---
title: "Johns Hopkins' ProSDD Catches Emotional Voice Deepfakes Where Others Fail"
subtitle: "A two-stage training recipe teaches deepfake detectors to model natural prosody first, cutting error rates by more than half on emotional spoofing attacks."
url: "https://alphasignal.ai/news/johns-hopkins-prosdd-catches-emotional-voice-deepfakes-where-others-fail"
author: "AlphaSignal Newsroom"
author_url: "https://alphasignal.ai/editorial-team"
company: "AlphaSignal"
company_url: "https://alphasignal.ai/company/alphasignal"
published: "2026-04-14T00:00:00.000Z"
updated: "2026-08-23T21:33:19.818Z"
type: "Paper"
topics:
  - "Audio"
  - "Security"
subtopics:
  - "Speech To Text"
  - "Red Teaming"
upvotes: 1890
source_url: "https://arxiv.org/abs/2604.13229"
comments: 0
truncated: true
---

# Johns Hopkins' ProSDD Catches Emotional Voice Deepfakes Where Others Fail

> A two-stage training recipe teaches deepfake detectors to model natural prosody first, cutting error rates by more than half on emotional spoofing attacks.

Speech deepfake detectors have a dirty little secret: they crush the benchmarks they were trained on, then fall apart the moment a synthesized voice starts to sound emotional. A new paper from Johns Hopkins and Hong Kong Polytechnic, [ProSDD](https://arxiv.org/abs/2604.13229), argues the fix is to stop obsessing over spoof artifacts and start teaching the model what real speech actually sounds like.

## Why current detectors keep getting fooled

Speech deepfake detection systems perform well on standard benchmarks but often fail to generalize against expressive and emotional spoofing attacks. Many methods rely on spoof-heavy training data, learning dataset-specific artifacts rather than transferable cues of natural speech. That works fine until a TTS system with laughs, whispers, or emotional inflection shows up, at which point the classifier has no idea what to do.

The authors flag two structural problems with the standard recipe. Fine-tuning a self-supervised backbone with only a spoof classification loss biases it toward whatever quirks live in the training set. And modern TTS and voice conversion systems still leak subtle prosodic inconsistencies, yet no one is training detectors to notice them. Humans do the opposite: they internalize what natural speech sounds like and flag anything that deviates.

_Preview only. Full text requires an AlphaSignal Pro account._

## Sources

- [ProSDD: Learning Prosodic Representations for Speech Deepfake Detection](https://arxiv.org/abs/2604.13229)
- [ProSDD full paper (HTML)](https://arxiv.org/html/2604.13229)
- [ProSDD project website](https://prosdd.github.io/ProSDD_website/)
- [Fraunhofer AISEC Related Work catalog](https://deepfake-demo.aisec.fraunhofer.de/related_work/)