GSM speech coding and speaker recognition

Besacier, Laurent; Grassi, S.; Dufaux, Alain; Ansorge, Michael; Pellandini, Fausto

doi:10.1109/icassp.2000.859152

Cited by 34 publications

(18 citation statements)

References 10 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Experiment 2, Simulation 1 testing and training the small source/microphone distance in Room 1 there are not room parameters significant levels of room reverberation and hence SI accuracy stage where speed is not a concern and storage requirements for speaker models are relatively small. The proposed method also requires N times as many likelihood calculations in the test stage [compare (5) with (4)] which will increase identification time. tion, the frequency with which the training rooms were assoTraining 2 (4.0, 3.0, 3.5) ciated with the identified speaker were tabulated and are given Training 3 (4.0, 3.0, 3.5)…”

Section: Test Signalmentioning

confidence: 99%

“…Because it is unrealistic to know in advance packet due to speech coding and distortions due to packet loss. In loss rates and somewhat difficult to accurately measure loss [5], the authors passed TIMIT signals through Global System rates, training channels can only approximate test channels. for Mobile (GSM) speech coders and measured SI accuracy Nevertheless, it was found that if a set of packet loss models of approximately 60% (40% lower compared to TIMIT).…”

Section: Introduction Weighted Mixture Of W Gaussian Pdfsmentioning

confidence: 99%

See 1 more Smart Citation

Speaker Identification in the Presence of Room Reverberation

León

Trevizo

2007

2007 Biometrics Symposium

View full text Add to dashboard Cite

Training Feature Xm} Speaker Speaker identification (SI) systems based on Gaussian MixSignal Extraction Modeling Speaker, s ture Models (GMMs) have demonstrated high levels of accuracy when both training and testing signals are acquired in (a) near ideal conditions. These same systems when trained and {Xg.. s } tested with signals acquired under non-ideal channels such as telephone have been shown to have markedly lower accuracy Test Signal Fte X Litklo Spae levels. In this paper, we consider a reverberant test environUnknown 3etraci n CopuaihoSdenai-ye ment and its impact on SI. We measure the degradation in SI Speaker, s v accuracy when the system is trained with clean signals but (b) tested with reverberant signals. Next, we propose a method whereby training signals are first filtered with a family of Fig. 1. (a) Training and (b) testing stages in SI reverberation filters prior to construction of speaker models; the reverberation filters are designed to approximate expected test room reverberation. Reverberant test signals are then scored against the family of speaker models and identification is made. Our research demonstrates that by approximatthe speaker's training feature vectors. The GMM-based aping test room reverberation in the training signals, the channel proach has shown to be very successful in accurately idenmismatch problem can be reduced and SI accuracy increased. tifying speakers from a large population [2]. In utilizing a GMM, we assume the probability density function (pdf) for feature vector x given speaker model A, can be modeled as a

show abstract

Section: Test Signalmentioning

confidence: 99%

Section: Introduction Weighted Mixture Of W Gaussian Pdfsmentioning

confidence: 99%

Speaker Identification in the Presence of Room Reverberation

León

Trevizo

2007

2007 Biometrics Symposium

View full text Add to dashboard Cite

show abstract

“…where Q nm denotes 4 bits of the channel quality identifier (CQI) value and A nm denotes the current channel utilization status [11].…”

Section: Fig1 Radio Resources Grid Of Lte Networkmentioning

confidence: 99%

Dynamic spectrum sharing algorithm for combined mobile networks

Klymash

Maksymyuk

et al. 2014

2014 20th International Conference on Microwaves, Radar and Wireless Communications (MIKON)

View full text Add to dashboard Cite

This paper proposes a new dynamic spectrumsharing algorithm for combined cellular networks which can allow users to access both networks simultaneously using the same spectrum resource. This article proposes an algorithm to save spectrum resources due to the limited and costly bands. The proposed method for enhancing the spectrum utilization of combined radio interface (CRI) is called the dynamic spectrum sharing algorithm (DSSA). The proposed algorithm dynamically reallocates the occupied channels to meet the quality of service (QoS) requirements for each user. This is a different idea from traditional static spectrum sharing algorithms (SSSA). In the proposed DSSA, spectrum utilization is enhanced by dividing the entire spectrum into three separate zones: determination zone, which is occupied concurrently by GSM or LTE users, and fixed zones just for GSM and LTE users, respectively.

show abstract

“…Speaker recognition performance with GSM coded speech has found to improve and become comparable to recognition on the original speech with the use of speech coding parameters for recognition e.g., cepstral coefficients derived from the LP spectrum [5] or Line Spectral Frequencies (LSFs) [6]. Similarly, speech recognition performance was also found to improve when using cepstral coefficients derived from speech codec parameters, compared to recognition on the decoded speech [7] [8].…”

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

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

GSM speech coding and speaker recognition

Cited by 34 publications

References 10 publications

Speaker Identification in the Presence of Room Reverberation

Speaker Identification in the Presence of Room Reverberation

Dynamic spectrum sharing algorithm for combined mobile networks

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

Contact Info

Product

Resources

About