cepstrum

(redirected from Cepstral)

cepstrum

[′sep·trəm]
(acoustics)
The Fourier transform of the logarithm of a speech power spectrum; used to separate vocal tract information from pitch excitation in voiced speech.
McGraw-Hill Dictionary of Scientific & Technical Terms, 6E, Copyright © 2003 by The McGraw-Hill Companies, Inc.

cepstrum

(mathematics)
(Coined in a 1963 paper by Bogert, Healey, and Tukey) The Fourier transform of the log-magnitude spectrum:

fFt(ln( | fFt(window . signal) | ))

This function is used in speech recognition and possibly elsewhere. Note that the outer transform is NOT an inverse Fourier transform (as reported in many respectable DSP texts).

This article is provided by FOLDOC - Free Online Dictionary of Computing (foldoc.org)
Mentioned in ?
References in periodicals archive ?
It provides a common application interface for Cepstral, IBM, Microsoft and Nuance speech technologies, Chant said.
TextAloud has a large variety of voices to select from because it works with a number of TTS vendors such as AT&T Natural Voices, NeoSpeech voices, Cepstral Voices, Scansoft RealSpeak Voices, and Acapelo Voices.
Additional premium voices offered optionally by NextUp.com include RealSpeak (R) and Cepstral (R).
[99] Objective voice measurements (cepstral peak and jitter) significantly improved in 12 patients 8 months after end-to-end anastomosis in a study by Olson et al.
The feature vector consisted of 16 cepstral coefficients, normalized logarithmic power, and their delta features (derivatives).
The world of millions of Fast Fourier Transforms per second and cepstral analyses, of automatic target recognition and sensor fusion, would be unachievable.
Fifteen audio event classifiers was trained using various feature vector combinations including the mel frequency cepstral coefficient (MFCC), perceptual linear prediction (PLP), and zero crossing rate (ZCR).
The commonly used feature parameters include Mel-frequency cepstral coefficient (MFCC) which has strong recognition performance and anti-noise capacity, linear predictive coefficient which has small computer load but general efficacy and accent sensitivity parameter which has favorable performance in recognition the middle frequency band of signals.
Mel-frequency cepstral coefficients (MFCC) [15], line spectral pair frequencies (LSPF) [16] and linear prediction coefficients (LPC) [17] are among the most popular spectral short-term features.
[5] studied the effect of a total of 143 audio features, showing that cepstrum-based features such as Mel-frequency cepstral coefficients (MFCCs) and Linear Predictive Cepstral Coefficients (LPCCs) are more effective than short-term and spectral features in audio classification.
Algorithm implementing this processing-intensive task most commonly combines spectrotemporal features drawn from the short-term Fourier transform (STFT) [10-14], Mel-frequency cepstral domain (MFC) [15, 16], wavelet transform [17, 18], empirical mode decomposition [19], and a variety of classification schemes, including decision trees [10, 12], neural networks [18], and support vector machine [13-15, 18].