Signal Modelling for the Digital Reconstruction of Gramophone Noise

The revenue generated from gramophone records surpassed that of ad-supported online streaming services in the United States in 2015. With the increase in popularity, a lot of old records are dug out and digitized by private audiophiles, or remastered by music labels. These old records are often scratched and damaged due to mishandling and extensive playback. This article analyses various models and algorithms than can be used to digitally refurbish the noise from gramophone records which are often perceived as crackles and pops. Different duplication approaches, trigonometric functions, polynomials, and time series models are analysed according to their signal reconstruction ability. Some novel artificial neural networks are discussed and compared to the existing models. It was found that the neural networks outperformed the other mathematical models, producing a more natural reconstruction of the audio signal with little remaining noise perceivable by the human ear.

[1]  Harris S. Shultz,et al.  The Function sin x/x. , 1990 .

[2]  Robert L. Mason,et al.  Fractional factorial design , 2009 .

[3]  Simon Dixon,et al.  Improved music feature learning with deep neural networks , 2014, 2014 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP).

[4]  Guillermo Sapiro,et al.  Audio restoration from multiple copies , 2013, 2013 IEEE International Conference on Acoustics, Speech and Signal Processing.

[5]  Gwilym M. Jenkins,et al.  Time series analysis, forecasting and control , 1972 .

[6]  Hualou Liang,et al.  Short-window spectral analysis of cortical event-related potentials by adaptive multivariate autoregressive modeling: data preprocessing, model validation, and variability assessment , 2000, Biological Cybernetics.

[7]  K. P. Seng,et al.  Multimedia signal processing using AI , 2003, 9th Asia-Pacific Conference on Communications (IEEE Cat. No.03EX732).

[8]  Andrzej Czyzewski,et al.  Learning algorithms for audio signal enhancement. Part 1: Neural network implementation for the removal of impulse distortions , 1997 .

[9]  Anup Hosangadi,et al.  Arithmetic Optimization Techniques for Hardware and Software Design: Hardware synthesis , 2010 .

[10]  Andrzej Czyzewski,et al.  Neuro-rough control of masking thresholds for audio signal enhancement , 2001, Neurocomputing.

[11]  Bronwyn H Hall,et al.  Estimation and Inference in Nonlinear Structural Models , 1974 .

[12]  C. G. Broyden.  The Convergence of a Class of Double-rank Minimization Algorithms 1. General Considerations , 1970 .

[13]  Geoffrey E. Hinton,et al.  A time-delay neural network architecture for isolated word recognition , 1990, Neural Networks.

[14]  Stig Nord,et al.  Approximation properties of the spline fit , 1967 .

[15]  D. T. Whiteside,et al.  The mathematical papers of Isaac Newton , 1967 .

[16]  D. Goldfarb.  A family of variable-metric methods derived by variational means , 1970 .

[17]  Michael I. Jordan.  Attractor dynamics and parallelism in a connectionist sequential machine , 1990 .

[18]  R. Engle.  Autoregressive conditional heteroscedasticity with estimates of the variance of United Kingdom inflation , 1982 .

[19]  Gerald Friedland,et al.  Audio concept classification with Hierarchical Deep Neural Networks , 2014, 2014 22nd European Signal Processing Conference (EUSIPCO).

[20]  Jaerin Cho.  Optimal design in regression and spline smoothing , 2007 .

[21]  C. Duchon.  Lanczos Filtering in One and Two Dimensions , 1979 .

[22]  Maciej Niedzwiecki,et al.  Smart copying-a new approach to reconstruction of audio signals , 2001, IEEE Trans. Signal Process..

[23]  Dong Yu,et al.  Context-Dependent Pre-Trained Deep Neural Networks for Large-Vocabulary Speech Recognition , 2012, IEEE Transactions on Audio, Speech, and Language Processing.

[24]  D. Shanno.  Conditioning of Quasi-Newton Methods for Function Minimization , 1970 .

[25]  Edward K. Blum,et al.  Approximation theory and feedforward networks , 1991, Neural Networks.

[26]  Henry Hoffmann,et al.  Managing performance vs. accuracy trade-offs with loop perforation , 2011, ESEC/FSE '11.

[27]  M. Hermite,et al.  Sur la formule d'interpolation de Lagrange , 1878 .

[28]  R. Fletcher,et al.  A New Approach to Variable Metric Algorithms , 1970, Comput. J..

[29]  T. Bollerslev,et al.  Generalized autoregressive conditional heteroskedasticity , 1986 .

[30]  Andries P. Engelbrecht,et al.  Predictive Noise Detection in Gramophone Records , 2016 .

[31]  Aurelio Uncini,et al.  Subband neural networks prediction for on-line audio signal recovery , 2002, IEEE Trans. Neural Networks.

[32]  Eugen Slutzky.  Summation of random causes as the source of cyclic processes , 1937 .

[33]  Andries Petrus Engelbrecht,et al.  Gramophone Noise Detection and Reconstruction Using Time Delay Artificial Neural Networks , 2017, IEEE Transactions on Systems, Man, and Cybernetics: Systems.

[34]  Peter Whittle,et al.  Hypothesis Testing in Time Series Analysis. , 1951 .