A Comparative Analysis of Pitch Detection Methods Under the Influence of Different Noise Conditions

Sukhostat, Lyudmila; İmamverdiyev, Yadigar

doi:10.1016/j.jvoice.2014.09.016

Cited by 29 publications

(19 citation statements)

References 25 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…This suggests SWIPE to be more appropriate for pitch estimation for the classification task at hand. It has been reported in [23] that the voicing decision error made by PEFAC is higher than that of SWIPE under both clean and additive white Gaussian noise conditions. Since speaking pattern is expected to carry important information for ALS/PD vs. HC classification, the erroneous voicing decisions might lead to an inferior classification performance using PEFAC than SWIPE.…”

Section: Resultsmentioning

confidence: 92%

Effect of Noise and Model Complexity on Detection of Amyotrophic Lateral Sclerosis and Parkinson’s Disease Using Pitch and MFCC

Bhattacharjee

Mallela

Atchayarcmf

et al. 2021

ICASSP 2021 - 2021 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP)

View full text Add to dashboard Cite

Dysarthria due to Amyotrophic Lateral Sclerosis (ALS) and Parkinson's disease (PD) impacts both articulation and prosody in an individual's speech. Complex deep neural networks exploit these cues for detection of ALS and PD. These are typically done using recordings in laboratory condition. This study aims to examine the robustness of these cues against background noise and model complexity, which has not been investigated before. We perform classification experiments with pitch and Mel-frequency cepstral coefficients (MFCC) using models of three different complexities and additive white Gaussian noise in four signal-to-noise-ratio (SNR) conditions. The findings are as follows: 1) In clean condition, pitch performs similar to MFCC across most model complexities considered, suggesting that one-dimensional pitch pattern provides discriminative cues for the classification to an extent equal to that of multi-dimensional MFCC, 2) Similar trend is observed in noisy cases when classifiers are trained and tested in matched noise and SNR conditions, 3) When the classifiers trained on clean data are applied in noisy cases, pitch based average classification accuracies are found to be 20.09% and 24.73% higher than those using MFCC for ALS vs. healthy and PD vs. healthy, respectively, suggesting robustness of pitch based classifier against noise and model complexity.

show abstract

Section: Resultsmentioning

confidence: 92%

Effect of Noise and Model Complexity on Detection of Amyotrophic Lateral Sclerosis and Parkinson’s Disease Using Pitch and MFCC

Bhattacharjee

Mallela

Atchayarcmf

et al. 2021

ICASSP 2021 - 2021 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP)

View full text Add to dashboard Cite

show abstract

“…The pitch extraction performance of the conventional methods (YIN [4] and BaNa [20]) and the FROOT and FROOT+ methods was investigated in noisy environments. In [29], BaNa was assessed as the best pitch extractor in noisy environments among nine methods that were compared. YIN was the second-best method in [22], where the best one was DNN-based.…”

Section: Performance Comparisonmentioning

confidence: 99%

Pitch Extraction Using Fourth-Root Spectrum in Noisy Speech

Rahman

Sugiura

Shimamura

2020

Journal of Signal Processing

View full text Add to dashboard Cite

In this paper, we present the use of the fourth-root spectrum instead of the log spectrum for pitch extraction in noisy environments. To obtain clear harmonics, lifter and clipping operations are performed. When the resulting spectrum is transformed into the time domain by the discrete Fourier transform, pitch detection is robust against narrow-band noise. When the same spectrum is amplified by power calculation and transformed into the time domain, pitch detection is robust against wide-band noise. These properties are investigated through exhaustive experiments in various noises. The required computational time is also studied.

show abstract

“…Los artículos seleccionados se clasifican en tres grandes grupos, de acuerdo a su aporte en la resolución de alguna de las tres preguntas de investigación. Una segunda clasificación se realiza entre los artículos para determinar cuáles son los principales métodos empleados en la detección de la frecuencia fundamental de la voz; las categorías definidas en esta clasificación se basan en la planteada en [11]. La lectura de los resúmenes de los artículos permite dilucidar en cada caso el método base para dicha detección.…”

Section: Esquema De Clasificaciónunclassified

Técnicas de detección de la frecuencia fundamental de la voz en entornos reales

et al. 2017

View full text Add to dashboard Cite

Introducción: este artículo de revisión se desarrolló como parte de un trabajo de maestría de la Universidad del Cauca en el 2017. Buscó encontrar los métodos de detección de la frecuencia fundamental de la voz más propicios para implementar en entornos reales como parte de una solución para mejorar la comunicación de las personas con discapacidad auditiva e incluirlas en la sociedad, ya que la mayoría de las propuestas solo plantean mejorar el canal de comunicación en el que la persona con discapacidad auditiva es el transmisor. Metodología: se realizó una revisión actualizada de la literatura, por lo que se basó principalmente en artículos científicos publicados en los últimos cinco años. Para la inclusión de artículos se hizo un mapeo sistemático sobre los diferentes métodos de detección de la frecuencia fundamental de la voz. Resultados: los fenómenos contemplados por los diferentes algoritmos para definir el entorno van desde el ruido y la interferencia hasta la reverberación; el desempeño del algoritmo depende de la calidad del audio grabado, lo que se ve en las variaciones obtenidas que dependen de la base de datos utilizada; se pueden detectar hasta dos frecuencias fundamentales diferentes. Conclusiones: se han implementado novedosos métodos para hacer más eficiente la detección de la frecuencia fundamental de la voz; sin embargo, aún queda mucho trabajo por hacer en esta área.

show abstract

A Comparative Analysis of Pitch Detection Methods Under the Influence of Different Noise Conditions

Cited by 29 publications

References 25 publications

Effect of Noise and Model Complexity on Detection of Amyotrophic Lateral Sclerosis and Parkinson’s Disease Using Pitch and MFCC

Effect of Noise and Model Complexity on Detection of Amyotrophic Lateral Sclerosis and Parkinson’s Disease Using Pitch and MFCC

Pitch Extraction Using Fourth-Root Spectrum in Noisy Speech

Técnicas de detección de la frecuencia fundamental de la voz en entornos reales

Contact Info

Product

Resources

About