Comparison of Speech Features on the Speech Recognition Task

Abstract

In the present work we overview some recently proposed discrete Fourier transform (DFT)- and discrete wavelet packet transform (DWPT)-based speech parameterization methods and evaluate their performance on the speech recognition task. Specifically, in order to assess the practical value of these less studied speech parameterization methods, we evaluate them in a common experimental setup and compare their performance against traditional techniques, such as the Mel-frequency cepstral coefficients (MFCC) and perceptual linear predictive (PLP) cepstral coefficients which presently dominate the speech recognition field. In particular, utilizing the well established TIMIT speech corpus and employing the Sphinx-III speech recognizer, we present comparative results of 8 different speech parameterization techniques

Robust feature extraction methods for speech recognition in noisy environments

“…Literature states that MFCC is most popular method because of its speech recognition performance [2]. The recognition performance mainly depends upon the choice of features and its extraction from the speech signal such that it gives a high recognition rate with less amount of computation [3]. This paper is organized as follows: Section 2 gives a brief idea about the feature extraction techniques.…”

Reconocimiento de voz basado en MFCC, SBC y Espectrogramas

“…En [1] se muestra una comparativa de distintos métodos de parametrización de voz, teniendo como base de comparación los Coeficientes Cepstrales en la Frecuencia de Mel, siendo estos los más usados debido a su bajo costo computacional y a su robustez. Con base en los resultados, se decidió analizar estas técnicas dentro del reconocimiento del habla robusto a cambios de tono en la voz, utilizando técnicas con distintas transformaciones, como la Transformada Discreta de Fourier (DFT) y la Transformada Discreta de Paquetes Wavelet (DWPT).…”

Reviewing Human-Machine Interaction through Speech Recognition approaches and Analyzing an approach for Designing an Efficient System

“…This helps in providing a closer response of human auditory system than the linearly spaced frequency bands which are derived from FFT (Fast Fourier Transform) and DCT [7] (Discrete Cosine Transform). Thus a mel frequency cepstrum results in more accurate processing of data.…”