Significance of the Modified Group Delay Feature in Speech Recognition

Hegde, Rajesh M.; Murthy, Hema A.; Gadde, Venkata Ramana Rao

doi:10.1109/tasl.2006.876858

Cited by 153 publications

(78 citation statements)

References 19 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…In general, the group delay function can be effectively utilised only if the signal under consideration is a minimum-phase signal [Hegde et al (2007b)]. 7 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64A signal is defined as a minimum phase signal only if both itself and its inverse are causal and stable.…”

Section: Representations Of Group Delay Functionsmentioning

confidence: 99%

“…In this method, the zeros near the unit circle in Equation 3 are suppressed by smoothing the magnitude spectrum X(w) [Murthy and Gadde (2003); Hegde et al (2007b)]. The resulting group delay function is well behaved.…”

Section: Modified Group Delay Function (Modgd)mentioning

confidence: 99%

“…Studies such as [Hegde et al (2007b)] have shown the have shown the equivalence between MINGD and MODGD in a least squares sense, when the signal is not minimum phase. Furthermore, it is also demonstrated in [Hegde et al (2007b)] the application of the MODGD for tasks such as syllable recognition, language identification and speaker recognition. Given these reasons, we validate the robustness property of group delay functions on the MODGD by experimental evaluation.…”

Section: Modified Group Delay Function (Modgd)mentioning

confidence: 99%

“…The speech signal was pre-emphasised, after using a frame size of 25 ms and a frame shift of 10 ms and applying a Hamming window. As specified in [Hegde et al (2007b)] the MODGDF parameters α and γ were set to 0.4 and 0.9 respectively, and the parameter lifter ω was set to 8. 13-dimensional MODGDF features were extracted.…”

Section: Utilising Modgdf For Vadmentioning

confidence: 99%

“…The group delay function has been used in applications like signal reconstruction [Yegnanarayana et al (1984)], formant extraction [Murthy and Yegnanarayana (1991)] and spectrum estimation [Yegnanarayana and Murthy (1992)]. Researchers have also come up with conventional short time feature representations derived from the group delay function [Zhu and Paliwal (2004); Hegde et al (2007b)]. Moreover, it is shown that for applications like speech and speaker recognition, magnitude-based and phase-based features achieve comparable performance, and combining them results in improvement [Zhu and Paliwal (2004); Hegde et al (2007a)].…”

Section: Introductionmentioning

confidence: 99%

See 4 more Smart Citations

Robustness of group delay representations for noisy speech signals

Parthasarathi

Rajan

Murthy

2011

Int J Speech Technol

Self Cite

View full text Add to dashboard Cite

This paper demonstrates the robustness of group delay based features to additive noise. First, we analytically show the robustness of group delay based representations. The analysis makes use of the fact that, for minimum-phase signals, the group delay function can be represented in terms of the cepstral coefficients of the log-magnitude spectrum. Such a representation results in the speech spectrum dominating over the noise spectrum, both at low and high SNRs. Further, we experimentally demonstrate the robustness of the representation on a voice activity detection (VAD) task, comparing a group delay based VAD algorithm with standard VAD methods as well as a magnitude-spectrum based method.

show abstract

Section: Representations Of Group Delay Functionsmentioning

confidence: 99%

Section: Modified Group Delay Function (Modgd)mentioning

confidence: 99%