mlsanityData to 19 Oct 2026

Too Emotional to Be Fake? Intra-Utterance Emotion Shifts Expose Failure Modes in Speech Deepfake Detectors

arXiv 2610.22832 ↗First posted 16 Oct 2026cs.SD cs.CLcited by 0

On the map: Speech and audio language models → Emotional speech synthesis and recognition

Abstract (arXiv)

Expressive speech synthesis can now change emotion within a single utterance, yet its impact on deepfake detection remains unclear. We introduce EmoShift, a benchmark comparing emotion-stable (AA) and emotion-changing (AB) speech to examine these effects while accounting for static emotion identity and acoustic confounds. It comprises 7,240 bona fide utterances from four source datasets and 9,720 synthetic utterances generated using four expressive TTS systems. Across seven detectors, we find that emotion changes can make synthetic speech appear more genuine, increasing spoof miss rates by up to 25.3 percentage points on OMG-Emotion at fixed thresholds. Conversely, emotion-changing bona fide speech increases false-alarm rates by up to 24.4 percentage points on IEMOCAP. Further analyses reveal detector-specific sensitivities to emotion trajectories, not fully explained by static emotion identity or splicing artifacts. These findings underscore the importance of emotion dynamics in deepfake detection.

Key facts. No citing papers in the map yet (data through 19 Oct 2026). Placed in Emotional speech synthesis and recognition, one of 11 directions in Speech and audio language models. Builds on 14 of its 28 references that are in the map. First posted on arXiv 16 Oct 2026.

Builds on

14 of 28 references.

Closest ancestors · 2

Same direction, newest first

  1. 2026Sequential Trajectories and Simultaneous Blending: Multi-Emotion Modeling for Instruction-Following TTS31 Aug
  2. 2026EmoTra-TTS: Smooth Intra-Utterance Emotion Transitions for Speech Synthesis24 Aug

Other references · 12

  1. 2021XLS-R: Self-supervised Cross-lingual Speech Representation Learning at Scalecited by 678
  2. 2020End-to-end anti-spoofing with RawNet2cited by 196
  3. 2020Seen and Unseen emotional style transfer for voice conversion with a new emotional speech datasetcited by 153
  4. 2024ASVspoof 5: Crowdsourced Speech Data, Deepfakes, and Adversarial Attacks at Scalecited by 121
  5. 2024XLSR-Mamba: A Dual-Column Bidirectional State Space Model for Spoofing Attack Detectioncited by 39
  6. 2025Do Audio LLMs Really LISTEN, or Just Transcribe? Measuring Lexical vs. Acoustic Emotion Cues Reliancecited by 34
  7. 2018The OMG-Emotion Behavior Datasetcited by 26
  8. 2026ProSDD: Learning Prosodic Representations for Speech Deepfake Detection against Expressive and Emotional Attackscited by 5
  9. 2025Why Speech Deepfake Detectors Won't Generalize: The Limits of Detection in an Open Worldcited by 2
  10. 2026Alethia: A Foundational Encoder for Voice Deepfakescited by 2
  11. 2026AffectDF: The Most Comprehensive Benchmark for Speech Deepfake Detection against Emotionally Expressive Attackscited by 2
  12. 2026The Last Mile of Deepfake Speech Detection: An Industry-Academia Experience Reportcited by 1

Built on by

Not cited yet.

Similar, not cited

Close in wording, no citation either way

  1. 2025The Affective Bridge: Preserving Speech Representations while Enhancing Deepfake Detection vian emotional Constraints
  2. 2026Eroding Trust in Real Speech: A Large-Scale Study of Human Audio Deepfake Perception
  3. 2026SEmoEdit: Probing and Harnessing the Editability of Pre-trained Speech Flows
  4. 2025Evaluating Emotion Recognition in Spoken Language Models on Emotionally Incongruent Speech
  5. 2026On the Emotion Understanding of Synthesized Speech
  6. 2023EmoSpeech: Guiding FastSpeech2 Towards Emotional Text to Speech
  7. 2026EmoTransCap: Dataset and Pipeline for Emotion Transition-Aware Speech Captioning in Discourses
  8. 2024EmoKnob: Enhance Voice Cloning with Fine-Grained Emotion Control