We propose a time-domain audio source separation method using down-sampling (DS) and up-sampling (US) layers based on a discrete wavelet transform (DWT). The proposed method is based on one of the state-of-the-art deep neural networks, Wave-U-Net, which successively down-samples and up-samples feature maps. We find that this architecture resembles that of multiresolution analysis, and reveal that the DS layers of Wave-U-Net cause aliasing and may discard information useful for the separation. Although the effects of these problems may be reduced by training, to achieve a more reliable source separation method, we should design DS layers capable of overcoming the problems. With this belief, focusing on the fact that the DWT has an anti-aliasing filter and the perfect reconstruction property, we design the proposed layers. Experiments on music source separation show the efficacy of the proposed method and the importance of simultaneously considering the anti-aliasing filters and the perfect reconstruction property.
Ambisonics is a complete theory for spatial audio whose building blocks are the spherical harmonics. Some of the drawbacks of low order Ambisonics, like poor source directivity and small sweet-spot, are directly related to the properties of spherical harmonics. In this thesis we illustrate a novel spatial audio framework similar in spirit to Ambisonics that replaces the spherical harmonics by an alternative set of functions with compact support: the spherical wavelets. We develop a complete audio chain from encoding to decoding, using discrete spherical wavelets built on a multiresolution mesh…
Comparative Analysis of Mel-Frequency Cepstral Coefficients and Wavelet Based Audio Signal Processing for Emotion Detection and Mental Health Assessment in Spoken Speech
Dipartimento di Ingegneria Navale, Elettrica, Elettronica e delle Telecomunicazioni (DITEN)（海军、电气、电子与电信工程系）；Universidad EAFIT（EAFIT大学）
We propose a study of the mathematical properties of voice as an audio signal. This work includes signals in which the channel conditions are not ideal for emotion recognition. Multiresolution analysis discrete wavelet transform was performed through the use of Daubechies Wavelet Family (Db1-Haar, Db 6, Db8, Db10) allowing the decomposition of the initial audio signal into sets of coefficients on which a set of features was extracted and analyzed statistically in order to differentiate emotional states. ANNs proved to be a system that allows an appropriate classification of such states. This s…
We introduce the joint time-frequency scattering transform, a time shift invariant descriptor of time-frequency structure for audio classification. It is obtained by applying a two-dimensional wavelet transform in time and log-frequency to a time-frequency wavelet scalogram. We show that this descriptor successfully characterizes complex time-frequency phenomena such as time-varying filters and frequency modulated excitations. State-of-the-art results are achieved for signal reconstruction and phone segment classification on the TIMIT dataset.
Signal denoising is a key preprocessing step for many applications, as the performance of a learning task is closely related to the quality of the input data. In this paper, we apply a signal processing based deep neural network architecture, a learnable extension of the wavelet packet transform. As main advantages, this model has few parameters, an intuitive initialization and strong learning capabilities. Moreover, we show that it is possible to easily modify the parameters of the model after the training step to tailor to different noise intensities. Two case studies are conducted to compar…
African Institute for Mathematical Sciences（非洲数学科学研究所）；ETH Zurich（苏黎世联邦理工学院）；Rochester Institute of Technology（罗切斯特理工学院）
Facing the diversity and growth of the musical field nowadays, the search for precise songs becomes more and more complex. The identity of the singer facilitates this search. In this project, we focus on the problem of identifying the singer by using different methods for feature extraction. Particularly, we introduce the Discrete Wavelet Transform (DWT) for this purpose. To the best of our knowledge, DWT has never been used this way before in the context of singer identification. This process consists of three crucial parts. First, the vocal signal is separated from the background music by us…
Institute of Mechatronic Systems（机电系统研究所）；School of Engineering（工程学院）；Zurich University of Applied Sciences (ZHAW)（苏黎世应用科技大学（ZHAW））
In the present paper, details are given on the implementation of a wavelet-based analysis tailored to the processing of acoustical signals. The family of the suitable wavelets (`Reimann wavelets') are obtained in the time domain from a Fourier transform, extracted in Ref.~\cite{r1} after invoking theoretical principles and time-frequency localisation constraints. A scheme is set forth to determine the optimal values of the parameters of this type of wavelet on the basis of the goodness of the reproduction of a $30$-s audio file containing harmonic signals corresponding to six successive $A$ no…
Environmental Sound Classification (ESC) is a rapidly evolving field that recently demonstrated the advantages of application of visual domain techniques to the audio-related tasks. Previous studies indicate that the domain-specific modification of cross-domain approaches show a promise in pushing the whole area of ESC forward. In this paper, we present a new time-frequency transformation layer that is based on complex frequency B-spline (fbsp) wavelets. Being used with a high-performance audio classification model, the proposed fbsp-layer provides an accuracy improvement over the previously u…
NYU, Courant Institute of Mathematical Sciences（纽约大学库朗数学科学研究所）；Facebook AI Research（Facebook人工智能研究院）
In this report we describe an ongoing line of research for solving single-channel source separation problems. Many monaural signal decomposition techniques proposed in the literature operate on a feature space consisting of a time-frequency representation of the input data. A challenge faced by these approaches is to effectively exploit the temporal dependencies of the signals at scales larger than the duration of a time-frame. In this work we propose to tackle this problem by modeling the signals using a time-frequency representation with multiple temporal resolutions. The proposed representa…
Detection and Classification of Cetacean Echolocation Clicks using Image-based Object Detection Methods applied to Advanced Wavelet-based Transformations
Flatiron Institute（熨斗研究所）；Cornell Lab of Ornithology, Cornell University（康奈尔大学康奈尔鸟类学实验室）；Music and Audio Research Laboratory, New York University（纽约大学音乐与音频研究实验室）；Ecole Normale Supérieure（巴黎高等师范学院）；Collège de France（法兰西公学院）
In time series classification and regression, signals are typically mapped into some intermediate representation used for constructing models. Since the underlying task is often insensitive to time shifts, these representations are required to be time-shift invariant. We introduce the joint time-frequency scattering transform, a time-shift invariant representation which characterizes the multiscale energy distribution of a signal in time and frequency. It is computed through wavelet convolutions and modulus non-linearities and may therefore be implemented as a deep convolutional neural network…
