaction between the effects of appliances and noise on intensity. Children in the High and Low Articulation Skills groups did not differ in response to appliances or noise. The articulation results are very similar to results described earlier in an adult group. The children's vocal intensity changes were almost twice as large as those reported for the adult group. There may be a complex interaction between speech measure, type of feedback manipulation, and stage of development.

Auditory brainstem anomalies in human albinos.

Brainstem auditory evoked potentials recorded from human albinos indicate significant hemispheric asymmetry. The asymmetry is symptomatic of differences between decussated and nondecussated auditory pathways in albino and pigmented humans at approximately the level of the superior olivary nuclei. Abnormal decussation of auditory pathways in albinos probably coincides with known visual system anomalies.

The effects of feedback filtering on speaker intelligibility.

Subjects read intelligibility tests while hearing their voices low- or high-pass filtered. The tests were presented to listeners to assess speaker intelligibility. Feedback filtering had a variable effect on intelligibility. When speakers were permitted to increase their intensities in low-pass filtering conditions, their intelligibility also increased. Intelligibility did not increase when intensity was controlled. These results do not support the hypothesis that speakers modulate their voices in order to maintain intelligible communication with a presumed listener. IT is suggested instead that speakers alter their intensities in order to regulate the loudness of their voices in their own ears.

The use of visual feedback to control vocal intensity and nasalization.

A device which provides visual feedback of vocal intensity, nasalization, or a ratio of nasalization to vocal intensity, is described. The device was used in several experiments in which control of vocal intensity was a necessary component. Normal and hypernasal speakers were able to control their vocal intensities successfully when using the visual display. The display also was used in an experiment designed to determine how well normal speakers can manipulate their nasalization without visual feedback and whether they improve when feedback is provided. Subjects manipulated their nasalization first without feedback, then with feedback, then, again, without feedback. The subjects' performance improved when feedback was provided and determined when feedback was removed. It appears that this device is useful for many applications in a variety of settings.

The effects of feedback filtering on nasalization in normal and hypernasal speakers.

Two groups of subjects, one with normal speech and one with hypernasal speech, spoke while hearing their voices unfiltered, low-pass filtered with cut-off frequencies of 1000, 500, and 300 Hz and high-pass filtered with cut-off frequencies of 500, 1000, and 2000 Hz. Measurements of nasalization were made with a miniature accelerometer attached to the side of the subject's nose. Both groups of subjects decreased nasalization when hearing their voices low-pass filtered with a cut-off frequency of 300 Hz. This decrease was statistically significant. The results were interpreted as supporting the hypothesis that nasalization is under feedback control.

Intensity of guitar playing as a function of auditory feedback.

Subjects played an electric guitar while auditory feedback was attenuated or amplified at seven sidetone levels varying 10-dB steps around a comfortable listening level. The sidetone signal was presented in quiet (experiment I) and several levels of white noise (experiment II). Subjects compensated for feedback changes, demonstrating a sidetone amplification as well as a Lombard effect. The similarity of these results to those found previously for speech suggests that guitar playing can be a useful analog for the function of auditory feedback in speech production. Unlike previous findings for speech, the sidetone-amplification effect was not potentiated by masking, consistent with a hypothesis that potentiation in speech is attributable to interference with bone conduction caused by the masking noise.