Forensic Speech Science

The purpose of this paper is to preliminarily present the status of the paradigm shift or new paradigm in forensic voice comparison in Peru. First, the ambiguous terminology (such as forensic acoustics, voice identification or forensic phonoaudiology as forensic voice comparison) used to avoid its inappropriate use in academic and judicial situations is reviewed. Second, forensic phonetics and its areas of practice are presented in general terms. Third, the methodological approaches in voice comparison and the way of expressing the results are presented and it is explained why the forensic expert in voice comparison should opt for the analysis and results through the likelihood ratio. Fourth, the basis of the paradigm shift and its influence on forensic voice comparison is discussed in depth. Fifth, the current state of research in forensic voice comparison in Peru is presented. As a result of this research, it can be argued that the current studies demonstrate a transition towards a paradigm shift, especially through the use of automated programs.

Refining the relevant population in forensic voice comparison – A response to Hicks et alii (2015) The importance of distinguishing information from evidence/observations when formulating propositions

Language mismatching is considered one of the biggest challenges in achieving adequate speaker verification. The number of bilingual speakers worldwide is increasing, making speaker verification for speech technology more challenging. The main objective of this study is to examine the effect of language mismatch between training and test conditions on the performance of the speaker verification model, with a specific focus on children's speech by utilizing pre-trained and fine-tuned deep speaker embedding models and by investigating two child speech datasets (Samromur and kidsTALC). This work utilises two time-delay neural networks (TDNN): X-Vector and Emphasized Channel Attention, Propagation and Aggregation (ECAPA) to extract the embedding features of speech samples. For evaluation of the performance of speaker verification, we used the likelihood-ratio framework by using the likelihood-ratio score calculation method based on children’s voices and employing measures such as log-likelihood ratio cost (Cllr) and equal error rate (EER). The experimental results indicate that the language variety between training and testing utterances significantly degrades speaker verification performance compared to multi-language training; however, fine-tuning still performs better than pretrained models by 11.5% in the best performance of ECAPA-TDNN.

In forensic voice comparison, deep learning has become widely popular recently. It is mainly used to learn speaker representations, called embeddings or embedding vectors. Speaker embeddings are often trained using corpora mostly containing widely spoken languages. Thus, language dependency is an important factor in automatic forensic voice comparison, especially when the target language is linguistically very different from that the model is trained on. In the case of a low-resource language, developing a corpus for forensic purposes containing enough speakers to train deep learning models is costly. This study aims to investigate whether a model pre-trained on multilingual (mostly English) corpus can be used on a target low-resource language (here, Hungarian), not represented by the model. Often multiple samples are not available from the offender (unknown speaker). Samples are therefore compared pairwise with and without speaker enrollment for suspect (known) speakers. Two corpora are used that were developed especially for forensic purposes and a third that is meant for traditional speaker verification. Speaker embedding vectors are extracted by the x-vector and ECAPA-TDNN techniques. Speaker verification was evaluated in the likelihood-ratio framework. A comparison is made between the language combinations (modeling, LR calibration, and evaluation). The results were evaluated by Cllrmin and EER metrics. It was found that the model pre-trained on a different language but on a corpus with a significant number of speakers can be used on samples with language mismatch. Sample duration and speaking style also seem to affect the performance.

Forensic voice comparison (FVC) is gaining global recognition as a scientific method for speaker identification in legal proceedings. In Indonesia, however, the application of FVC remains underdeveloped, despite increasing reliance on digital and audio based evidence in criminal cases. This study presents a legally and methodologically robust framework for implementing FVC within the Indonesian judicial context. The research integrates Indonesian legal standards (KUHAP Articles 183–184, UU ITE) with international forensic protocols, including ISO/IEC 27037 for digital evidence handling. A combination of acoustic features MFCCs, F0, formants, and VTLN was extracted from disputed and reference voice samples. Likelihood ratios (LRs) were calculated using Gaussian Mixture Models, with score calibration via logistic regression. Results showed mean log LR values of 2.1 for genuine trials and –1.8 for impostor trials, with an ROC AUC of 0.91. Visual tools, including Tippett plots and ROC curves, were used to interpret and communicate evidence reliability. The findings confirm that calibrated, probabilistic FVC methods are feasible and legally admissible in Indonesia. However, challenges remain in handling low quality recordings, maintaining chain of custody, and bridging communication gaps between scientific experts and legal practitioners. The study recommends structured training, standardized protocols, and the use of visual aids to enhance evidentiary transparency. This framework lays the foundation for a scalable, court ready FVC system aligned with national law and global best practices. It supports interdisciplinary cooperation aimed at strengthening Indonesia’s forensic infrastructure.

Use of relevant data, quantitative measurements, and statistical models to calculate a likelihood ratio for a Chinese forensic voice comparison case involving two sisters

In this presentation I argue that one should not attempt to directly assess whether a forensic analysis technique is scientifically acceptable. Rather one should first specify what one considers to be appropriate principles governing acceptable practice, then consider any particular approach in light of those principles. I focus on one principle: The validity and reliability of an approach should be empirically tested under conditions reflecting those of the case under investigation using test data drawn from the relevant population. Versions of this principle have been key elements in several reports on forensic science, including forensic voice comparison, published over the last four-and-a-half decades. I consider the aural-spectrographic approach to forensic voice comparison (also known as "voiceprint" or "voicegram" examination) in light of this principle, and also the currently widely practiced auditory-acoustic-phonetic approach (these two approaches do not appear to be mutually exclusive). Finally, I challenge the audience members to consider what each of them thinks constitutes the relevant principles regarding acceptable practice, and then consider their own approach to forensic-acoustic analysis in light of those principles.

Multi-laboratory evaluation of forensic voice comparison systems under conditions reflecting those of a real forensic case (forensic_eval_01) – Introduction

Voice biometrics are being implemented globally in large scale applications such as remote banking, government e-services, transportation and building security access, autonomous vehicles, and healthcare. They have been integrated in numerous apps, often coupled with face biometrics and artificial intelligence methods. Voice biometrics products and solutions must meet three key re-quirements for the success in their deployment: they must be highly trustable regarding privacy protection; easy to use and always be available. This edited book presents the state of the art in voice biometrics research and technologies including implementation and deployment challenges in terms of interoperability, scalability and performance, and security. The team of editors and chapter authors combine a wealth of expertise from academia and the industry. Topics covered include the fundamentals of voice biometrics; design of countermeasures for replay attack; attacker's perspective for voice biometrics; voice biometrics; speaker de-identification; performance evaluation of voice biometrics solutions; standardization of voice biometrics technology; industry perspectives; joining forces of voice and facial biometrics; and future trends and challenges in voice biometrics. Providing comprehensive coverage of the field of voice biometrics, this authoritative volume will be of great interest to researchers, scientists, engineers, practitioners and advanced students involved in the fields of security, biometrics, forensic sciences, human computer interaction, speech processing, acoustics, multimedia, pattern recognition, and privacy-preserving, digital signal processing and speech technologies. It will also be of interest to researchers and professionals working in law and criminology.

Mismatched conditions between the recordings of suspect, offender and relevant background population represent a typical scenario in real forensic casework. In this paper, we investigate the impact of mismatch conditions associated with mobile phone speech recordings on forensic voice comparison (FVC). The two major mobile phone technologies currently in use today are the Global System for Mobile Communications (GSM) and Code Division Multiple Access (CDMA). These are fundamentally different in the way in which they handle the speech signal, which in turn will lead to significant mismatch between speech recordings. Our results suggest that the resulting degradation on the accuracy of a FVC analysis can be very significant (as high as 150%). Surprisingly, though, our results also suggest that the reliability of a FVC analysis may actually improve. We propose a strategy for lessening this impact by passing the suspect speech data through the GSM or CDMA codecs, depending on the network origin of the offender data, prior to the FVC analysis. Though this goes a long way to mitigating the impact (a reduction in loss of accuracy from 150% to 80%), it is still not as good as analysis under matched conditions.

Forensic phonetic features have been being the most commonly used features in forensic voice comparison technology, and few dispute that they are dramatically affected by real-world conditions, which leads to within-speaker variations and consequently reduces the reliability of forensic voice comparison results as evidence. In this study two experiments are designed and the conversations recorded in pure real-world conditions are adopted as experiment materials to respectively demonstrate the effects of real-world conditions on quantitative phonetic feature of formant trajectory and qualitative phonetic feature of formants pattern that have been tested valid in researches or practice. The results of experiments vividly display that pure real-world conditions do reduce both the amount and quality of available quantitative and qualitative phonetic features, which consequently makes forensic voice comparison impossible or forensic voice comparison results less valid and reliable.

This book is the 11th volume in a series that comprises the Benchmark Papers in Acoustics. The present volume is a collection of 39 previously published papers and one original contribution. Of the previously published papers, about half have ap peared in the<i>Journal of the Acoustical Society of America</i>. The selected articles represent, in the editor's opinion, the most significant developments in the area of speech intelligibility and speaker recognition. The editor is Mones E. Hawley, chairman, in 1960, of the Committee for the Development of a Standard Method for Measurement of Monosyllabic Word Intelligibility, American National Standards Institute. The volume focuses primarily on the contributions of scientists from universities and laboratories in America. However, articles from speech scientists and engineers in seven other countries are included.

This article investigates the development and evaluation of a speaker identification system using deep learning techniques, with a focus on Convolutional Neural Networks (CNNs) and the audioMNIST dataset. The study reveals significant advancements in speaker identification, demonstrating substantial improvements over state-of-the-art models. Our system achieves high accuracy and reliability in distinguishing speakers, showcasing its potential applications in forensic science, security, and privacy protection. The paper thoroughly examines audio signal representation, preprocessing techniques, and feature extraction methods, highlighting how these components contribute to the system's effectiveness. By leveraging CNNs, the proposed system provides highly accurate speaker identification and exhibits robustness in various conditions, including noise and varying speech patterns. The findings underscore the system’s capability to enhance security measures and forensic research, paving the way for future optimizations and broader applications. This contribution expands the knowledge base in speaker identification technology, offering scalable and efficient solutions for real-world scenarios. Future research directions include refining the dataset, exploring advanced optimization techniques, and addressing ethical considerations to ensure the system's robustness and practical utility in diverse applications.

The National Institute of Standards and Technology (NIST) has coordinated numerous evaluations of speech and language technology over the past two decades. We discuss here evaluations in three general areas of speech processing that have security and forensic applications: large vocabulary conversational speech recognition (LVCSR), speaker recognition, and language recognition.

This article examines the effectiveness of artificial neural networks (ANNs) as forensic voice comparison techniques. This study specifically considers feed-forward multilayer perceptron (MLP) and radial basic function (RBF) network models. Formant frequencies of Polish vowel e (stressed or unstressed) in selected contexts were used as predictors. This has already been confirmed in an earlier investigation that determined that dynamic formant frequencies of vowels are powerful elements in distinguishing the voice. It has been concluded that neural networks might assist in distinguishing speakers from the others with very good accuracy, reaching 100%. MLP models should be given preference. The results of the investigation have shown the influence of vowel e triads on the effectiveness of correct classification rates. In addition, the authors have determined that the accuracy of classification is greater when based on a single context than for similar input data aggregated over several different contexts.