Effects of speech coding on text-dependent speaker recognition

Phythian, Mark; Ingram, John C. L.; Sridharan, Sridha

doi:10.1109/tencon.1997.647276

Cited by 9 publications

(8 citation statements)

References 2 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…In brief, CELP consists of two stages (please refer to [9][10] for complete coder details): Linear Prediction (LP) analysis followed by an adaptive and fixed codebook vector quantization of the excitation (error residual) signal. AMR (at 8kHz) utilizes a 10 th order LP analysis process and AMR-WB (at 16kHz) 16 th order LP, given by: (1) where â i denotes the ith LP coefficient. LP coefficients are calculated for every analysis frame, where both AMR and AMR-WB use 20ms frames.…”

Section: Amr and Amr-wb Speech Codecsmentioning

confidence: 99%

“…The effect of mobile telephony speech compression on speech parameters has been previously studied: Phythian et al [1] found the Code-Excited Linear Prediction (CELP) and Linear Prediction (LP) based GSM speech coders to negatively impact the vowel Fpatterns and fundamental frequency (F0). Further, McClelland [2] found an increase of up to 30Hz on F0 estimations from mobile telephony over landline quality speech, whilst Byrne and Foulkes [3] found 29% higher vowel F1 formants from mobile telephony speech compared to direct speech recordings.…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

On the effect of amr and AMR-WB GSM compression on overlapped speech for forensic analysis

Cheng

Burnett

2011

2011 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP)

View full text Add to dashboard Cite

The recent ubiquity of mobile telephony has posed the challenge of forensic speech analysis on compressed speech content. Whilst existing research studies have investigated the effect of mobile speech compression on speaker and speech parameters, this paper addresses the effect of speech compression on parameters when an interfering background speaker is present in clean and noisy conditions. Preliminary evaluations presented in this paper study the effect of the Adaptive Multi-Rate (AMR) and Adaptive MultiRate Wideband (AMR-WB) speech coders on the Linear Prediction (LP) speech spectrum, Line Spectral Frequencies (LSFs), and Mel Frequency Cepstral Coefficients (MFCCs). Results indicate that due caution should be employed for the forensic analysis of mobile telephony speech: speech coder parameters are significantly degraded when an interfering speaker or noise is present, compared to parameters obtained from the main speaker alone. Moreover, at high SNR the speech parameters exhibit values that gradually transition from those ideally and independently obtained from the main speaker to those of the background speaker as the amplitude of the background interfering speaker increases.

show abstract

Section: Amr and Amr-wb Speech Codecsmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

On the effect of amr and AMR-WB GSM compression on overlapped speech for forensic analysis

Cheng

Burnett

2011

2011 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP)

View full text Add to dashboard Cite

show abstract

“…In recent years, with the rapid development of network technology, the speech is encoded with compression for effective transmit, and the bit rate is relatively low, which results in the distortion of speech signal and the decline of speaker recognition performance. Particularly, when the condition of training and testing is codec mismatch, the performance is even worse [3][4][5]. The compensation method of audio coding influence has been attracting more attentions of a variety of researchers.…”

Section: Introductionmentioning

confidence: 99%

MAP-based Audio Coding Compensation for Speaker Recognition

Jiang¹,

Han²

2011

JSIP

View full text Add to dashboard Cite

show abstract

“…According to [10], there are two methods for solving AR equations, autocorrelation method and covariance method. Out of these two, autocorrelation is preferred because it gives somewhat stable filters and is computationally efficient [10]. AR equations for autocorrelation method are:…”

Section: Lpc-an Introductionmentioning

confidence: 99%

“…Where R is Toeplitz matrix, a is the vector of the LPC coefficients and r is the autocorrelation. The autocorrelation sequence is given by The redundancy of AR equations can efficiently be computed by Levinson Derbin Regression [10].…”

Section: Lpc-an Introductionmentioning

confidence: 99%

An Automatic Speaker Recognition System

Akrouf¹,

Mehamel²,

Benhamouda³

et al. 2009

International Conference on Advanced Computer Theory and Engineering (ICACTE 2009)

View full text Add to dashboard Cite

Speech is becoming the privileged mean of communication. With the advent of computers, researchers and industrials are using it like a mean of communication in the Human Machine Interface domain. All the current systems, based on the natural language, are far from assimilating all the characteristics of the language. Nowadays The Automatic Treatment of speech is progressing, in particular in the fields of Automatic Speech Recognition “ASR” and Speech Synthesis. The automatic speaker recognition is represented like a particular pattern recognition task. It associates the problems relating to the speaker identification or verification using information found in the acoustic signal: we have to recognize a person by using his voice. ASR is used in many fields, like domestic, military or jurisprudence applications. In this work we explore the automatic speaker recognition and present an Automatic Speaker Recognition System. We will use speaker recognition in text independent mode since we dispose of very few training data. We have to estimate with few data a robust speaker model to allow the recognition of the speaker.

show abstract

Effects of speech coding on text-dependent speaker recognition

Cited by 9 publications

References 2 publications

On the effect of amr and AMR-WB GSM compression on overlapped speech for forensic analysis

On the effect of amr and AMR-WB GSM compression on overlapped speech for forensic analysis

MAP-based Audio Coding Compensation for Speaker Recognition

An Automatic Speaker Recognition System

Contact Info

Product

Resources

About