Presence Of Masking Noise Research Articles

マコガレイの背景雑音によるマスキング

Using a classical cardiac conditioning technique, the auditory thresholds of marbled sole Pleuronectes yokohamae were measured in the presence of masking noise in the spectrum level range of 70-90dB re 1μPa/√Hz. Critical ratios (signal-to-noise ratio) were approximately 27dB at 60-100Hz and 32-58dB at 150-700Hz. The ratio increased almost linearly with increasing frequency. The noise spectrum level at the start of masking was 62-70dB within the frequency range of 60-700Hz. This suggests that hearing of the marbled sole is masked in the natural environment with the noise spectrum level above 70dB.

Age-related loss of activity of auditory-nerve fibers.

1. Characteristic frequencies (CF), spontaneous rates (SR), and thresholds were recorded from single fibers in the auditory nerves of gerbils aged for 36 mo in a quiet vivarium. The data from the quiet-aged animals were compared with similar data obtained previously from young controls. Fibers were classified as "low-SR" if their spontaneous rates were < or = 18 spikes/s and "high SR" for higher rates. 2. For CFs > 6 kHz, the percentage of low-SR fibers contacted declined from 57% of the population in young gerbils to 29% in the aged gerbils. This population change is statistically significant (P < 0.01). At CFs < 6 kHz, the population demographics did not change significantly with age, with the low-SR fibers comprising 30 and 39% of the population, respectively, for the young and aged animals. 3. To further test the hypothesis that low-SR fibers with CFs > 6 kHz become less active with age, additional experiments were conducted to examine the recovery of the compound action potential (CAP) response from prior high-level stimuli. Previous work has shown that the CAP recovery curve has two segments: a fast segment associated with the high-SR fibers and a slow segment associated with the low-SR fibers. The curves obtained from quiet aged gerbils show a faster recovery than young controls for probe tones at 8 and 16 kHz, but not at 2 and 4 kHz. Thus these results agree with our single-fiber data indicating that there is a loss of low-SR activity for CFs > 6 kHz in the aged animals. 4. Low-SR fibers typically have larger dynamic ranges than those of high-SR fibers, are better able to preserve information concerning stimulus timing and amplitude modulation, and their responses are more robust in the presence of masking noise. Moreover, low-SR fibers are likely inputs to the crossed-olivocochlear reflex, a reflex that serves an antimasking role in the detection of sounds in a binaural noise field. If true for humans, the loss of the low-SR system could explain many of the hearing deficits often seen in older individuals; e.g., decreased ability to understand speech in noise, changes in masking level differences, and decreased ability to localize sound sources using binaural cues.

スケトウダラの音響生態学的研究 Ｖ． スケトウダラの聴覚しきい値と雑音スペクトラムレベル比

Using a cardiac conditioning technique, auditory thresholds in walleye pollock Theragra chalcogramma were determined in the presence of masking noise. Signal frequencies of 60, 120, 200, 300, 400, and 500Hz were used. Unfiltered random white noise was introduced as constant background noise. Thresholds ranged from 17 to 66dB above the noise spectrum level in the region of the signal frequency. At the most sensitive frequency (the 60Hz signal), critical ratio (signal/noise ratio) was 17dB, and tends to be larger along with higher frequencies that start lowering the audible sense. It was possible to detect signals in the frequency range of 60 to 500Hz in the presence of background noise. These findings are similar to results from the previous study conducted on cod with similar hearing ability and critical ratios.

DETECTION OF SPECIES-SPECIFIC CALLS IN NOISE BY ZEBRA FINCHESPOEPHILA GUTTATAAND BUDGERIGARSMELOPSITTACUS UNDULATUS:TIME OF FREQUENCY DOMAIN?

ABSTRACT Two zebra finches and two budgerigars were trained, by operant conditioning, to detect autogenous (self-generated) distance calls in the presence of masking noise. For both species, there were no differences in detection thresholds for normal calls compared to time- reversed calls. Thresholds for autogenous calls were also compared with thresholds of the other species of birds listening to the same call. When detecting a zebra finch call, budgerigars had slightly lower thresholds than that of the zebra finch. On the other hand, when detecting a budgerigar call, zebra finches showed significantly higher thresholds than the budgerigar. From these results, and from what is known about basic hearing capabilities in these species, we conclude that these birds are not using a mechanism which utilizes “matched” or cross-correlational filtering. It is more likely that they are using “frequency-based” filtering in detecting calls in noise.

Development of a speech intelligibility test for hearing aid research

An English language speech comprehension test has been developed to address shortcomings of current speech tests as measures of hearing handicap [R. Plomp, J. Acoust. Soc. Am. 63, 533–549 (1979)]. The Bamford‐Kowal‐Bench sentences, as revised to conform to American English usage, were modified to create 28 lists of 12 sentences of approximately equivalent length, difficulty, and phonetic content. These materials were then sampled at 20 kHz with 8‐kHz bandwidth for measurement, editing, and analysis. Following procedures used with Dutch materials [R. Plomp and A. M. Mimpen, Audiology 18, 43–52 (1979)], Leq for each sentence was matched as closely as possible. Speech reception thresholds were measured in the presence of masking noise that matched the average long‐term spectrum of all sentences as a means of norming the difficulty of each sentence list. Results of the norming tests will be presented and the use of these materials in hearing aid research will be discussed.

Word recognition by humans and machines: Tests on a multitalker, multistyle database

Experiments comparing isolated word recognition by human listners with automatic speech recognition systems are valuable because error analyses may lead to improvements in speech recognition technology. Isolated word recognition in adult human listeners has been compared with recognition performance by two commercially available speech-recognition systems. The test stimuli were drawn from the Lincoln Laboratory Stressed-Speech database. The database consists of 6930 stimuli (two iterations of each of 35 words spoken by nine different people in 11 different speaking styles). The vocabulary contains confusable words (i.e., go, hello, oh, no, and zero); the speaking styles include a wide range of naturally occurring variations (i.e., normal, slow, fast, soft, loud, angry). Analyses show that the acoustic characteristics of individual words vary considerably across talkers, and across styles within talkers. Performance of human listeners and the two machine-based recognition systems was tested in a single-talker, multistyle condition, and in a multitalker, multistyle condition. All tests were conducted under two listening conditions: normal, and in the presence of masking noise. The data to be presented are the error patterns of human listeners, versus the machine-recognition systems, exhibited across talkers, across speaking styles, and across training conditions (multitalker, multistyle training versus single talker, single style training). [Work supported by Boeing Aerospace and Electronics.]

Masked cochlear whole-nerve response intensity functions altered by electrical stimulation of the crossed olivocochlear bundle.

Cochlear whole-nerve response (CAP) intensity functions were recorded from the guinea pig round window. The intensity functions were obtained in the presence of masking noise, with electrical stimulation of the crossed olivocochlear bundle (COCB), and the combination of masking noise and COCB stimulation. Electrical stimulation of the COCB produced the expected reduction of CAP magnitude for low- to moderate-intensity tone bursts. Masking noise produced reductions in CAP magnitude over the whole signal intensity range used in this study. The combination of electrical stimulation of the COCB with the masking noise produced CAP magnitude changes graded between reduction and enhancement dependent on signal and masker levels. In general, at low signal levels, the masked CAP magnitude is reduced compared to the masked alone condition. At high signal levels, the masked CAP is increased in magnitude. These results confirm and extend the earlier observations of Nieder and Nieder [Exp. Neurol. 28, 174-188 (1970); also, Nature 227, 184-185 (1970)].

Detection of complex echoes in noise by an echolocating dolphin.

Dolphins echolocate with short broadband acoustic signals that have good time resolution properties. Received echoes are often complex, with many resolvable highlights or components caused by reflection of the incident signal from external and internal boundaries of a target and from different propagational modes within a target. A series of experiments was performed to investigate how dolphins perceive complex echoes. Echoes were produced by a microprocessor-controlled electronic target simulator that captured each emitted click and retransmitted the signal back to the animal after an appropriate time delay. The use of this "phantom" target allowed for precise control of the number of highlights, the time separation between highlights, and the relative amplitudes of highlights in the simulated echoes. An echolocating dolphin was trained to perform a target detection task in the presence of masking noise using these phantom echoes. The properties of simulated echoes were systematically varied, and corresponding shifts in the dolphin's detection threshold were observed, allowing for inferences of how the dolphin perceived echoes. The dolphin performed like an energy detector with an integration time of approximately 264 microseconds.

Comparison of target detection capabilities of the beluga and bottlenose dolphin.

The echolocation detection capabilities of a beluga (Delphinapterus leucas) and an Atlantic bottlenose dolphin (Tursiops truncatus) were directly compared in a target detection experiment. Both animals were trained to detect targets in the presence of masking noise. Targets were stainless-steel, water-filled spheres 7.62 and 22.86 cm in diameter. Target ranges of 16.5 and 40 m were used with the 7.62-cm sphere and 80 m with the 22.86-cm sphere. Masking noise with a flat spectrum from 40-160 kHz was projected from a spherical transducer placed 4 or 5 m, depending on the target distance, from the animal hoop station in line with the target. Target detection performance was determined as a function of masking noise level at each target range. The echo-to-noise ratio (Ee/No)max for the beluga at the 75% correct response threshold was approximately 1.0 dB compared to about 10 dB for the dolphin. The differences of each animal's detection performance across the three ranges were consistent with target strength and transmission loss differences. It is speculated that the difference in performance between the two species may be due to differences in critical bandwidth, signal processing capability, or echolocation strategy.

The preception of complex echoes by an echolocating dolphin

Dolphins echolocate with short broadband acoustic signals which have good time resolution properties. Received echoes are often complex, with many resolvable highlights or components caused by reflection of the incident signal from external and internal boundaries of a target and from different propagational modes within a target. A series of experiments was performed to investigate how dolphins perceive complex echoes. Echoes were produced by a microprocessor‐controlled electronic target simulator which captured each emitted click and retransmitted the signal back to the animal after an appropriate delay time. The use of a “phantom” target allowed for precise control of the number of highlights, the time separation between highlights and the relative amplitudes of highlights in the simulated echoes. An echolocating dolphin was trained to perform a target detection task in the presence of masking noise using these phantom echoes. The properties of simulated echoes were systematically varied, and corresponding shifts in the dolphin's detection threshold were observed, allowing the inference of how the dolphin perceived the echoes. The dolphin performed like an energy detector with an integration time of approximately 290 μs.

Acoustic‐phonetic properties of vowels produced in noise

It is well known that speech produced in noise has higher intensity, longer duration, and higher fundamental frequency than speech produced in quiet. However, recent studies in our laboratory have indicated that changes in the acoustic properties of speech produced in noise are not limited to changes in the glottal source but may also include significant changes in the acoustic‐phonetic structure of speech reflected in the characteristics of vocal tract transfer function. We investigated the effects of wideband noise on the acoustic properties of vowels. Talkers produced ten American‐English vowels in /hvd/ context. Ten tokens of each vowel were produced in quiet and in the presence of masking noise (90 dB SPL). The results indicate a rather complex pattern of changes in formant frequencies. First, with the exception of the vowels /æ̈ɑ, ɔ/, whenever a significant change in F1 was observed. F1 frequency was higher in the noise condition. Second, F2 frequencies decreased in front vowels and increased in back vowels. The net results of these changes is a more compact vowel space.

Temporal summation in the auditory system of the goldfish

Previous investigations of temporal integration in fishes have presented conflicting results. Popper [J. Acoust. Soc. Am. 52, 596–602 (1972)] found no evidence for integration for signals between 10 and 500 ms in duration, while Offutt [J. Acoust. Soc. Am. 41, 13–19 (1967)] clearly did. Since one of the differences between these studies was the level of a noise background, we determined to measure the effect of noise on the detectability of tone pulses as a function of pulse duration. Three goldfish (Carassius auratus) were classically conditioned to detect tone pulses ranging from 2.5 to 100 ms in duration for repetition periods of 100 and 1000 msec under both ambient (unmasked) and broadband noise conditions. The threshold function of signal duration under noise showed nearly perfect integration (3 dB decrease in threshold per doubling of duration). The function under ambient conditions showed no consistent effects of signal duration, and thus no evidence for the type of integration expected from simple models [R. Plomp, J. Acoust. Soc. Am. 33, 1561–1569 (1961)]. We conclude that the presence of masking noise may account for the previous conflicting results, and that the nature of the neural processes used in signal detection by the goldfish are fundamentally different under noisy and quiet conditions. [Work supported by N.I.H. grants to R.R.F.]

Intelligibility of spoken letters of the alphabet

A subjective experiment was carried out to assess the effect of spelling context on the ability of listeners to identify spoken letters of the alphabet. Individual letters, spoken by a male talker and a female talker, were concatenated to form strings of up to ten letters representing the spelling of a last name (surname) and one or two initials. Three spelling contexts were observed. First, real names were selected at random from the 20 000‐entry Bell Labs telephone directory. Second, strings were constructed using trigram statistics for letter sequences tabulated from the same directory. Third, letters were selected at random to form strings. In addition, three masking noise conditions were observed: no noise, low‐level, and high‐level broadband additive masking noise. The results show that the listeners made fewer errors identifying letters imbedded in real name or trigram strings than in random letter strings. The effect is especially pronounced in the presence of masking noise. Confusions among the spelled letters are also tabulated.

Stimulus dominance in dichotic listening

Twenty-four listeners received 20 dichotic-listening trials of natural stop-vowel syllables aligned by reference to onset of noise burst. Listeners marked two responses for each of the 30 pairs of syllables per trial, which provided 480 pairs of responses for each pair of syllables. The direction of ear advantage varied considerably across the 30 pairs; 11 had a significant REA, 10 had a significant LEA, and a nonsignificant ear advantage occurred for nine pairs. These differences in direction arose mainly from stimulus-dominance effects in which the advantage belonged to one member of a pair independent of which ear received the syllable. Stimulus dominance will be discussed from two perspectives. The first is feature dominance; voiceless stops, on the average, produced higher scores than voiced stops in voiced-voiceless pairs. The second is syllable dominance; performance was sufficiently low for two syllables, /ba/ and /da/, to account for a substantial amount of the observed stimulus dominance. These two alternatives will be interpreted with respect to binaural intelligibility of the syllables in the presence of masking noise. [Supported by PHS Grant No. NS-12125.]

Auditory discrimination among speakers in the presence of masking noise

Subjects judged whether two masked utterances were spoken by the same or different speakers in an experiment designed to measure the influence of masking on speaker discrimination. Performance decreased from 80 percent correct judgments for a signal-to-noise ratio of +20 dB to 65 percent correct for a signal-to-noise ratio of -10 dB. Multidimensional scaling was used to determine speakers most sensitive to masking noise.

Spectral and aerodynamic properties of normal vs defective /s/

Temporary anterior open-bite malocclusion was experimentally induced in 15 normal speakers by introducing posterior bite blocks. Speech samples produced at 0, 1, 3, and 6 mm incisal spacing, both with and without peripheral masking, were analyzed for spectral and aerodynamic features. Selected /s/ sounds were judged for acceptability by naive and phonetically trained listeners. Preliminary analysis indicates (1) judged dysfluency increased both with increasing anterior spacing and in the presence of masking noise; (2) the presence of masking noise had little effect on oral constriction; (3) the relationship between anterior spacing and degree of constriction was complex, speakers with wider than average constriction widening further with increasing spacing, those with normally tight constriction narrowing still further; (4) dysfluent /s/ spectra tended to have lower amplitude and frequency than normal productions by the same speaker. These data indicate differences among speakers in adaptive response to anterior spacing. [Supported by NIH Grant No. AF446, NIDR Grant No. DE03533, and SRS Grant No. 44P20078.]

Most Comfortable Loudness for Pure Tones and Speech in the Presence of Masking Noise

Pure-tone and speech most-comfortable-loudness (MCL) levels were determined in noise and quiet. A Bekesy-type tracking task was used to determine the MCLs under all conditions. The subjects (N = 12) tracked their MCLs for five frequencies (250, 500, 1000, 20008, and 4000 Hz) in the presence of four white-noise levels (55, 65, 75, and 85 dB SPL) and in quiet. Speech MCLs were tracked under the identical noise conditions. The results indicated that the mean MCLs for tones between 500 and 4000 Hz, and for the speech stimuli, systematically increased as a function of the masking level. The rate of this increase was 7 dB per 10-dB noise increment for both stimulus modes when the level of the noise exceeded 65 dB SPL. At levels below 65 dB SPL, however, pure-tone MCL increases were only about 3 dB per 10-dB noise increment, while the slope of the speech MCL function remained relatively unchanged. The variability of MCL judgments for both stimulus modes decreased as a function of increased masking.

Most comfortable loudness for pure tones sad speech in the presence of masking noise

Pure-tone and speech most-comfortable-loudness levels (MCLs) were determined in noise and quiet. A Békésy-type tracking task was used to determine the MCLs under all conditions. During the first phase, the subjects (N = 12) tracked their MCLs for five pure-tone frequencies (250, 500, 1000, 2000, and 4000 Hz) in the presence of four white-noise levels (55, 65, 75, and 85 dB SPL) and in quiet. During the second phase, speech MCLs were tracked under the noise conditions identical to those of the tones. The results indicated that the mean MCLs for tones between 500–4000 Hz, and for the speech stimuli, systematically increased as a function of the masker level. The rate of this increase was about 7 dB/10-dB noise increment for both stimulus modes when the level of the noise exceeded 65 dB. At noise levels below 65 dB SPL, pure-tone MCL increases were only about 3 dB/10-dB noise increment, while the slope of the speech MCL function remained relatively unchanged. The results further indicated that the variability of MCL judgments for both stimulus modes decreased as a function of increased masking. The present MCL data were related to other investigations of loudness growth in noise.

Auditory discrimination among speakers in the presence of masking noise

The influence of masking noise on the ability of listeners to distinguish among speakers was measured for noise-to-signal ratios of − 20 dB to + 10 dB. The prerecorded stimulus ensemble contained 15 utterances: three test sentences, each spoken by five speakers. Each trial consisted of two different test sentences, spoken by the same or different speakers with equal probability. Masker level was the same for the two utterances in a trial. The subjects' task was to judge whether the two utterances were spoken by the same or by different speakers. Five listeners unfamiliar with the speakers took part in the experiment. Percentage of correct responses decreased from 80% for a noise-to-signal ratio of − 20 dB to 65% for a noise-to-signal ratio of + 10 dB (chance performance = 50%), while response time increased monotonically. Multidimensional scaling techniques were used to determine those speakers for whom discrimination performance was most sensitive to masking noise. Continuing experiments with other signal degradations are outlined.

Masking Of Action Potentials In The Guinea Pig Cochlea, Its Relation To Adaptation

The amplitude, AN1, and the latency, τN1, of the N1 component of round-window-recorded APs in the guinea pig are modified by simultaneous background noise and also for some time afterward by a preceding noise burst ('forward masking') White-noise masking acts differently below and above 60 dB with respect to its influence on AN1 and τN1.Forward masking is more effective if the delay time is short. On the basis of different slopes of the curves relating AN1 to noise intensity, the time during which forward masking has an effect can be subdivided into three periods. The period of shor delay is dominated by refractory mechanisms; during the period with long delays only adaptive mechanisms cause the masking; in the middle period both mechanisms are at work.Adaptation experiments performed in the presence of masking noise indicate a competitive action between masking and adaptation. At equal stimulus intensities there is a definite relation between S/N value and ISI value.The results of masking of the compound...