IAFPA is inviting applications for (max. £1,300) research grants.
The current year’s suggested (but by no means mandatory) topics are:
large scale data analysis
validation of methods, and
The deadlines for applications will be as follows:
submission of proposals: Oct 8, 2018
reviews by Research Committee returned to applicants: Oct 22
resubmission of applications (for which resubmission has been requested): Nov 5
notification of awards: by Nov 12
Please make your application using the form provided to members by the secretary. The instructions are given on the form itself. If you have any questions about the application and review process, please do not hesitate to email Vica Papp @ email@example.com
Voxalys AB and
Division of Speech and Language Pathology
Department of Clinical Neuroscience and Rehabilitation
Institute of Neuroscience and Physiology
Sahlgrenska Academy at University of Gothenburg
This study aims at evaluating the effects of recording material consisting of what can be called and described as ‘double filtering’. This filtering effect can here be defined as sound transmitted via GSM communication (first filter), which then passes an indeterminable distance through the air prior to being captured by another recording device, such as a mobile phone or handheld recorder’s microphone second filter). Several cases have been received recently with material either known to have been subdued to this effect or suspected to be. To date and to the authors knowledge there has been little or no focus on analyzing the reliability and effects of this type of recorded material, which is why the aim is to conduct a pilot study where the so called ‘double filtering’ effect is evaluated primarily using Automatic Voice Comparison (AVC). However, a database such as the described can in the future be used to evaluate comparisons made using phonetic and linguistic analysis.
Investigators: Erica Gold and Vincent Hughes
Affiliation: Department of Language and Linguistic Science, The University of York
Building on a pilot study carried out by the applicants (Gold and Hughes 2012), the project sets out to investigate two aspects of correlation between speech parameters. The first involves empirical testing of data from a homogeneous group of speakers (DyViS: Nolan 2009) to reveal correlations that may exist between traditional acoustic-phonetic parameters commonly used in forensic speaker comparisons. Secondly, we aim to address theoretical issues underlying the application of logistic regression fusion (Brümmer et al 2007) in a likelihood ratio (LR) framework, by comparing the levels of correlations found in the data against the levels of correlations found for LRs computed by a given system. The results have two sets of implications. Firstly, the results will provide an empirically-based starting point for making informed decisions concerned with the combination of parameters in real forensic speaker comparisons. This applies both to experts working in a LR framework who must account for naïve Bayes, as well as those working in other frameworks where the expert personally selects parameters to consider and combine in casework. Those making uninformed assumptions in regards to correlations that may exist between parameters used in casework could potentially carry out analyses which lead to miscarriages of justice in the representation of the strength of evidence. Secondly, the results are relevant in addressing how appropriate fusion is as a method for combining dependent parameters using LRs. For those working within a LR framework, the results are intended to provide a basis for the development of a Bayesian network (Taroni et al 2006) to create a ‘front-end’ mathematical model of interdependencies between speech parameters in order to appropriately combine parameters.
Department of Language and Linguistic Science, University of York, UK
This study examines the extent to which socio-indexical information can be carried in a highly degraded signal. It is also concerned with the role prosody plays in the description and identification of varieties. Five varieties of British English are used to test two hypotheses: (1) listeners are able to identify accents of British English based on the prosodic features of speech alone, (2) listeners are better able to identify accents closer to their own than those geographically removed. Samples of speech from speakers of the varieties in question will be low-pass filtered at 350Hz, leaving segmental content absent or unintelligible. A variety of answering options will be used in the experimental design to test the degree of fine-grained resolution listeners are able to demonstrate in the identification task.
By investigating listeners’ ability to identify regional accents in samples of speech in which segmental content is unintelligible, this study investigates the recoverability of socio-indexical information from a highly degraded signal. The study seeks to demonstrate that prosodic characteristics of regional accent groups carry sufficient information to be considered valid cues to speaker origin in their own right.