Azimuth sound localization in many animals relies on the processing of differences in time-of-arrival of the low-frequency sounds at both ears: the interaural time differences (ITD). It was observed in some species that this cue depends on the spectrum of the signal emitted by the source. Yet, this variation is often discarded, as humans and animals are assumed to be insensitive to it. The purpose of this thesis is to assess this dependency using acoustical techniques, and explore the consequences of this additional complexity on the neurophysiology and psychophysics of sound localization. In the vicinity of rigid spheres, a sound field is diffracted, leading to frequency-dependent wave propagation regimes. Therefore, when the head is modeled as a rigid sphere, the ITD for a given position is a frequency-dependent quantity. I show that this is indeed reflected on human ITDs by studying acoustical recordings for a large number of human and animal subjects. Furthermore, I explain the effect of this variation at two scales. Locally in frequency the ITD introduces different envelope and fine structure delays in the signals reaching the ears. Second the ITD for low-frequency sounds is generally bigger than for high frequency sounds coming from the same position. In a second part, I introduce and discuss the current views on the binaural ITD-sensitive system in mammals. I expose that the heterogenous responses of such cells are well predicted when it is assumed that they are tuned to frequency-dependent ITDs. Furthermore, I discuss how those cells can be made to be tuned to a particular position in space irregardless of the frequency content of the stimulus. Overall, I argue that current data in mammals is consistent with the hypothesis that cells are tuned to a single position in space. Finally, I explore the impact of the frequency-dependence of ITD on human behavior, using psychoacoustical techniques. Subjects are asked to match the lateral position of sounds presented with different frequency content. Those results suggest that humans perceive sounds with different frequency contents at the same position provided that they have different ITDs, as predicted from acoustical data. The extent to which this occurs is well predicted by a spherical model of the head. Combining approaches from different fields, I show that the binaural system is remarkably adapted to the cues available in its environment. This processing strategy used by animals can be of great inspiration to the design of robotic systems.

The increasing number of electric and hybrid vehicles in urban areas has shown to be beneficial in reducing both air and noise pollution. However, the lack of sound when driving at low speed has negatively affected the pedestrian safety since many rely on the vehicle sound for orientation. Regulatory bodies have therefore introduced minimum sound requirements for all silent vehicles, which has resulted in a key challenge for car manufacturers to develop external warning signature sounds. The objective of this project has been to study how these signals should sound in order to fit the image of electric and hybrid vehicles and minimize noise and annoyance. To complete the study, five sound concepts with different characteristics and rhythms were developed using the concept generation process. In order to gain subjective impressions of the sounds, three measurement methods were utilized. The first listening test was performed to eliminate the possibility that the sound samples would be perceived variously loud in the upcoming tests. The second listening test was performed to gain information on how suitable the signals are for electric and hybrid vehicles, and the third listening test was conducted to obtain information on how much annoyance the signals contributed with. A reference signal representing an internal combustion engine was included in the tests as well. The results of the measurement methods showed that the sound samples with long tone sequences were more preferred than those with short tone sequences, and that the artificial approach was more popular than the sound of an internal combustion engine vehicle. It was also established that additional tests need to be performed in order to confirm that these results are valid. It was suggested that field tests should be conducted and that new sound concepts should be developed based on the results of these tests.

A habilidade auditiva de resolução temporal consiste no tempo mínimo requerido para segregar ou resolver eventos acústicos. Esta habilidade é fundamental para a compreensão da fala humana, constituindo-se num pré-requisito para as habilidades lingüísticas, bem como para a leitura. Em 2003, Musiek desenvolveu um teste para avaliar os limiares de detecção de gap a ser utilizado na prática clínica - o GIN - Gap In Noise (Musiek et al., 2004). Para que o teste GIN possa ser incorporado à bateria de testes para avaliação do processamento auditivo, é necessário que existam critérios de normalidade para ouvintes sem alterações auditivas. O objetivo geral do presente trabalho é estabelecer critérios de normalidade para o teste GIN, em adultos com audição normal. Como objetivos específicos, têm-se: obter as médias dos limiares de detecção de gap, a porcentagem média de acertos, bem como definir um intervalo de confiança para cada uma das faixas-teste que compõem o GIN; obter o desempenho por intervalo de gap; verificar o efeito das variáveis orelha, gênero e faixa-teste. O teste GIN foi aplicado em 100 indivíduos (50 do gênero feminino e 50 do gênero masculino), de faixa-etária entre 18 e 31 anos, após a realização de outros testes audiológicos para descartar possíveis alterações auditivas e/ou do processamento auditivo, que pudessem comprometer os resultados. Como resultados gerais, foram observados limiares de detecção de gap e porcentagens médias de acertos semelhantes para as orelhas direita e esquerda, para os gêneros masculino e feminino e para as quatro faixas-teste testadas. A média geral dos limiares de gap foi de 3,98 ms, enquanto a média das porcentagens de acertos foi de 78,89%. Foi definido um intervalo de confiança (limite mínimo e limite máximo) para cada uma das faixas-teste (Média dos limiares de detecção de gap - faixa-teste 1: 3,73 - 4,01 ms; faixa-teste 2: 3,9 - 4,18 ms; faixa-teste 3: 3,88 - 4,14 ms; faixa-teste 4: 3,9 - 4,14 ms; Porcentagens médias de acertos - faixa-teste 1: 78,14 - 80,52%; faixa-teste 2: 77,34 - 79,66%; faixa-teste 3: 77,73 - 79,83%; faixa-teste 4: 77,82 - 80,14%). A curva do desempenho dos participantes por intervalo de gap foi estabelecida da seguinte forma: para gaps de 2 ms, a porcentagem de acertos foi sempre igual ou menor do que 5%; para 3 ms, esta porcentagem já fica em torno de 10 a 30%; para intervalos de 4 ms, as porcentagens de acerto chegam ao redor de 60 a 70%; para intervalos de gap iguais ou maiores do que 5 ms, a porcentagem de acertos alcança 90% ou mais. Todos estes resultados poderão ser utilizados como parâmetros de normalidade. Desta forma, o teste em questão mostrou-se consistente e com baixa variabilidade, em relação aos dados obtidos para os 100 indivíduos. / Auditory temporal resolution ability refers to the shortest time required to segregate or to resolve acoustic events. This ability is important to human speech comprehension and it is a prerequisite condition for both linguistic and reading abilities. In 2003, Musiek developed a clinical test to measure gap detection thresholds - the GIN Test - Gap In Noise (Musiek et al., 2004). In order to incorporate the GIN test to auditory processing evaluation, it is necessary to established norms in normal hearing subjects. The aims of this study were: to establish parameters for the GIN test in normal-hearing young adults; to obtain both gap detection threshold and percentage of correct responses mean; to define a confidence interval to the four lists that compound GIN test; to get a performance slope for each gap interval; to verify the variable effect concerning ear, gender and list. Proceed by an audiological evaluation to exclude hearing loss and/or auditory processing disorders, the GIN test was applied in 100 subjects (50 females and 50 males), ranged from 18 to 31 years old. Results indicated that the gap detection threshold and the percentage of correct responses means were quite similar in both right and left ears, in male and female gender and in the four tested lists. The gap detection threshold mean was 3,98 ms, and the percentage of correct responses mean was 78,89%. A confidence interval was defined (minimum and maximum limits) to each one of four lists (Gap detection threshold mean - List 1: 3,73 - 4,01 ms; List 2: 3,9 - 4,18 ms; List 3: 3,88 - 4,14 ms; List 4: 3,9 - 4,14 ms; Percentage of correct responses mean - List 1: 78,14 - 80,52%; List 2: 77,34 - 79,66%; List 3: 77,73 - 79,83%; List 4: 77,82 - 80,14%). The performance slope for each gap interval was determined according to the following criteria: to 2 ms gap interval the percentage of accurate responses were less or equal 5%; to 3 ms gap interval it varied from 10% to 30%; to 4 ms gap interval, accurate responses were from 60 to 70%; and, finally, from 5 ms and to upper intervals a 90 or more percentage was achieved. All achieved results may be applied as standard parameters. Finally, the GIN test demonstrated be consistent and to have low variability in relation to the data from the 100 subjects.

Audio coding is used to compress digital audio signals, thereby reducing the amount of bits needed to transmit or to store an audio signal. This is useful when network bandwidth or storage capacity is very limited. Audio compression algorithms are based on an encoding and decoding process. In the encoding step, the uncompressed audio signal is transformed into a coded representation, thereby compressing the audio signal. Thereafter, the coded audio signal eventually needs to be restored (e.g. for playing back) through decoding of the coded audio signal. The decoder receives the bitstream and reconverts it into an uncompressed signal. ISO-MPEG is a standard for high-quality, low bit-rate video and audio coding. The audio part of the standard is composed by algorithms for high-quality low-bit-rate audio coding, i.e. algorithms that reduce the original bit-rate, while guaranteeing high quality of the audio signal. The audio coding algorithms consists of MPEG-1 (with three different layers), MPEG-2, MPEG-2 AAC, and MPEG-4. This work presents a study of the MPEG-4 AAC audio coding algorithm. Besides, it presents the implementation of the AAC algorithm on different platforms, and comparisons among implementations. The implementations are in C language, in Assembly of Intel Pentium, in C-language using DSP processor, and in HDL. Since each implementation has its own application niche, each one is valid as a final solution. Moreover, another purpose of this work is the comparison among these implementations, considering estimated costs, execution time, and advantages and disadvantages of each one.

Audio coding is used to compress digital audio signals, thereby reducing the amount of bits needed to transmit or to store an audio signal. This is useful when network bandwidth or storage capacity is very limited. Audio compression algorithms are based on an encoding and decoding process. In the encoding step, the uncompressed audio signal is transformed into a coded representation, thereby compressing the audio signal. Thereafter, the coded audio signal eventually needs to be restored (e.g. for playing back) through decoding of the coded audio signal. The decoder receives the bitstream and reconverts it into an uncompressed signal. ISO-MPEG is a standard for high-quality, low bit-rate video and audio coding. The audio part of the standard is composed by algorithms for high-quality low-bit-rate audio coding, i.e. algorithms that reduce the original bit-rate, while guaranteeing high quality of the audio signal. The audio coding algorithms consists of MPEG-1 (with three different layers), MPEG-2, MPEG-2 AAC, and MPEG-4. This work presents a study of the MPEG-4 AAC audio coding algorithm. Besides, it presents the implementation of the AAC algorithm on different platforms, and comparisons among implementations. The implementations are in C language, in Assembly of Intel Pentium, in C-language using DSP processor, and in HDL. Since each implementation has its own application niche, each one is valid as a final solution. Moreover, another purpose of this work is the comparison among these implementations, considering estimated costs, execution time, and advantages and disadvantages of each one.

Audio coding is used to compress digital audio signals, thereby reducing the amount of bits needed to transmit or to store an audio signal. This is useful when network bandwidth or storage capacity is very limited. Audio compression algorithms are based on an encoding and decoding process. In the encoding step, the uncompressed audio signal is transformed into a coded representation, thereby compressing the audio signal. Thereafter, the coded audio signal eventually needs to be restored (e.g. for playing back) through decoding of the coded audio signal. The decoder receives the bitstream and reconverts it into an uncompressed signal. ISO-MPEG is a standard for high-quality, low bit-rate video and audio coding. The audio part of the standard is composed by algorithms for high-quality low-bit-rate audio coding, i.e. algorithms that reduce the original bit-rate, while guaranteeing high quality of the audio signal. The audio coding algorithms consists of MPEG-1 (with three different layers), MPEG-2, MPEG-2 AAC, and MPEG-4. This work presents a study of the MPEG-4 AAC audio coding algorithm. Besides, it presents the implementation of the AAC algorithm on different platforms, and comparisons among implementations. The implementations are in C language, in Assembly of Intel Pentium, in C-language using DSP processor, and in HDL. Since each implementation has its own application niche, each one is valid as a final solution. Moreover, another purpose of this work is the comparison among these implementations, considering estimated costs, execution time, and advantages and disadvantages of each one.

A habilidade auditiva de resolução temporal consiste no tempo mínimo requerido para segregar ou resolver eventos acústicos. Esta habilidade é fundamental para a compreensão da fala humana, constituindo-se num pré-requisito para as habilidades lingüísticas, bem como para a leitura. Em 2003, Musiek desenvolveu um teste para avaliar os limiares de detecção de gap a ser utilizado na prática clínica - o GIN - Gap In Noise (Musiek et al., 2004). Para que o teste GIN possa ser incorporado à bateria de testes para avaliação do processamento auditivo, é necessário que existam critérios de normalidade para ouvintes sem alterações auditivas. O objetivo geral do presente trabalho é estabelecer critérios de normalidade para o teste GIN, em adultos com audição normal. Como objetivos específicos, têm-se: obter as médias dos limiares de detecção de gap, a porcentagem média de acertos, bem como definir um intervalo de confiança para cada uma das faixas-teste que compõem o GIN; obter o desempenho por intervalo de gap; verificar o efeito das variáveis orelha, gênero e faixa-teste. O teste GIN foi aplicado em 100 indivíduos (50 do gênero feminino e 50 do gênero masculino), de faixa-etária entre 18 e 31 anos, após a realização de outros testes audiológicos para descartar possíveis alterações auditivas e/ou do processamento auditivo, que pudessem comprometer os resultados. Como resultados gerais, foram observados limiares de detecção de gap e porcentagens médias de acertos semelhantes para as orelhas direita e esquerda, para os gêneros masculino e feminino e para as quatro faixas-teste testadas. A média geral dos limiares de gap foi de 3,98 ms, enquanto a média das porcentagens de acertos foi de 78,89%. Foi definido um intervalo de confiança (limite mínimo e limite máximo) para cada uma das faixas-teste (Média dos limiares de detecção de gap - faixa-teste 1: 3,73 - 4,01 ms; faixa-teste 2: 3,9 - 4,18 ms; faixa-teste 3: 3,88 - 4,14 ms; faixa-teste 4: 3,9 - 4,14 ms; Porcentagens médias de acertos - faixa-teste 1: 78,14 - 80,52%; faixa-teste 2: 77,34 - 79,66%; faixa-teste 3: 77,73 - 79,83%; faixa-teste 4: 77,82 - 80,14%). A curva do desempenho dos participantes por intervalo de gap foi estabelecida da seguinte forma: para gaps de 2 ms, a porcentagem de acertos foi sempre igual ou menor do que 5%; para 3 ms, esta porcentagem já fica em torno de 10 a 30%; para intervalos de 4 ms, as porcentagens de acerto chegam ao redor de 60 a 70%; para intervalos de gap iguais ou maiores do que 5 ms, a porcentagem de acertos alcança 90% ou mais. Todos estes resultados poderão ser utilizados como parâmetros de normalidade. Desta forma, o teste em questão mostrou-se consistente e com baixa variabilidade, em relação aos dados obtidos para os 100 indivíduos. / Auditory temporal resolution ability refers to the shortest time required to segregate or to resolve acoustic events. This ability is important to human speech comprehension and it is a prerequisite condition for both linguistic and reading abilities. In 2003, Musiek developed a clinical test to measure gap detection thresholds - the GIN Test - Gap In Noise (Musiek et al., 2004). In order to incorporate the GIN test to auditory processing evaluation, it is necessary to established norms in normal hearing subjects. The aims of this study were: to establish parameters for the GIN test in normal-hearing young adults; to obtain both gap detection threshold and percentage of correct responses mean; to define a confidence interval to the four lists that compound GIN test; to get a performance slope for each gap interval; to verify the variable effect concerning ear, gender and list. Proceed by an audiological evaluation to exclude hearing loss and/or auditory processing disorders, the GIN test was applied in 100 subjects (50 females and 50 males), ranged from 18 to 31 years old. Results indicated that the gap detection threshold and the percentage of correct responses means were quite similar in both right and left ears, in male and female gender and in the four tested lists. The gap detection threshold mean was 3,98 ms, and the percentage of correct responses mean was 78,89%. A confidence interval was defined (minimum and maximum limits) to each one of four lists (Gap detection threshold mean - List 1: 3,73 - 4,01 ms; List 2: 3,9 - 4,18 ms; List 3: 3,88 - 4,14 ms; List 4: 3,9 - 4,14 ms; Percentage of correct responses mean - List 1: 78,14 - 80,52%; List 2: 77,34 - 79,66%; List 3: 77,73 - 79,83%; List 4: 77,82 - 80,14%). The performance slope for each gap interval was determined according to the following criteria: to 2 ms gap interval the percentage of accurate responses were less or equal 5%; to 3 ms gap interval it varied from 10% to 30%; to 4 ms gap interval, accurate responses were from 60 to 70%; and, finally, from 5 ms and to upper intervals a 90 or more percentage was achieved. All achieved results may be applied as standard parameters. Finally, the GIN test demonstrated be consistent and to have low variability in relation to the data from the 100 subjects.

Orientador: Luís Geraldo Pedroso Meloni / Dissertação (mestrado) - Universidade Estadual de Campinas, Faculdade de Engenharia Elétrica e de Computação / Made available in DSpace on 2018-08-21T21:07:48Z (GMT). No. of bitstreams: 1 Chanquini_JuliaJacobsenDornelles_M.pdf: 2607975 bytes, checksum: f9b57a1325c9977a5bfd0cdb69a56661 (MD5) Previous issue date: 2012 / Resumo: Este trabalho apresenta a adequação de um codificador de áudio padrão MPEG-4 AAC para aderência à norma brasileira do SBTVD. Também apresenta um estudo dos conceitos envolvidos em codificadores de áudio perceptuais com enfoque no codificador MPEG-4 AAC e também inclui a parte de multiplexação e sincronia do MPEG-4. Para o desenvolvimento do projeto foram estudados alguns códigos abertos de codificadores AAC: FAAD, 3GPP e o código de referência do padrão MPEG-4, especialmente a parte referente ao LATM/LOAS. O decodificador de áudio padrão MPEG-4 AAC que foi modificado para suportar a camada LATM/ LOAS foi o FAAD. Foi calculado o tempo adicional que o decodificador modificado leva para decodificar o áudio com a camada LATM/LOAS, sem ser notado um aumento significativo que não permite a decodificação em tempo real do áudio / Abstract: This work presents an adaptation of a standard MPEG-4 AAC audio coder to conform to the Brazilian digital TV standard SBTVD. It also presents a study of the concepts involved in perceptual audio coders focusing on MPEG-4 AAC and also including the multiplexing and synchronization part of the MPEG-4 standard. To develop this project, open source AAC coders were studied: FAAD, 3GPP and the MPEG-4 reference software code specially the part concerning LATM/LOAS. The AAC audio decoder which was modified to support the LATM / LOAS layer was FAAD. The additional time that the modified decoder needs to decode a sample audio with LATM / LOAS was calculated, and it did not introduce a large enough delay that would restrict real time audio decoding / Mestrado / Telecomunicações e Telemática / Mestra em Engenharia Elétrica

This Thesis deals with a method to enforce the intellectual property rights and protect digital media from tampering – Digital Audio Watermarking. The main aim of this work is implement an audio watermarking algorithm. The theoretical part defined basic terms, methods and processes, which are used in this area. The practical part shows a process of embedding the digital signature into a host signal and her backward extraction. The embedding rule used spread spectrum technique and a psychoacoustic model. The implemented psychoacoustic model involves two properties of the human auditory system which are frequency masking and representation the frequency scale on limited bands called critical bands. The model is relatively new and based on the DWPT. In terms of above model is then the digital watermark embedded in the wavelet domain. This algorithm is implemented in technical software MATLAB. One part of this work focuses on robustness tests of the algorithm. Common signal processing modifications are applied to the watermarked audio as follows: Cutting of the audio, re-sampling, lossy compression, filtering, equalization, modulation effects, noise addition. The last part of the thesis presents subjective and objective methods usable in order to judge the influence of watermarking embedding on the quality of audio tracks called transparency.

Hearing loss afflicts as many as 10\% of our population.Fortunately, technologies designed to alleviate the effects ofhearing loss are improving rapidly, including cochlear implantsand the increasing computing power of digital hearing aids. Thisthesis focuses on theoretically sound methods for improvinghearing aid technology. The main contributions are documented inthree research articles, which treat two separate topics:modelling of human speech recognition (Papers A and B) andoptimization of diagnostic methods for hearing loss (Paper C).Papers A and B present a hidden Markov model-based framework forsimulating speech recognition in noisy conditions using auditorymodels and signal detection theory. In Paper A, a model of normaland impaired hearing is employed, in which a subject's pure-tonehearing thresholds are used to adapt the model to the individual.In Paper B, the framework is modified to simulate hearing with acochlear implant (CI). Two models of hearing with CI arepresented: a simple, functional model and a biologically inspiredmodel. The models are adapted to the individual CI user bysimulating a spectral discrimination test. The framework canestimate speech recognition ability for a given hearing impairmentor cochlear implant user. This estimate could potentially be usedto optimize hearing aid settings.Paper C presents a novel method for sequentially choosing thesound level and frequency for pure-tone audiometry. A Gaussianmixture model (GMM) is used to represent the probabilitydistribution of hearing thresholds at 8 frequencies. The GMM isfitted to over 100,000 hearing thresholds from a clinicaldatabase. After each response, the GMM is updated using Bayesianinference. The sound level and frequency are chosen so as tomaximize a predefined objective function, such as the entropy ofthe probability distribution. It is found through simulation thatan average of 48 tone presentations are needed to achieve the sameaccuracy as the standard method, which requires an average of 135presentations.