A Study on How Pre-whitening Influences Fundamental Frequency Estimation

Jaramillo, Alfredo Esquivel; Nielsen, Jesper Kjær; Christensen, Mads Græsbøll

doi:10.1109/icassp.2019.8683653

Cited by 9 publications

(11 citation statements)

References 23 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The harmonic distortion of a DUT can distort the original harmonic structure of the input signal, and cause performance degradation by affecting the balance of Mel-frequency cepstrum coefficient (MFCC) [36][37][38]. Then, the noise of the DUT will flood the harmonic structural features to some extent, making it more difficult to be extracted [39]. Therefore, the SNR, SINAD, THD and ADDR performance of the DUTs indirectly affect the accuracy of ASR.…”

Section: Simulationmentioning

confidence: 99%

An Audio Distortion Dynamic Range Index for Evaluating the Performance of Audio Systems

et al. 2020

View full text Add to dashboard Cite

The dynamic range (DR) index lacking of an official definition leads to ambiguities in performance evaluation. The existing measurement methods of DR do not always match with the various actual application conditions, and some detailed distortion behavior of the device under test (DUT) is not extracted. In this paper, a new index for evaluating the DR performance of audio systems is proposed and validated, herein referred to as the audio distortion dynamic range (ADDR). It reduces the uncertainty of measurement conditions by an explicit definition and unifies the signal-to-noise ratio (SNR) and the signal-to-noise-and-distortion ratio (SINAD) indexes if under the same measurement condition. Moreover, to comprehensively reflect the impact of harmonic, spurious, and noise components on the DUT, the definitions of the traditional indexes based on classification of distorted components are replaced by the variable thresholds in the ADDR definition. Subsequently, the detailed steps of ADDR and the critical factors influencing its accuracy, are analyzed and then the optimized measurement conditions are given. Experiments based on simulated DUTs show the ADDR index can distinguish performance difference that the traditional indexes cannot distinguish, which proves it is an effective supplementary to the existing indexes in some real applications.

show abstract

Section: Simulationmentioning

confidence: 99%

An Audio Distortion Dynamic Range Index for Evaluating the Performance of Audio Systems

et al. 2020

View full text Add to dashboard Cite

show abstract

“…Therefore, a prewhitening step is required to deal with the inconsistency between the white Gaussian noise model assumption and real life noise model. A linear prediction (LP) based prewhitening step is applied to each frame to deal with the non-white Gaussian noise (see [9], [43] for detail). The power spectral density (PSD) of the noise given noisy signals is estimated using the minimum mean-square error (MMSE) estimator [44].…”

Section: Prewhiteningmentioning

confidence: 99%

Robust Bayesian Pitch Tracking Based on the Harmonic Model

Shi

Nielsen

Jensen

et al. 2019

IEEE/ACM Trans. Audio Speech Lang. Process.

View full text Add to dashboard Cite

Fundamental frequency is one of the most important characteristics of speech and audio signals. Harmonic modelbased fundamental frequency estimators offer a higher estimation accuracy and robustness against noise than the widely used autocorrelation-based methods. However, the traditional harmonic model-based estimators do not take the temporal smoothness of the fundamental frequency, the model order, and the voicing into account as they process each data segment independently. In this paper, a fully Bayesian fundamental frequency tracking algorithm based on the harmonic model and a first-order Markov process model is proposed. Smoothness priors are imposed on the fundamental frequencies, model orders, and voicing using first-order Markov process models. Using these Markov models, fundamental frequency estimation and voicing detection errors can be reduced. Using the harmonic model, the proposed fundamental frequency tracker has an improved robustness to noise. An analytical form of the likelihood function, which can be computed efficiently, is derived. Compared to the state-of-the-art neural network and non-parametric approaches, the proposed fundamental frequency tracking algorithm has superior performance in almost all investigated scenarios, especially in noisy conditions. For example, under 0 dB white Gaussian noise, the proposed algorithm reduces the mean absolute errors and gross errors by 15% and 20% on the Keele pitch database and 36% and 26% on sustained /a/ sounds from a database of Parkinson's disease voices. A MATLAB version of the proposed algorithm is made freely available for reproduction of the results 1 .

show abstract

“…One example is when sub-harmonic errors appear when estimating the fundamental frequency (a.k.a. pitch) of voiced speech segments [6], [7] from estimators which assume WGN. A preprocessor which renders the coloured noise closer to white, namely a pre-whitener, can alleviate this problem.…”

Section: Introductionmentioning

confidence: 99%

“…A preprocessor which renders the coloured noise closer to white, namely a pre-whitener, can alleviate this problem. Applying pre-whitening using a linear filter is advantageous compared to a general linear transformation with, e.g., the Cholesky factor [4], since the effect of linear filtering can be modeled by only changing the sinusoidal amplitudes and phases [6], [7]. Unlike general linear transformations, linear filtering thus enables us to use many existing model-based estimators based on a WGN assumption.…”

Section: Introductionmentioning

confidence: 99%

Adaptive Pre-whitening Based on Parametric NMF

Jaramillo

Nielsen

Christensen

2019

2019 27th European Signal Processing Conference (EUSIPCO)

Self Cite

View full text Add to dashboard Cite

Several speech processing methods assume that a clean signal is observed in white Gaussian noise (WGN). An argument against those methods is that the WGN assumption is not valid in many real acoustic scenarios. To take into account the coloured nature of the noise, a pre-whitening filter which renders the background noise closer to white can be applied. This paper introduces an adaptive pre-whitener based on a supervised non-negative matrix factorization (NMF), in which a pre-trained dictionary includes parametrized spectral information about the noise and speech sources in the form of autoregressive (AR) coefficients. Results show that the noise can get closer to white, in comparison to pre-whiteners based on conventional noise power spectral density (PSD) estimates such as minimum statistics and MMSE. A better pitch estimation accuracy can be achieved as well. Speech enhancement based on the WGN assumption shows a similar performance to the conventional enhancement which makes use of the background noise PSD estimate, which reveals that the proposed pre-whitener can preserve the signal of interest.

show abstract

A Study on How Pre-whitening Influences Fundamental Frequency Estimation

Cited by 9 publications

References 23 publications

An Audio Distortion Dynamic Range Index for Evaluating the Performance of Audio Systems

An Audio Distortion Dynamic Range Index for Evaluating the Performance of Audio Systems

Robust Bayesian Pitch Tracking Based on the Harmonic Model

Adaptive Pre-whitening Based on Parametric NMF

Contact Info

Product

Resources

About