Effects of telephone transmission on the performance of formant-trajectory-based forensic voice comparison – Female voices

Zhang, Cuiling; Morrison, Geoffrey Stewart; Enzinger, Ewald; Ochoa, Felipe

doi:10.1016/j.specom.2013.01.011

Cited by 26 publications

(18 citation statements)

References 43 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…A similar conclusion can be derived from another forensically-oriented study by Zhang et al (2013), who compare formant measurements for the Mandarin triphthong /iau/ using manual and a range of automated measurements. Although the manual measurements were more accurate and yielded the best models, the authors concluded that the manual measurements were only a marginal improvement and thus did not justify the expense of using human analysts in preference to automatic formant tracking.…”

Section: Robustness Of Statistical Modelssupporting

confidence: 72%

Three steps forward for predictability. Consideration of methodological robustness, indexical and prosodic factors, and replication in the laboratory

Foulkes

Docherty

Hufnagel

et al. 2018

Linguistics Vanguard

View full text Add to dashboard Cite

AbstractThere is now abundant evidence that phonetic forms are shaped by probabilistic effects reflecting predictability or informativity. We outline a number of challenges for such work, where theoretical claims are often based on small differences in acoustic measurements, or interpretations of small statistical effect sizes. We outline caveats about the methods and assumptions encountered in many studies of predictability effects, particularly regarding corpus-based approaches. We consider the wide range of factors that influence patterns of variability in phonetic forms, taking a broad perspective on what is meant by “the message” in order to show that predictability effects need to be considered alongside many others, including indexical and prosodic factors. We suggest a number of ways forward to extend our understanding of the form-predictability relationship.

show abstract

Section: Robustness Of Statistical Modelssupporting

confidence: 72%

Three steps forward for predictability. Consideration of methodological robustness, indexical and prosodic factors, and replication in the laboratory

Foulkes

Docherty

Hufnagel

et al. 2018

Linguistics Vanguard

View full text Add to dashboard Cite

show abstract

“…This is due to the bandwidth restrictions imposed upon voice recordings [20], e.g., when analyzing telephone conversations. Modern telephone networks limit the available bandwidth to approximately 200-3500 Hz [3,32]. This hampers speaker identification because research indicates that higher formants are often peculiar to specific individuals [13].…”

Section: Discussionmentioning

confidence: 99%

Speaker identification based on artificial neural networks. Case study: the Polish vowel (pilot study)

Sałapa

Trawińska²,

Roterman

et al. 2014

Bio-Algorithms and Med-Systems

View full text Add to dashboard Cite

Current statistical methods and technologies used for speaker identification via dynamic formant frequency often involve classic multivariate analyses that must meet a number of criteria in order to be considered trustworthy. The authors propose more advanced classification techniques, including artificial neural networks. Owing to iterative learning algorithms, neural networks can be trained to detect highly complex, nonlinear relations hidden in input data. This study specifically considers feed-forward multilayer perceptron and radial basic function network models. The investigation involves an analysis of the Polish vowel ɑ (stressed or unstressed) in selected contexts described by the four lowest formant frequencies. Results indicate high accuracy of neural networks as a speaker identification tool reaching up to 100%. In addition, the authors have determined that the accuracy of classification is similar when based on a single context to when input data are aggregated over several different contexts.

show abstract

“…See [8] for details of the procedure. In line with previous studies [11,12,21], the zeroth through fourth DCT coefficient values from F2 and F3 were used as variables in the present study.…”

Section: Formant-trajectory Measurement and Parameterizationmentioning

confidence: 96%

“…Human-supervised measurements of the trajectories of the first three formants (F1, F2, and F3) of each vowel token were made using FORMANTMEASURER [24]. See [12,11] for details on the procedure for human-supervised formant-trajectory measurement. Discrete cosine transforms (DCTs) were fitted to the measured formant trajectories of all /iau/ tokens.…”

Section: Formant-trajectory Measurement and Parameterizationmentioning

confidence: 99%

“…A likelihood ratio is the probability of the observed evidence if the prosecution hypothesis were true versus if the defense hypothesis were true [4,5]. Acoustic-phonetic approaches to forensic voice comparison predominantly used the multivariate kernel density (MVKD [6]) model to calculate likelihood ratios [7,8,9,10,11,12] (see also a review in [13]). Concerns about the model's robustness, in particular numerical stability issues with the model's implementation and the difficulty of kernel density estimation, in particular when using a large number of features, prompted the development of the principal component analysis kernel density likelihood ratio (PCAKLR [14]) model.…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Likelihood Ratio Calculation in Acoustic-Phonetic Forensic Voice Comparison: Comparison of Three Statistical Modelling Approaches

Enzinger

2016

Interspeech 2016

Self Cite

View full text Add to dashboard Cite

This study compares three statistical models used to calculate likelihood ratios in acoustic-phonetic forensic-voicecomparison systems: Multivariate kernel density, principal component analysis kernel density, and a multivariate normal model. The data were coefficient values obtained from discrete cosine transforms fitted to human-supervised formant-trajectory measurements of tokens of /iau/ from a database of recordings of 60 female speakers of Chinese. Tests were conducted using high-quality recordings as nominal suspect samples and mobileto-landline transmitted recordings as nominal offender samples. Performance was assessed before and after fusion with a baseline automatic mel frequency cepstral coefficient Gaussian mixture model universal background model system. In addition, Monte Carlo simulations were used to compare the output of the statistical models to true likelihood-ratio values calculated on the basis of the distribution specified for a simulated population.

show abstract

Effects of telephone transmission on the performance of formant-trajectory-based forensic voice comparison – Female voices

Cited by 26 publications

References 43 publications

Three steps forward for predictability. Consideration of methodological robustness, indexical and prosodic factors, and replication in the laboratory

Three steps forward for predictability. Consideration of methodological robustness, indexical and prosodic factors, and replication in the laboratory

Speaker identification based on artificial neural networks. Case study: the Polish vowel (pilot study)

Likelihood Ratio Calculation in Acoustic-Phonetic Forensic Voice Comparison: Comparison of Three Statistical Modelling Approaches

Contact Info

Product

Resources

About