1999
DOI: 10.1109/9780470544402
|View full text |Cite
|
Sign up to set email alerts
|

Discrete-Time Processing of Speech Signals

Help me understand this report

Search citation statements

Order By: Relevance

Paper Sections

Select...
1
1
1
1

Citation Types

5
927
0
71

Year Published

1999
1999
2012
2012

Publication Types

Select...
6
2

Relationship

0
8

Authors

Journals

citations
Cited by 1,228 publications
(1,046 citation statements)
references
References 0 publications
5
927
0
71
Order By: Relevance
“…Short-term spectral estimates became popular acoustic features, computed using filter banks at first, then using signal processing techniques like the Fast Fourier Transform (FFT). In the 1960's, homomorphic or cepstral analysis was developed [3]. Since then, cepstral-based feature sets have become the standard acoustic analysis technique for speech recognition.…”
Section: History Of Asrmentioning
confidence: 99%
See 2 more Smart Citations
“…Short-term spectral estimates became popular acoustic features, computed using filter banks at first, then using signal processing techniques like the Fast Fourier Transform (FFT). In the 1960's, homomorphic or cepstral analysis was developed [3]. Since then, cepstral-based feature sets have become the standard acoustic analysis technique for speech recognition.…”
Section: History Of Asrmentioning
confidence: 99%
“…These coefficients correspond directly to the general spectral envelope of a segment of speech. Because cepstral coefficients are computed using linear signal processing techniques, they are unable to capture nonlinear or higher-order statistics of the speech signals [3]. Nonlinear processes do not obey the principle of superposition, and cannot be fully represented by power spectra.…”
Section: Pulse Trainmentioning
confidence: 99%
See 1 more Smart Citation
“…Para tanto, são usadas técnicas no domí-nio do tempo (análise temporal) e/ou no domínio da frequência (análise espectral) [46]. Nesse âmbito, a energia de um sinal temporal x[·] com N amostras discretas é um escalar definido, segundo [7], como…”
Section: Parâmetros Utilizados Energia Temporalunclassified
“…The function of this preemphasis is to reduce the effects of the glottal pulses and radiation impedance. It also takes the focus to the spectral properties of the vocal tract [6]. This is followed by a bandpass filter, which removes constant and low frequency components of background, as well as high frequency noise and speech harmonics due to sampling and nonlinearity of the recording device.…”
Section: Implementation Examplementioning
confidence: 99%