Accounting for correlation in linguistic-acoustic likelihood ratio-based forensic speaker discrimination

Phil Rose

doi:10.1109/ODYSSEY.2006.248095

Accounting for correlation in linguistic-acoustic likelihood ratio-based forensic speaker discrimination

Phil Rose^*

^*Corresponding author for this work

Research output: Chapter in Book/Report/Conference proceeding › Conference Paper › peer-review

15 Citations (Scopus)

Abstract

The necessity of taking correlation between variables into account when estimating strength of forensic speaker recognition evidence is argued for. A modest forensic speaker discrimination experiment is described which investigates how well non-contemporaneous speech samples from the same speaker can be discriminated from different-speaker samples using bivariate kernel density likelihood ratios from F2 and F3 of the five monophthongal phonemes of General Australian English, spoken by 11 males. The experiment shows that an approach which takes the correlation of variables into account can yield useful strengths of evidence. It is also pointed out that the results of such tests still require evaluation with the appropriate confidence limits.

Original language	English
Title of host publication	IEEE Odyssey 2006
Subtitle of host publication	Workshop on Speaker and Language Recognition
DOIs	https://doi.org/10.1109/ODYSSEY.2006.248095
Publication status	Published - 2006
Event	IEEE Odyssey 2006: Workshop on Speaker and Language Recognition - San Juan, Puerto Rico Duration: 28 Jun 2006 → 30 Jun 2006

Publication series

Name	IEEE Odyssey 2006: Workshop on Speaker and Language Recognition

Conference

Conference	IEEE Odyssey 2006: Workshop on Speaker and Language Recognition
Country/Territory	Puerto Rico
City	San Juan
Period	28/06/06 → 30/06/06

Access to Document

10.1109/ODYSSEY.2006.248095

Cite this

@inproceedings{9e9a77b887724b1a80ebd8f17d393c7a,

title = "Accounting for correlation in linguistic-acoustic likelihood ratio-based forensic speaker discrimination",

abstract = "The necessity of taking correlation between variables into account when estimating strength of forensic speaker recognition evidence is argued for. A modest forensic speaker discrimination experiment is described which investigates how well non-contemporaneous speech samples from the same speaker can be discriminated from different-speaker samples using bivariate kernel density likelihood ratios from F2 and F3 of the five monophthongal phonemes of General Australian English, spoken by 11 males. The experiment shows that an approach which takes the correlation of variables into account can yield useful strengths of evidence. It is also pointed out that the results of such tests still require evaluation with the appropriate confidence limits.",

author = "Phil Rose",

year = "2006",

doi = "10.1109/ODYSSEY.2006.248095",

language = "English",

isbn = "142440472X",

series = "IEEE Odyssey 2006: Workshop on Speaker and Language Recognition",

booktitle = "IEEE Odyssey 2006",

note = "IEEE Odyssey 2006: Workshop on Speaker and Language Recognition ; Conference date: 28-06-2006 Through 30-06-2006",

}

Rose, P 2006, Accounting for correlation in linguistic-acoustic likelihood ratio-based forensic speaker discrimination. in IEEE Odyssey 2006: Workshop on Speaker and Language Recognition., 4013512, IEEE Odyssey 2006: Workshop on Speaker and Language Recognition, IEEE Odyssey 2006: Workshop on Speaker and Language Recognition, San Juan, Puerto Rico, 28/06/06. https://doi.org/10.1109/ODYSSEY.2006.248095

TY - GEN

T1 - Accounting for correlation in linguistic-acoustic likelihood ratio-based forensic speaker discrimination

AU - Rose, Phil

PY - 2006

Y1 - 2006

N2 - The necessity of taking correlation between variables into account when estimating strength of forensic speaker recognition evidence is argued for. A modest forensic speaker discrimination experiment is described which investigates how well non-contemporaneous speech samples from the same speaker can be discriminated from different-speaker samples using bivariate kernel density likelihood ratios from F2 and F3 of the five monophthongal phonemes of General Australian English, spoken by 11 males. The experiment shows that an approach which takes the correlation of variables into account can yield useful strengths of evidence. It is also pointed out that the results of such tests still require evaluation with the appropriate confidence limits.

AB - The necessity of taking correlation between variables into account when estimating strength of forensic speaker recognition evidence is argued for. A modest forensic speaker discrimination experiment is described which investigates how well non-contemporaneous speech samples from the same speaker can be discriminated from different-speaker samples using bivariate kernel density likelihood ratios from F2 and F3 of the five monophthongal phonemes of General Australian English, spoken by 11 males. The experiment shows that an approach which takes the correlation of variables into account can yield useful strengths of evidence. It is also pointed out that the results of such tests still require evaluation with the appropriate confidence limits.

UR - https://www.scopus.com/pages/publications/37649033138

U2 - 10.1109/ODYSSEY.2006.248095

DO - 10.1109/ODYSSEY.2006.248095

M3 - Conference Paper

SN - 142440472X

SN - 9781424404728

T3 - IEEE Odyssey 2006: Workshop on Speaker and Language Recognition

BT - IEEE Odyssey 2006

T2 - IEEE Odyssey 2006: Workshop on Speaker and Language Recognition

Y2 - 28 June 2006 through 30 June 2006

ER -

Accounting for correlation in linguistic-acoustic likelihood ratio-based forensic speaker discrimination

Abstract

Publication series

Conference

Access to Document

Other files and links

Fingerprint

Cite this