Loudness Variation Research Articles

The relationship between speech perception and electrode discrimination in cochlear implantees.

Speech Intelligibility Index (SII) procedures were used to measure the amount of speech information perceived in five frequency bands (170-570, 570-1170, 1170-1768, 1768-2680, and 2680-5744 Hz) by 15 users of the Cochlear Ltd. CI-22M implant and Spectra-22/SPEAK processor. The speech information perceived was compared to that perceived by normal-hearing listeners. The ability of these subjects to discriminate between stimulation on adjacent electrodes corresponding to each frequency band was also investigated, using a 4IFC procedure with random current level variations of between 0% and 60% of the dynamic range. Relative to normal-hearing listeners, speech information was, on average, significantly more reduced in the four frequency regions between 170 and 2680 Hz than in the region 2680-5744 Hz. There was a significant correlation between electrode discrimination ability (when the random level variation encompassed 20% or more of the dynamic range) and the amount of speech information perceived in the four frequency regions between 170 and 2680 Hz. There was no such correlation in the region 2680-5744 Hz, regardless of the extent of random level variation. These results indicate that speech information in the low to medium frequencies is more difficult for implantees to perceive, that this difficulty is correlated with the difficulty in discriminating electrode place in the presence of random loudness variations, and that fine spectral discrimination may be relatively more important in the vowel-formant regions than in higher frequency regions.

Voice characteristics in the progression of Parkinson's disease.

This study examined the acoustic and perceptual voice characteristics of patients with Parkinson's disease according to disease severity. The perceptual and acoustic voice characteristics of 30 patients with early stage PD and 30 patients with later stage PD were compared with data from 30 normal control subjects. Voice recordings consisted of prolongation of the vowel /a/, scale singing, and a 1-min monologue. In comparison with controls and previously published normative data, both early and later stage PD patients' voices were characterized perceptually by limited pitch and loudness variability, breathiness, harshness and reduced loudness. High modal pitch levels also characterized the voices of males in both early and later stages of PD. Acoustically, the voices of both groups of PD patients demonstrated lower mean intensity levels and reduced maximum phonational frequency ranges in comparison with normative data. Although less clear, the present data also suggested that the PD patients' voices were characterized by excess jitter, a high-speaking fundamental frequency for males and a reduced fundamental frequency variability for females. While several of these voice features did not appear to deteriorate with disease progression (i.e. harshness, high modal pitch and speaking fundamental frequency in males, fundamental frequency variability in females, low intensity and jitter), breathiness, monopitch and monoloudness, low loudness and reduced maximum phonational frequency range were all worse in the later stages of PD. Tremor was the sole voice feature which was associated only with later stage PD.

Constancy of loudness of pipe organ sounds at different locations in an auditorium

Pipe organ sounds, as judged by ear, tend to remain constant across different locations in an auditorium, yet the SPL of line spectra may vary by a maximum of 26 dB (mean 8.98 dB, s.d. 2.5), and the overall level may vary, typically, 10 to 12 dB from location to location. However, organs are designed, listened to, and regulated using the psychophysical units of loudness and timbre, and it is possible that the heard sound constancy exists at the psychophysical level. The present work recorded the sound of the C's and G's of pipe organ stops at three different locations in a church. The sound pressure levels were transformed to loudness. Similarity of loudness across the locations was measured two ways. First, the bass to treble distribution of loudness across the compass was measured using cross-correlation functions across pairs of locations. Second, the degree of similarity of loudness at the different locations was quantified by calculating ratios of loudness between the different locations. By these measures, the bass to treble loudness distribution and absolute loudness of the reeds were found to be nearly identical at the three locations. Two psychophysical processes were shown to be responsible for the loudness constancy. The first depended upon the power summation of harmonics within each third octave band above band 9 which contain two or more harmonics. The power summation of these harmonics greatly reduced the effect of SPL variability of the line spectra contained within these higher numbered bands. The second depended upon interharmonic loudness summation and upward masking of the first six harmonics. Greater loudness variability at the different locations was found after transforming the SPL measurements of two 8-ft diapasons to loudness compared with the reeds. The larger loudness variability was due to the smaller number of harmonics above the third of the diapasons compared with the reeds. The psychoacoustic measures indicate what a listener will hear as he/she moves among the locations.

Lexical boundary error analysis in hypokinetic and ataxic dysarthria.

This investigation is the second in a series to examine a potential source of reduced intelligibility in dysarthric speech, namely the mismatch between listeners' perceptual strategies and the acoustic information available in the dysarthric speech signal. Lexical boundary error (LBE) analysis was conducted on listener transcripts from phrases produced by speakers with hypokinetic dysarthria, ataxic dysarthria, and normal controls. By design, the hypokinetic and ataxic dysarthric tapes elicited similar intelligibility (words-correct) scores. However, they elicited different numbers and patterns of lexical boundary errors. The nature of the error pattern differences can be traced to the listeners' use of available syllabic strength information to segment the acoustic stream. Specifically, although both dysarthric speech samples elicited numerous lexical boundary errors, those for the hypokinetic speech generally conformed to predictions offered from studies of degraded normal speech. Those for the ataxic speech did not conform strongly to such predictions. It appears that the prosodic deficits of the ataxic speech (tendency toward syllabic isochrony, excessive loudness variation, and reduced vowel working space consequent to reductions in vowel strength) posed more of a problem for listeners than did the prosodic deficits of the hypokinetic speech (rapid rate, monotony, reduced vowel working space).

Optimizing sound localization with hearing AIDS.

Optimizing sound localization with hearing AIDS.

Multidimensional scaling of synthetic musical timbre: perception of spectral and temporal characteristics.

The perceptual correlates of acoustic parameters involved in musical timbre were investigated by examining judgements of timbre dissimilarity. Nine synthetic sounds were created, derived from crossing three levels of spectral and temporal parameters (number of harmonics and rise time, respectively). Two separate conditions were tested, one using single tones, the other using short melodies. Fifteen musically untrained subjects were presented with pairs of stimuli and asked to judge dissimilarity on an 8-point scale. The spatial configuration resulting from multidimensional analysis of the data was best fit by a three-dimensional model, with the first two dimensions accounting for most of the variance. The perceptual space derived from the analysis indicates that these two orthogonal dimensions corresponded closely to the spectral and temporal differences inherent to the stimuli. Similar results were obtained with both melodies and single tones. A second experiment replicated the findings despite the introduction of random loudness variation, indicating that timbre judgements can be made independently of loudness. We conclude that even musically unselected subjects are sensitive to spectral and temporal information in musical tones, and are able to use them independently in making perceptual judgements of musical timbre.

Glottal airflow voice source characteristics of 11-year-old singers

The voice source of 18 girls and 10 boys from a music school was analyzed by inverse filtering the flow signal during sustained phonation of the vowel /a/ at three pitches (F0 between 220 and 330 Hz) using soft, mid, and loud voices. Subglottal pressure was measured as the oral pressure during the /p/ occlusion preceding the vowel. As compared to previous data on adult voices, the pressure values were high, and boys tended to use slightly higher pressures than girls. The influence of pitch and loudness variation was similar to that previously reported for adult voices. The ac flow amplitude was substantially smaller than previously reported for adults and was significantly higher in the boys than in the girls, while no significant sex differences were found for the children’s closed quotient and for the level difference between the two lowest spectrum partials. Contrary to findings for adults, the dc flow during the quasiclosed phase was greater in the boys. Overall, the present findings were similar to those previously reported for untrained children’s voices. Thus children’s voice source characteristics can be more conveniently derived from singer subjects with their greater skill in independent control of pitch and loudness.

Midsagittal velar kinematics during production of VCV sequences.

We present data on velar speech movements examined by electromagnetic articulography (EMA), a tool for the study of velar kinematics. Based on investigations on 22 normal German subjects (15 male, 7 female), movement paths of the velum using a sensor positioned on the midsagittal anterior surface were recorded using a series of high and low vowel VCV utterances. Maximum velar elevation (mean = 13.5 +/- 0.6 mm [SD]) with the rest position as a reference was observed during the production /iti/. The average maximum articulatory velocity of velar movements during /iti/ production was 141.0 mm/sec. Maximum movement of the velum observed during production of /ana/ averaged 7.5 +/- 1.1 mm. The direction of velar movement ranged between 44.6 degrees and 50.4 degrees relative to the bite plane registered electromagnetically. The influence of loudness variation of sequence production is demonstrated and discussed with the literature.

Loudness growth in forward masking: Relation to intensity discrimination

The growth of loudness of a tone burst following an intense forward masker was measured as a function of the tone level. The level of the forward-masked tone was adjusted to balance the loudness of a standard tone presented without a forward masker, using a 2AFC, double-staircase, tracking procedure. The forward masker was a 90-dB SPL, 100-ms, 1000-Hz pure tone. The standard tone and the masked tone were both 25-ms, 1000-Hz pure tones. The forward masker and the masked tone were always presented in the first interval. With a 100-ms delay between them, there was little or no threshold elevation for the masked tone. However, the masker caused the masked tone to sound louder than it would if it had not been masked, a phenomenon termed "loudness enhancement" [Irwin and Zwislocki, Percept. Psychophys. 10, 189-192 (1971); Galambos et al., J. Acoust. Soc. Am. 52, 1127-1130 (1972)]. In addition, the present results show a nonmonotonic enhancement function that the forward masker introduced a 10-16-dB enhancement effect for tones of 40-65 dB SPL and no significant effect for the 30 and 90 dB SPL tones. The loudness variability in forward masking was estimated from the upper and lower sequences tracking the 21% and the 79% louder response levels on the psychometric function, respectively. The variability demonstrated a similar nonmonotonic function. In forward masking loudness grows more steeply at low-medium sensation levels, and merges with normal growth at high levels.(ABSTRACT TRUNCATED AT 250 WORDS)

Time-domain simulation of sound production in brass instruments

Sound production in brass instruments has been formulated physically, including two different models of lip vibration; the perpendicular model, where lips strike laterally to the direction of the air flow, and the ‘‘swinging door’’ model, where lips execute outwardly rolling motion. The above formulation was used to carry out time-domain simulation. As a result, self-excitation of the brass instrument was found to be possible with both of these lip models. By changing the lip eigenfrequency, sustained oscillations near at least the first through the eighth resonant impedance peaks of the instrument were realized. The simulated sound was shown to have variety in the harmonic structure accompanied with pitch and loudness variations as is seen in sound the natural brass instruments produce. Unusual regimes of oscillation were also obtained; the pedal tone, whose fundamental does not participate in its sound production, a tone octave below the third impedance peak, whose odd harmonics do not contribute to the regime of oscillation, and a multiphonic, which is an oscillation perceived having two fundamentals simultaneously.

Phonatory instability in ALS dysarthria: A case study

A subgroup of patients with dysarthria resulting from amyotrophic lateral sclerosis (ALS) presents with voice problems characterized primarily by cycle‐to‐cycle and long‐term phonatory instabilities. This paper reports acoustic analyses of the phonation of a 63‐yr‐old female 9 months post diagnosis of ALS. Perceptual characteristics of voice included inconsistent harshness and cyclic variations in both pitch and loudness. Spectral analysis of fundamental frequency (f0) and intensity contours revealed periodic instabilities at multiple temporal domains, including 0.4‐ and 0.8‐ Hz cycles in both f0 and intensity (‘‘wow’’), 2‐Hz cycles in f0 (‘‘tremor’’), and 11‐Hz cycles in intensity (‘‘flutter’’). Phase relations were noted among these cycles and with other acoustic analysis parameters such as perturbation and spectral energy distribution, indicating that occurrence of short‐term and cycle‐to‐cycle instabilities was modulated by the longer term cycles. The results are discussed with relation to video stroboscopic examinations of this patient, and also in comparison with phonatory characteristics of other patients in this subgroup (females with ALS related voice problems). [Work supported by NIH.]

Speaking behavior and voice sound characteristics in depressive patients during recovery

Based on a sample of 30 depressive patients, we have investigated the time course of recovery from depression in so far as this time course was assessable through changes in psychopathology syndrome scores and through changes in speaking behavior and voice sound characteristics. Specifically, our study design provided 6 repeated assessments over 2 weeks and at a fixed time in the morning each Monday, Wednesday and Friday, plus a final assesment at the patients' releases from hospital. Thus, we were able to determine the degree to which single- parameter approaches to speaking behavior and voice sound characteristics reflect the individual time course of recovery from depression. In this context, we could rely upon a calibration sample with repeated assessments on 192 healthy volunteers which yielded all necessary information concerning reproductibility and sensitivity of speech parameters. Our analysis revealed several prominent features of speaking behavior and voice sound characteristics to be closely related to the time course of recovery from depression. In particular, the parameters “F0-amplitude”, “F0-6db-bandwidth” and “F0-contour” which assess important characteristics of a speaker's voice timbre, as well as the parameters “energy” and “dynamics” which assess a speaker's mean loudness and the variation of loudness over time, displayed consistently high correlations with depression syndromes. Moreover, the results of single-case analysis turned out to be in remarkable accordance with those of the cross-sectional one: in almost two-thirds of patients there existed a significant relationship over time between the global depression scores and major speech parameters. As to the remaining one-third of patients who did not fit the picture of high correlations between psychopathology and speech parameters, we found an overproportionally large number of non-improvers characterized by irregular patterns of slight improvement with subsequent deterioration, or of deterioration followed by slight improvement. In other words, one-third of patients displayed time courses of depression whose psychopathology is difficult to assess through standard exploration techniques. Accordingly, it is not clear whether the observed lack of correlation in these patients is due to insufficient data or to an actual disordance between the time development of psychopathology and that of speech parameters.

Phonatory control in male singing: A study of the effects of subglottal pressure, fundamental frequency, and mode of phonation on the voice source

This article describes experiments carried out in order to gain a deeper understanding of the mechanisms underlying variation of vocal loudness in singers. Ten singers, two of whom are famous professional opera tenor soloists, phonated at different pitches and different loudnesses. Their voice source characteristics were analyzed by inverse filtering the oral airflow signal. It was found that the main physiological variable underlying loudness variation is subglottal pressure (Ps). The voice source property determining most of the loudness variation is the amplitude of the negative peak of the differentiated flow signal, as predicted by previous research. Increases in this amplitude are achieved by (a) increasing the pulse amplitude of the flow waveform; (b) moving the moment of vocal fold contact earlier in time, closer to the center of the pulse; and (c) skewing the pulses. The last mentioned alternative seems dependent on both Ps and the ratio between the fundamental frequency and the first formant. On the average, the singers doubled Ps when they increased fundamental frequency by one octave, and a doubling of the excess Ps over threshold caused the sound pressure level (SPL) to increase by 8-9 dB for neutral phonation, less if mode of phonation was changed to pressed. A shift of mode of phonation from flow over neutral to pressed was associated with a reduction of the peak glottal permittance i.e., the ratio between peak transglottal airflow to Ps. Flow phonation had the most favorable relationship between Ps and SPL.

Instrument modeling using sampled sound

The modern electronic keyboard instrument is asked to imitate all the familiar musical instruments. Virtually all rely on digital sampling. Compared to older synthesis techniques, sampling at first seemed to present startling realism. But the exact reproduction of a single recording of a single note does not fool the ear. Every musical instrument has a set of cues which, when absent, expose the imposter. This talk examines a number of techniques that are used to create more realistic simulations of acoustic instruments in Ensoniq’s latest sampling keyboards. The problems addressed include temporal variations in loudness, pitch, and brightness; formants; legato connections between notes; sympathetic resonances, restrike of a vibrating string, and nonlinearities. Expressive nuance in performance poses additional questions. How can a performer, squeezed through the pipe of digital information, create expression in a sampled sound? The answer must include ergonomic physical controllers linked to perceptually important sound parameters. Since samples are inherently static, special tricks are needed to give the illusion of a dynamic process.

Loudness and the binaural masking level difference

The large threshold difference found between a low-frequency antiphasic signal, Sπ, and in-phase signal, SO, in the presence of a homophasic noise masker (NO) is referred to as the binaural masking level difference (BMLD). The magnitude of the BMLD diminishes when the antiphasic signal is increased to suprathreshold levels. The decrease in effectiveness of a 180° phase-shifted signal, at suprathreshold levels, produces a concomitant decrease in perceived loudness of the signal. This decrease in effectiveness at the higher signal levels was examined in a loudness-matching paradigm. A change in neural processing was hypothesized to account for the variation in loudness, detectability, discrimination, and identification when the signal level approached 20 dB SL.

Effect of loudness and timing variation on rhythm perception

A rhythmically ambiguous musical passage, interpretable as being in either duple or triple time, was constructed. The loudness and timing of notes in the passage were varied in order to systematically explore the space of expressive variation of these parameters that would be used by a performer to induce a particular rhythmic interpretation. The resulting stimuli were played to musically trained experimental subjects, who were then asked to make a forced choice between duple and triple rhythmic interpretations. The subjects' responses were statistically analyzed and compared with a hypothetical model of their interpretive processes. It was found that, while loudness variation affected subjects' responses in a manner consistent with previous results and with the hypothetical model, timing variation, surprisingly, has no obvious effect on the subjects' responses. These results have implications for machine interpretation of rhythm, as well as for an understanding of human rhythm processing.

Perceiving fractal noises in pitch, loudness, and duration

Three experiments examined discrimination among a family of fractals possessing different power law spectra. In experiment 1, random numbers sequences were generated so that their spectra had slopes (in log‐log plane) or 0 (white noise), − 1 (flicker noise), and − 2 (brown noise). The distributions were used to produce sequences that could vary in pitch, loudness, or duration. Listeners categorized each sequence as being of one of the three noise types. Thurstonian scaling revealed that noise segment classification could be accomplished on the basis of pitch and loudness variation, but not on the basis of duration changes. Experiment 2 explored simultaneous pitch and loudness variation, finding dimensional redundancy for white/flicker discrimination suggestive of a pitch‐loudness interaction. Experiment 3 examined sensitivity for discrimination of noises as a function of the power law exponent. Listeners heard two sequences of tones and indicated whether they had the same or different power law exponent. The range of exponents spanned the interval [0, −3], with the stimuli under comparison differing by 0.2, 0.4, 0.6, or 0.8. ROC analyses revealed some variation in sensitivity to the fractal exponents as a function of position in the [0, −3] interval.

Acoustic studies of the voice source in male singers

This paper reviews various recent investigations with various co‐workers of male singers' vocal fold function mainly studied by inverse filtering for deriving the waveform of the transglottal airflow during phonation. Theoretical and experimental studies confirmed the dominating acoustic importance of (1) the peak airflow Ûg for the amplitude of the voice source fundamental, and (2) the negative peak amplitude of the differentiated waveform dU/dt for the SPL of the radiated sound. Unlike nonsingers, some professional baritone and bass singers have been found to retain a high Ûg during loud tones at high pitches. Different strategies are possible for increasing dUg/dt for loudness variation: increasing Ug, increasing the duration of the closed phase, and skewing the pulse shape. Different singers use different combinations of these alternatives depending on pitch, loudness range, and voice type. Some of the physiological correlates of these strategies can be postulated. Here, Ûg seems highly dependent on the degree of glottal adduction and dUg/dt on subglottal pressure. However, influence of the breathing strategy has also been quantitatively demonstrated.

Psychophysical studies evaluating the feasibility of a speech processing strategy for a multiple-channel cochlear implant.

This paper reports further psychophysical studies on a multiple-channel cochlear implant patient evaluating the feasibility of a speech processing strategy which converts the acoustic fundamental frequency to electric repetition rate, the second-formant frequency to electrode position, and the acoustic amplitude to current level. The first four studies evaluated the use of a special pulse pattern to minimize the loudness variation with electric repetition rate. The chosen pulse pattern consisted of multiple pulses occurring in the first half of each repetitive period (MPP) in contrast to the more conventional pattern with a single pulse per period (SPP). The results showed that MPP approximately equalized the loudness variation with repetition rate. The dynamic range of current, the pitch variation with repetition rate, and the difference limens for repetition rate were found to be similar to MPP and SPP. Two other studies investigated interaction between electrode position and repetition rate (RR). The first of these showed that the patient could make use of information provided by rising or falling RR trajectories superimposed on individual electrodes or electrode trajectories as an indicator of the direction of intonation variation. The second of these studies showed that the dissimilarities amongst the hearing sensations produced by steady-state stimuli differing in electrode position and repetition rate were characterized by two perceptual components, relating to the two electric parameters, respectively.

Turn-competitive incomings

Abstract A recurrent feature of multi-party conversation is that one speaker comes in prior to the completion of another's turn and can be heard as directly competing with the other for possession of the turn. That is, the incomer can be heard as wanting the floor to himself not when the current speaker has finished but now, at this point in the conversation. Our analysis reveals that in managing talk of this kind participants, methodically produce and monitor for certain prosodic features of speech. These features, which have hitherto received scant attention in analyses of interaction, involve pitch-height, tempo and loudness variations. By deploying these prosodic features participants can constitute their incomings as competitive for the turn irrespective of the lexico-syntactic or illocutionary characteristics of their talk.