Thesis (M.Phil.)--Hong Kong University of Science and Technology, 2008. / Includes bibliographical references (leaves 42-44). Also available in electronic version.

In passive sonar systems, knowledge of low-frequency shipping noise is an important factor for target detection performance. However, an accurate model for the shipping noise structure is difficult to obtain, due to the varying distributions of ships and complicated underwater environment. This work characterizes low-frequency distant shipping noise observed in deep water environments as a function of receiver depth and vertical arrival structure for the case of a receiver below the conjugate depth. Surface shipping noise is examined using Monte Carlo simulations using a normal mode propagation model based on random distribution of ships and realistic parameters. The depth dependence of the simulated distant shipping noise is in agreement with published experimental measurements. A Vertical Line Array (VLA) is used to produce vertical beams that isolate the surface interference from nearby targets. Simulation results quantifying the beamformer output as a function of ocean environment, receiver aperture, and frequency are presented for both conventional and adaptive beamformers. The results suggest that conventional beamforming could detect the noisy target from both direct arrival and bottom bounce in the presence of distant shipping interferers and wind noise. However, the beamwidth of conventional beamforming is wider than that of adaptive beamforming. Once the motion effects of nearby ship interferences are considered, the adaptive beamforming using diagonal loading provides better detection performance. Preliminary adaptive beamforming results corresponding to different snapshot times show that motion effects can be minimized by using short observation times.

Recent theoretical and experimental studies have demonstrated that an estimate of the Green's function between two hydrophones can be extracted passively from the cross‐correlation of ambient noise recorded at these two points. Hence monitoring the temporal evolution of these estimated Green's functions can provide a means for noise‐based acoustic tomography using a distributed sensor network. However, obtaining unbiased Green's function estimate requires a sufficiently spatially and temporally diffuse ambient noise field. Broadband ambient noise ([200 Hz-20 kHz]) was recorded continuously for 2 days during the SWAMSI09 experiment (next to Panama City, FL) using two moored vertical line arrays (VLAs) spanning 7.5m of the 20‐m water column and separated by 150 m. The feasibility of noise‐based acoustic tomography ([300-1000 Hz]) was assessed in this dynamic coastal environment over the whole recording period. Furthermore, coherent array processing of the computed ocean noise cross‐correlations between all pairwise combinations of hydrophones was used to separate acoustic variations between the VLAs caused by genuine environmental fluctuations-such as internal waves-from the apparent variations in the same coherent arrivals caused when the ambient noise field becomes strongly directional, e.g., due to an isolated ship passing in the vicinity of the VLAs.

An approximate of the Green's function can be obtained by taking the cross-correlation of ambient noise that has been simultaneously recorded on separate sensors. This method is applied for two experiments, which illustrate the advantages and challenges of this technique. The first experiment is in the ultrasonic regime [5-30] MHz and uses capacitive micromachined ultrasonic transducer arrays to image the near field and compares the passive imaging to the conventional pulse-echo imaging. Both the array and target are immersed in a fluid with the sensors recording the fluid's random thermal-mechanical motion as the ambient noise. The second experiment is a passive ocean monitoring experiment, which uses spatiotemporal filtering to rapidly extract coherent arrivals between two vertical line arrays. In this case the ambient noise in the frequency band [250 1500] Hz is dominated by non-stationary shipping noise. For imaging purposes, the cross-correlation needs to extract the Green's function so that the imaging can be done correctly. While for monitoring purposes, the important feature is the change in arrivals, which corresponds to the environment changing. Results of both experiments are presented along with the advantages of this passive method over the more accepted active methods.

In this thesis, we are interested in processing noisy speech signals that are meant to be heard by humans, and hence we approach the noise-suppression problem from a perceptual perspective. We develop a noise-suppression paradigm that is based on a model of the human auditory system, where we process signals in a way that is natural to the human ear. Under this paradigm, we transform an audio signal in to a perceptual domain, and processes the signal in this perceptual domain. This approach allows us to reduce the background noise and the audible artifacts that are seen in traditional noise-suppression algorithms, while preserving the quality of the processed speech. We develop a single- and dual-microphone algorithm based on this perceptual paradigm, and conduct subjecting tests to show that this approach outperforms traditional noise-suppression techniques. Moreover, we investigate the cause of audible artifacts that are generated as a result of suppressing the noise in noisy signals, and introduce constraints on the noise-suppression gain such that these artifacts are reduced.

In this thesis, we tackle key challenges in processing audio signals, specifically focusing on speech and music. These signals are crucial for human interaction with both the environment and machines. Our research addresses three core topics: speech denoising, speech dereverberation, and music-dance generation, each of which plays a vital role in enhancing the harmony between audio and human interaction. Leveraging machine learning and human-centric approaches inspired by classical algorithms, we develop methods to mitigate common audio degradations, such as additive noise and multiplicative reverberation, delivering high-quality audio suitable for human use and applications. Furthermore, we introduce a real-time, music-responsive system for generating 3D dance animations, advancing the integration of audio signals with human engagement. The first focus of our thesis is the elimination of additive noise from audio signals by focusing on short pauses, or silent intervals, in human speech. These brief pauses provide key insights into the noise profile, enabling our model to dynamically reduce ambient noise from speech. Tested across diverse datasets, our method outperforms traditional and audiovisual denoising techniques, showcasing its effectiveness and adaptability across different languages and even musical contexts. In the second work of our research, we address reverberation removal from audio signals, a task traditionally reliant on knowing the environment's exact impulse response—a requirement often impractical in real-world settings. Our novel solution combines the strengths of classical and learning-based approaches, tailored for online communication contexts. This human-centric method includes a one-time personalization step, adapting to specific environments and human speakers. The two-stage model, integrating feature-based Wiener deconvolution and network refinement, has shown through extensive experiments to outperform current methods, both in effectiveness and user preference. Transitioning from foundational audio signal enhancement and analysis to a more dynamic realm, our research culminates in a novel, interactive system for real-time 3D human dance generation. Contrasting with the passive human-centric assumptions of our previous works, this final work actively engages users, enabling direct interaction with a system that synchronizes expressive dance movements to live music, spanning various musical elements like type, tempo, and energy. This innovative approach, diverging from traditional choreography methods, leverages spontaneous improvisation to generate unique dance sequences. These sequences, a mix of pre-recorded choreographies and algorithm-generated transitions, adapt to real-time audio inputs, offering customization through personal 3D avatars. This system's user-centric design and interactivity are validated by user studies, confirming its effectiveness in creating an immersive and engaging user experience.

Nightclubs and restaurant bars have become major sources of noise pollution particularly in areas close to residential dwellings. The purpose of this study was to investigate the impact of noise emanating from nightclubs and restaurant bars on the community of Melville, Johannesburg. This study followed both qualitative and quantitative research methods. A total of 100 respondents were randomly sampled within the study area. Qualitative data was collected using a structured questionnaire. A calibrated sound level meter was used to measure environmental noise levels at 10 different measuring points. The research finding revealed that about:  87% of noise levels measured with the sound level meter did not comply with officially acceptable levels of 40dB at night.  69% of respondents indicated that the main source of noise is pollution is nightclubs.  78% of respondents described noise as annoying, disturbing and unwanted.  57% of respondents indicated that members of their household have suffered from sleeping disorders due to noise activities at night disrupting their sleep patterns and resulting in irritability and fatigue. The noise measurements were taken on weekends and public holidays during the day from 10h00 to 14h30 and at night from 22h00 to 02h30. The research findings revealed that the residents of Melville experienced high level of noise at night with nightclub as major source of noise and as a result the majority of the sampled population complained about irritability, fatigue and sleeping disorders due to exposure to noise. The outcome of this research indicated the need of health education on the adverse effects of noise pollution and the need of sound insulation at places of entertainment. Implementation of a noise management policy is needed in order to effectively control and manage the noise pollution in its area of jurisdiction and regular noise level monitoring by constantly taking noise measurements by law enforcements officers. / Department of Environmental Sciences / M. Sc. (Environmental Management)