Human perception of audio deepfakes: the role of language and speaking style

Audio deepfakes have reached a level of realism that makes it increasingly difficult to distinguish between human and artificial voices, which poses risks such as identity theft or spread of disinformation. Despite these concerns, research on humans' ability to identify deepfakes is limited, with most studies focusing on English and very few exploring the reasons behind listeners' perceptual decisions. This study addresses this gap through a perceptual experiment in which 54 listeners (28 native Spanish speakers and 26 native Japanese speakers) classified voices as natural or synthetic, and justified their choices. The experiment included 80 stimuli (50% artificial), organized according to three variables: language (Spanish/Japanese), speech style (audiobooks/interviews), and familiarity with the voice (familiar/unfamiliar). The goal was to examine how these variables influence detection and to analyze qualitatively the reasoning behind listeners' perceptual decisions. Results indicate an average accuracy of 59.11%, with higher performance on authentic samples. Judgments of vocal naturalness rely on a combination of linguistic and non-linguistic cues. Comparing Japanese and Spanish listeners, our qualitative analysis further reveals both shared cues and notable cross-linguistic differences in how listeners conceptualize the "humanness" of speech. Overall, participants relied primarily on suprasegmental and higher-level or extralinguistic characteristics - such as intonation, rhythm, fluency, pauses, speed, breathing, and laughter - over segmental features. These findings underscore the complexity of human perceptual strategies in distinguishing natural from artificial speech and align partly with prior research emphasizing the importance of prosody and phenomena typical of spontaneous speech, such as disfluencies.

From Sharpness to Better Generalization for Speech Deepfake Detection

Generalization remains a critical challenge in speech deepfake detection (SDD). While various approaches aim to improve robustness, generalization is typically assessed through performance metrics like equal error rate without a theoretical framework to explain model performance. This work investigates sharpness as a theoretical proxy for generalization in SDD. We analyze how sharpness responds to domain shifts and find it increases in unseen conditions, indicating higher model sensitivity. Based on this, we apply Sharpness-Aware Minimization (SAM) to reduce sharpness explicitly, leading to better and more stable performance across diverse unseen test sets. Furthermore, correlation analysis confirms a statistically significant relationship between sharpness and generalization in most test settings. These findings suggest that sharpness can serve as a theoretical indicator for generalization in SDD and that sharpness-aware training offers a promising strategy for improving robustness.

Quantifying Source Speaker Leakage in One-to-One Voice Conversion

Using a multi-accented corpus of parallel utterances for use with commercial speech devices, we present a case study to show that it is possible to quantify a degree of confidence about a source speaker's identity in the case of one-to-one voice conversion. Following voice conversion using a HiFi-GAN vocoder, we compare information leakage for a range speaker characteristics; assuming a "worst-case" white-box scenario, we quantify our confidence to perform inference and narrow the pool of likely source speakers, reinforcing the regulatory obligation and moral duty that providers of synthetic voices have to ensure the privacy of their speakers' data.

QualiSpeech: A Speech Quality Assessment Dataset with Natural Language Reasoning and Descriptions

This paper explores a novel perspective to speech quality assessment by leveraging natural language descriptions, offering richer, more nuanced insights than traditional numerical scoring methods. Natural language feedback provides instructive recommendations and detailed evaluations, yet existing datasets lack the comprehensive annotations needed for this approach. To bridge this gap, we introduce QualiSpeech, a comprehensive low-level speech quality assessment dataset encompassing 11 key aspects and detailed natural language comments that include reasoning and contextual insights. Additionally, we propose the QualiSpeech Benchmark to evaluate the low-level speech understanding capabilities of auditory large language models (LLMs). Experimental results demonstrate that finetuned auditory LLMs can reliably generate detailed descriptions of noise and distortion, effectively identifying their types and temporal characteristics. The results further highlight the potential for incorporating reasoning to enhance the accuracy and reliability of quality assessments. The dataset will be released at https://huggingface.co/datasets/tsinghua-ee/QualiSpeech.

Good practices for evaluation of synthesized speech

This document is provided as a guideline for reviewers of papers about speech synthesis. We outline some best practices and common pitfalls for papers about speech synthesis, with a particular focus on evaluation. We also recommend that reviewers check the guidelines for authors written in the paper kit and consider those as reviewing criteria as well. This is intended to be a living document, and it will be updated as we receive comments and feedback from readers. We note that this document is meant to provide guidance only, and that reviewers should ultimately use their own discretion when evaluating papers.

ASVspoof 5: Design, Collection and Validation of Resources for Spoofing, Deepfake, and Adversarial Attack Detection Using Crowdsourced Speech

ASVspoof 5 is the fifth edition in a series of challenges which promote the study of speech spoofing and deepfake attacks as well as the design of detection solutions. We introduce the ASVspoof 5 database which is generated in crowdsourced fashion from data collected in diverse acoustic conditions (cf. studio-quality data for earlier ASVspoof databases) and from ~2,000 speakers (cf. ~100 earlier). The database contains attacks generated with 32 different algorithms, also crowdsourced, and optimised to varying degrees using new surrogate detection models. Among them are attacks generated with a mix of legacy and contemporary text-to-speech synthesis and voice conversion models, in addition to adversarial attacks which are incorporated for the first time. ASVspoof 5 protocols comprise seven speaker-disjoint partitions. They include two distinct partitions for the training of different sets of attack models, two more for the development and evaluation of surrogate detection models, and then three additional partitions which comprise the ASVspoof 5 training, development and evaluation sets. An auxiliary set of data collected from an additional 30k speakers can also be used to train speaker encoders for the implementation of attack algorithms. Also described herein is an experimental validation of the new ASVspoof 5 database using a set of automatic speaker verification and spoof/deepfake baseline detectors. With the exception of protocols and tools for the generation of spoofed/deepfake speech, the resources described in this paper, already used by participants of the ASVspoof 5 challenge in 2024, are now all freely available to the community.

Towards Automated Fact-Checking of Real-World Claims: Exploring Task Formulation and Assessment with LLMs

Fact-checking is necessary to address the increasing volume of misinformation. Traditional fact-checking relies on manual analysis to verify claims, but it is slow and resource-intensive. This study establishes baseline comparisons for Automated Fact-Checking (AFC) using Large Language Models (LLMs) across multiple labeling schemes (binary, three-class, five-class) and extends traditional claim verification by incorporating analysis, verdict classification, and explanation in a structured setup to provide comprehensive justifications for real-world claims. We evaluate Llama-3 models of varying sizes (3B, 8B, 70B) on 17,856 claims collected from PolitiFact (2007-2024) using evidence retrieved via restricted web searches. We utilize TIGERScore as a reference-free evaluation metric to score the justifications. Our results show that larger LLMs consistently outperform smaller LLMs in classification accuracy and justification quality without fine-tuning. We find that smaller LLMs in a one-shot scenario provide comparable task performance to fine-tuned Small Language Models (SLMs) with large context sizes, while larger LLMs consistently surpass them. Evidence integration improves performance across all models, with larger LLMs benefiting most. Distinguishing between nuanced labels remains challenging, emphasizing the need for further exploration of labeling schemes and alignment with evidences. Our findings demonstrate the potential of retrieval-augmented AFC with LLMs.

Exploring Active Data Selection Strategies for Continuous Training in Deepfake Detection

In deepfake detection, it is essential to maintain high performance by adjusting the parameters of the detector as new deepfake methods emerge. In this paper, we propose a method to automatically and actively select the small amount of additional data required for the continuous training of deepfake detection models in situations where deepfake detection models are regularly updated. The proposed method automatically selects new training data from a \textit{redundant} pool set containing a large number of images generated by new deepfake methods and real images, using the confidence score of the deepfake detection model as a metric. Experimental results show that the deepfake detection model, continuously trained with a small amount of additional data automatically selected and added to the original training set, significantly and efficiently improved the detection performance, achieving an EER of 2.5% with only 15% of the amount of data in the pool set.

Towards An Integrated Approach for Expressive Piano Performance Synthesis from Music Scores

This paper presents an integrated system that transforms symbolic music scores into expressive piano performance audio. By combining a Transformer-based Expressive Performance Rendering (EPR) model with a fine-tuned neural MIDI synthesiser, our approach directly generates expressive audio performances from score inputs. To the best of our knowledge, this is the first system to offer a streamlined method for converting score MIDI files lacking expression control into rich, expressive piano performances. We conducted experiments using subsets of the ATEPP dataset, evaluating the system with both objective metrics and subjective listening tests. Our system not only accurately reconstructs human-like expressiveness, but also captures the acoustic ambience of environments such as concert halls and recording studios. Additionally, the proposed system demonstrates its ability to achieve musical expressiveness while ensuring good audio quality in its outputs.

Explaining Speaker and Spoof Embeddings via Probing

This study investigates the explainability of embedding representations, specifically those used in modern audio spoofing detection systems based on deep neural networks, known as spoof embeddings. Building on established work in speaker embedding explainability, we examine how well these spoof embeddings capture speaker-related information. We train simple neural classifiers using either speaker or spoof embeddings as input, with speaker-related attributes as target labels. These attributes are categorized into two groups: metadata-based traits (e.g., gender, age) and acoustic traits (e.g., fundamental frequency, speaking rate). Our experiments on the ASVspoof 2019 LA evaluation set demonstrate that spoof embeddings preserve several key traits, including gender, speaking rate, F0, and duration. Further analysis of gender and speaking rate indicates that the spoofing detector partially preserves these traits, potentially to ensure the decision process remains robust against them.