Ambient Noise Spectrum Research Articles

Speech Reinforcement Based on Soft Decision under Far-End Noise Environments

In this letter, we propose a speech reinforcement technique based on soft decision under both the far-end and near-end noise environments. We amplify the estimated clean speech signal at the far-end based on the estimated ambient noise spectrum at the near-end, as opposed to reinforcing the noisy far-end speech signal, so that it can be heard more intelligibly in far-end noisy environments. To obtain an effective reinforcement technique, we adopt the soft decision scheme incorporating a speech absence probability (SAP) in the frequency dependent signal-to-noise ratio (SNR) recovery method where the clean speech spectrum is estimated and the reinforcement gain is inherently derived and modified within the unified framework. Performance of the proposed method is evaluated by a subjective testing under various noisy environments. This is an improvement over previous approaches.

Oceanic ambient noise as a background to acoustic neutrino detection

Ambient noise measured in the deep ocean is studied in the context of a search for signals from ultra-high energy cosmic ray neutrinos. The spectral shape of the noise at the relevant high frequencies is found to be very stable for an extensive data set collected over several months from 49 hydrophones mounted near the bottom of the ocean at ~1600 m depth. The slopes of the ambient noise spectra above 15 kHz are found to roll-off faster than the -6 dB/octave seen in Knudsen spectra. A model attributing the source to an uniform distribution of surface noise that includes frequency-dependent absorption at large depth is found to fit the data well up to 25 kHz. This depth dependent model should therefore be used in analysis methods of acoustic neutrino pulse detection that require the expected noise spectra.

Analysis of wind generated shallow water ambient noise

Underwater ambient noise measurements were carried out in very shallow water (10‐20 m) in the archipelago of the Gulf of Finland during the full year. Weather conditions varied from calm sea to near gale winds. The lowest spectral levels were obtained under the ice cover during the coldest winter months. No seasonal effect was observed in the measured spectra. The role of water depth was modeled with several bottom sediments in order to identify possible channel effects at lower frequencies of the noise spectra. The ambient noise spectra are typically bandlimited. A bandpass filter model was thus fitted to the noise spectra. The filter model provides several useful parameters to characterize the effect of wind speed on the noise spectra. The cutoff frequency of the high‐pass part of the spectra decreases from ca 400 Hz down to below 200 Hz as wind increases from light breeze to near gale force. The corresponding spectral slopes get steeper with increasing winds saturating to the value of 12 dB/oct already in fresh breeze. The other parameters used in the analysis are high frequency (low‐pass) spectral slope, the frequency of maximum spectral level, noise bandwidth and total power in the band.

Stream ambient noise, spectrum and propagation of sounds in the goby Padogobius martensii: Sound pressure and particle velocity

The most sensitive hearing and peak frequencies of courtship calls of the stream goby, Padogobius martensii, fall within a quiet window at around 100 Hz in the ambient noise spectrum. Acoustic pressure was previously measured although Padogobius likely responds to particle motion. In this study a combination pressure (p) and particle velocity (u) detector was utilized to describe ambient noise of the habitat, the characteristics of the goby's sounds and their attenuation with distance. The ambient noise (AN) spectrum is generally similar for p and u (including the quiet window at noisy locations), although the energy distribution of u spectrum is shifted up by 50-100 Hz. The energy distribution of the goby's sounds is similar for p and u spectra of the Tonal sound, whereas the pulse-train sound exhibits larger p-u differences. Transmission loss was high for sound p and u: energy decays 6-10 dB10 cm, and sound pu ratio does not change with distance from the source in the nearfield. The measurement of particle velocity of stream AN and P. martensii sounds indicates that this species is well adapted to communicate acoustically in a complex noisy shallow-water environment.

Comparison of wind driven ambient noise in shallow and deep-water ocean environments

Measured sound pressure time series in shallow and deep-water ocean environments are used to investigate the character of wind-generated noise over a range of wind speed of 2–40 knots and a frequency band of 10–3,000 Hz. In 2006, acoustic data were collected during tropical storm Ernesto off the New Jersey continental shelf in shallow (70 m) water. Ambient noise from these data is compared to deep-water measurements at several locations in the Northwest and the Northeast Pacific. The results also are compared to previously reported measurements by [Kuperman and Ferla (1985)], [Wenz (1962)], and others. The analysis includes a technique for separating background shipping noise from the wind-induced component of ambient noise. The system electronic noise floor and other extraneous sources of noise are evaluated to assure that the ambient noise measurements are not corrupted. Physical mechanisms examined include a wind-saturation effect that limits the ability to acoustically monitor distant storms and the effect of shallow water propagation on the character of wind noise. The identification of the distant shipping part of the spectrum allows for inferences to be made on the low frequency part of the ambient noise spectrum. [Work supported by ONR.]

Wind dependence of ambient noise in shallow water of Bay of Bengal

This work reports on investigations into the wind dependence of ambient noise in the Bay of Bengal. Ambient noise measurements were made in the shallow water of Bay of Bengal using a portable broadband, high frequency data acquisition system together with a sensitive hydrophone suspended from the measuring platform at a depth of 5 m from the surface where the ocean depth was 25 m. Periodic measurements were carried out for one year corresponding to a wind speed range between 2 m/s and 9 m/s during summer, monsoon and winter seasons. The proportionality of the noise level with wind speed for frequencies ranging from 500 Hz to 6 kHz for each season was studied. The analysis reveals that the correlation between the wind speed and the ambient noise spectrum level was higher at lower frequencies. The results of empirical fitting based on analysis were used for noise level prediction and the model predictions compare well with the measured noise level. Further it was observed that the wind generated noise level measured during summer was approximately 8 dB less than that in other seasons. On the other hand the proportionality between the noise level and the wind speed was less during winter.

Underwater noise as source of information on conditions and dynamics of ocean environments

Leonid Brekhovskikh wrote in his book The Ocean and the Human (1987): ‘‘Ocean noise is as important oceanographic parameter as temperature, current, and wind.’’ Brekhovskikh created and headed the Laboratory of Acoustic Methods of Ocean Research in 1956. One of the scientific directions of the Laboratory was investigation of underwater noise both as interference for sound reception and a source of environmental information. Long-term studies on the unique acoustic research vessels created under the initiative of Brekhovskikh resulted in numerous important findings, including ambient noise spectra and envelopes of acoustic fluctuations, depth dependence of noise directivity, and mechanisms of ambient noise generation. Brekhovskikh was always eager to find practical applications of scientific results. Different methods of ensuring noise immunity of hydroacoustic arrays were developed under his supervision. Passive methods of acoustic navigation based on use of natural noise were suggested. Techniques for underwater acoustic monitoring of the ocean based either on ambient noise analysis or reemission of noise from a point away from the receiving system have been developed. The success of the team of scientists headed by Brekhovskikh was determined by the creative atmosphere around him: there was neither competition nor commercial interests. The common goal was knowledge of the ocean.

Song Sparrow (Melospiza Melodia) Song Varies with Urban Noise

AbstractIn urban environments, anthropogenic noise may mask bird song, especially the notes occurring at lower frequencies (1–2 kHz). Birds living in urban environments may modify their songs, particularly the low-frequency portions, to minimize masking by anthropogenic noise. Such modifications have been observed in Great Tits (Parus major) in The Netherlands, as well as in some mammals. We studied Song Sparrows (Melospiza melodia), which are common in both urban and rural environments in much of North America, and recorded the songs of 28 free- living males in Portland, Oregon. We also measured the amplitude and spectrum of ambient noise at singing locations. Song Sparrows singing at noisier locations exhibited higher-frequency low notes and had relatively less energy (amplitude) in the low-frequency range of their songs (1–4 kHz), where most anthropogenic noise also occurred. Although the mechanism(s) producing the correlation are as yet undetermined, the observed match between song and noise may result from behavioral plasticity. We discuss explanations for these patterns and how to test them.Le Chant de Melospiza melodia Varie avec le Bruit Urbain

SONG SPARROW (MELOSPIZA MELODIA) SONG VARIES WITH URBAN NOISE

Abstract In urban environments, anthropogenic noise may mask bird song, especially the notes occurring at lower frequencies (1–2 kHz). Birds living in urban environments may modify their songs, particularly the low-frequency portions, to minimize masking by anthropogenic noise. Such modifications have been observed in Great Tits (Parus major) in The Netherlands, as well as in some mammals. We studied Song Sparrows (Melospiza melodia), which are common in both urban and rural environments in much of North America, and recorded the songs of 28 free- living males in Portland, Oregon. We also measured the amplitude and spectrum of ambient noise at singing locations. Song Sparrows singing at noisier locations exhibited higher-frequency low notes and had relatively less energy (amplitude) in the low-frequency range of their songs (1–4 kHz), where most anthropogenic noise also occurred. Although the mechanism(s) producing the correlation are as yet undetermined, the observed match between song and noise may resu...

Measurements of Ambient Noise Spectra in the South Norwegian Sea

Measurements have been made of ambient sea noise north of the Shetland Islands in the southern part of the Norwegian Sea. Shore-terminated, omnidirectional hydrophones, bottomed in 240 and 350 fathoms, were used. Twenty minute samples of single hydrophone outputs were recorded during daylight hours about every 3 days over a 4-year period from the summer of 1957 to 1961. Amplitude levels were measured at selected frequencies in the band 30 cps to 1 kc. Matching observations of wind and weather were coordinated with the noise data. The noise spectrum was found to be relatively flat compared with the spectrum measured in deep ocean areas of the western North Atlantic. Also, the pressure levels at frequencies above 100 cps are considerably higher. Seasonal changes were observed in the spectrum below 400 cps. The annual excursion in level from winter high to midsummer low was about 5 db at 50 cps and 3.5 db at 100 cps. Although no seasonal changes occurred in the upper spectrum, day-to-day variations in level at the higher frequencies generally followed patterns of change in local weather. There was good correlation between wind velocity and ambient noise at 800 and at 1000 cps. The character of sea-noise in the area seems to be shaped to two factors. First, a combination of thermal structure and a very gradual bottom slope result in high propagation losses. This in turn restricts the acoustic access at a bottomed hydrophone to a localized field. Second, except for a few summer calms, weather in the islands is severe, and strong winds and high sea states are sustained throughout the year.

Acoustic pressure-vector sensor array

Pressure-vector sensors measure both scalar and vector components of the acoustic field. December 2003 measurements at the NUWC Seneca Lake test facility verify previous observations that acoustic ambient noise spectrum levels measured by acoustic intensity sensors are reduced relative to either acoustic pressure or acoustic vector sensor spectrum levels. The Seneca measurements indicate a reduction by as much as 15 dB at the upper measurement frequency of 2500 Hz. A nonlinear array synthesis theory for pressure-vector sensors will be introduced that allows smaller apertures to achieve narrow beams. The significantly reduced ambient noise of individual pressure-vector elements observed in the ocean by others, and now at Seneca Lake, should allow a nonlinearly combined array to detect significantly lower levels than has been observed in previous multiplicative processing of pressure sensors alone. Nonlinear array synthesis of pressure-vector sensors differs from conventional super-directive algorithms that linearly combine pressure elements with positive and negative weights, thereby reducing the sensitivity of conventional super-directive arrays. The much smaller aperture of acoustic pressure-vector sensor arrays will be attractive for acoustic systems on underwater vehicles, as well as for other applications that require narrow beam acoustic receivers. [The authors gratefully acknowledge the support of ONR and NUWC.]

Scaling laws for bubble acoustic radiation strength in whitecaps

The past decade has seen significant advances in our understanding of open-ocean bubble plumes, including scaling laws for whitecap properties [W. K. Melville and P. Matusov, Nature (London) 417, 58–63 (2002)] and the physics of bubble creation [Garrett et al., J. Phys. Oceanogr. 30, 2163–2171 (2000), and G. B. Deane and M. D. Stokes, Nature (London) 418, 839–844 (2002)]. Despite this progress, we do not yet have a complete understanding of the relationship between noise radiated by whitecaps and the bubble creation processes occurring within the wave crest. Understanding this relationship is important: it would allow bubble-mediated gas transfer rates to be estimated from wind speed, for example, and provide a model for the Knudsen spectrum of underwater ambient noise. An important piece of this puzzle is the dependence of bubble acoustic radiation strength on bubble radius. We will present estimates of the scaling laws for this relationship based on open-ocean measurements of whitecap noise and laboratory measurements of bubble creation rates in plunging breakers. [Work supported by ONR and NSF.]

Protection from acoustic trauma is not a primary function of the medial olivocochlear efferent system.

The medial olivocochlear (MOC) efferent system is an important component of an active mechanical outer hair cell system in mammals. An extensive neurophysiological literature demonstrates that the MOC system attenuates the response of the cochlea to sound by reducing the gain of the outer hair cell mechanical response to stimulation. Despite a growing understanding of MOC physiology, the biological role of the MOC system in mammalian audition remains uncertain. Some evidence suggests that the MOC system functions in a protective role by acting to reduce receptor damage during intense acoustic exposure. For the MOC system to have evolved as a protective mechanism, however, the inner ears of mammals must be exposed to potentially damaging sources of noise that can elicit MOC-mediated protective effects under natural conditions. In this review, we evaluate the possibility that the MOC system evolved to protect the inner ear from naturally occurring environmental noise. Our survey of nonanthropogenic noise levels shows that while sustained sources of broadband noise are found in nearly all natural acoustic environments, frequency-averaged ambient noise levels in these environments rarely exceed 70 dB SPL. Similarly, sources reporting ambient noise spectra in natural acoustic environments suggest that noise levels within narrow frequency bands are typically low in intensity (<40 dB SPL). Only in rare instances (e.g., during frog choruses) are ambient noise levels sustained at moderately high intensities (~70-90 dB SPL). By contrast, all experiments in which an MOC-mediated protective effect was demonstrated used much higher sound intensities to traumatize the cochlea (100-150 dB SPL). This substantial difference between natural ambient noise levels and the experimental conditions necessary to evoke MOC-mediated protection suggests that even the noisiest natural acoustic environments are not sufficiently intense to have selected for the evolution of the MOC system as a protective mechanism. Furthermore, although relatively intense noise environments do exist in nature, they are insufficiently distributed to account for the widespread distribution of the MOC system in mammals. The paucity of high-intensity noise and the near ubiquity of low-level noise in natural environments supports the hypothesis that the MOC system evolved as a mechanism for "unmasking" biologically significant acoustic stimuli by reducing the response of the cochlea to simultaneous low-level noise. This suggested role enjoys widespread experimental support.

Habitat-dependent song divergence in the little greenbul: an analysis of environmental selection pressures on acoustic signals.

Bird song is a sexual trait important in mate choice and known to be shaped by environmental selection. Here we investigate the ecological factors shaping song variation across a rainforest gradient in central Africa. We show that the little greenbul (Andropadus virens), previously shown to vary morphologically across the gradient in fitness-related characters, also varies with respect to song characteristics. Acoustic features, including minimum and maximum frequency, and delivery rate of song notes showed significant differences between habitats. In contrast, we found dialectal variation independent of habitat in population-typical songtype sequences. This pattern is consistent with ongoing gene flow across habitats and in line with the view that song variation in the order in which songtypes are produced is not dependent on habitat characteristics in the same way physical song characteristics are. Sound transmission characteristics of the two habitats did not vary significantly, but analyses of ambient noise spectra revealed dramatic and consistent habitat-dependent differences. Matching between low ambient noise levels for low frequencies in the rainforest and lower minimal frequencies in greenbul songs in this habitat suggests that part of the song divergence may be driven by habitat-dependent ambient noise patterns. These results suggest that habitat-dependent selection may act simultaneously on traits of ecological importance and those important in prezygotic isolation, leading to an association between morphological and acoustic divergence. Such an association may promote assortative mating and may be a mechanism driving reproductive divergence across ecological gradients.

Environmental correlates of vocal communication of wild pygmy marmosets, Cebuella pygmaea

We quantified the acoustic characteristics of the habitats of two wild populations of pygmy marmosets, in Amazonian Ecuador to evaluate their effects on vocal signal structure. We obtained measures of ambient noise, sound attenuation and reverberation through recordings and broadcasts of exemplars of two short-range vocalizations and one long-range vocalization of the marmosets. Ambient noise levels differed among habitats. The calls of pygmy marmosets had frequencies that coincided with relatively quiet regions of the ambient noise spectra of the habitats. The three vocalization types were degraded similarly in all habitats. The two short-range signals, with a pulsatile structure, were more affected by reverberation than was the long-range, less pulsatile vocalization. This degradation could be used by the marmosets to estimate the distance of the caller animals. We obtained data on context of vocalizations from six groups of pygmy marmosets, three from each population, in both the dry and rainy seasons. The use of Trills, J calls and Long calls was related to the distance between the calling animal and the potential receivers suggesting that marmosets are using the calls in a way appropriate to the effects of habitat acoustics. Copyright 2002 The Association for the Study of Animal Behaviour. Published by Elsevier Science Ltd. All rights reserved.

Channel-tolerant FH-MFSK acoustic signaling for undersea communications and networks

Undersea acoustic channels can exhibit multipath propagation with impulse-response duration and coherence time both of the order of tens to hundreds of milliseconds. Signal reception is further impaired by the presence of time-varying nonwhite ambient-noise spectra having a dynamic range of 30 dB or more. Acoustic communication requires appropriate waveform design and associated signal processing to accommodate these adverse transmission characteristics while also providing desired performance features such as low-probability-of-detection (LPD) and multi-access networking. Adaptive-equalization techniques provide good performance only for channels with stable multi-paths and high signal-to-noise ratios (SNR's) accommodating the signaling rates needed to sample and compensate for rapid changes. An alternative approach is to design for robustness against channel fluctuations. This paper describes a channel-tolerant approach identified as "telesonar type-B signaling." The technique has been designed to accommodate network architectures requiring multiple access to the channel while simultaneously providing covertness and energy efficiency. Specialized frequency-hopped M-ary frequency-shift-key (FH-MFSK) waveforms are combined with related signal processing, including nonlinear adaptive techniques to mitigate the effects of all types of interference. This effectively results in a channel that has uniformly distributed noise in both time and frequency. Powerful error-correction coding permits low SNR transmissions. Nonbinary, long-constraint-length, convolutional coding and related sequential decoding is a classical solution for difficult low-rate channels. The probability of bit errors below 10/sup -10/ is obtainable, even in Rayleigh-faded channels near the computational cutoff rate, and the probability of failure to decode frames of data is extremely small. Both simulations and analyses of at-sea experiments demonstrate the performance of this noncoherent approach to reliable acoustic communications.

The two domains in envelope spectrum of ocean ambient noise

Analyzing level fluctuations of the natural ambient underwater noise shows that the spectrum of the envelope usually consists of two frequency domains of different physical nature. The low-frequency domain is governed by variations in wind speed, that is, variations in surface density of the source power. The high-frequency domain is related to rough-surface sound scattering. The boundary between the domains correspond to frequencies of 0.05–0.1 Hz. Experiments have shown that both cross-frequency noise correlation and correlation between the noise and wind speed fluctuations are high if only frequencies lower than 0.1 Hz are retained in the spectrum, these correlations being low for unfiltered noise. Thus, if there is no ship-traffic noise, the frequency correlation coefficient remains to be 0.6–0.8 within 0.1–10 kHz if the high-frequency domain is cut off and it is close to zero in disjoint filtering bands for unfiltered noise. The experimental data are discussed and interpreted with the use of the noise model allowing for space–time parameters of the wind speed.

Adaptive pulselength correction (APLECORR): a strategy for waveform optimization in ultrawideband active sonar

This paper describes adaptive pulselength correction (APLECORR), an environmentally adaptive technique for optimizing the detection performance in wide-band active sonars in so-called doubly spread channels. It works by allocating available transmit energy to frequency bands according to the in situ measured reverberation and ambient noise spectra. It optimizes the waveform and the detection processor at the same time. It is appealing in its simplicity and achieves significant gains whenever the reverberation-to-noise ratio is not constant across frequency, thus its applicability to wide-band systems. The method extends easily to PRN and other non-FM waveforms. The paper includes a proof that time spreading and frequency-spreading distortion have an approximately equivalent effect if the waveform is linear or hyperbolic frequency modulation.

Ambient noise in the natural surf zone: wave-breaking frequencies

Ambient noise in the surf zone, in the frequency range 120 Hz to 5 kHz, was recorded using a broad-band hydrophone, located approximately 1 m above bottom and 1-2 m below the mean sea surface. The predominant source of this noise is breaking waves. Analysis of simultaneous land-based video observations of the sea surface in the region of the hydrophone, along with wave height data, reveals quantitative correlation between wave-breaking events and the hydrophone signal. In energetic surf, locally breaking waves appear as discrete events in the ambient noise spectra. Distant breaking events do not appear to be detected, as distinct events above the ambient background noise, by the hydrophone. The noise events associated with local breakers are characterized by an asymmetry in the time envelope: low frequencies (less than 500 Hz) are observed leading the breaking crest, followed by a broader range of frequencies peaking in intensity with the passage of the wave crest above the hydrophone, and then decreasing abruptly at all frequencies. Low frequencies are generally not observed trailing the breaking wave. The detection by the hydrophone of breaking waves in the immediate vicinity implies that ambient noise in heavy surf provides a means of studying breaking-wave statistics in the surf zone in situ: in particular, the frequency of occurrence of local breaking.

Monitoring ocean seismicity using autonomous hydrophone moorings

Following the successful use of the U.S. Navy’s SOSUS for monitoring low-level seismicity in the northeast Pacific, NOAA developed a strategy for monitoring remote areas of the world ocean not covered by fixed hydrophone arrays. This strategy requires the deployment of autonomous, hydrophone moorings able to record low-frequency acoustic energy in the SOFAR channel over extended periods. The instrument includes a single ceramic hydrophone, filter/amplifier stage designed to prewhiten the ambient noise spectrum, accurate clock (1 s/yr), and processor modified from off-the-shelf hardware. The instrument is capable of recording at 100 Hz (1 to 40-Hz bandpass) for periods of up to 1 year and is designed to be serviced at sea. A 2-month field test was conducted in the northeast Pacific during May–July 1995 to allow a direct comparison of the method with existing SOSUS arrays and other seismic experiments along the Juan de Fuca Ridge. The results indicate that the sensitivity of the moored systems is comparable to SOSUS detections. In May 1996, the array was deployed in the eastern equatorial Pacific to begin long-term monitoring of the East Pacific Rise between 20N and 20S, the Galapagos Ridge, and the Mid-America Trench.