EmoNet-Voice: A Fine-Grained, Expert-Verified Benchmark for Speech Emotion Detection
Fuente:
arXiv
Guardado en:
| Autores principales: | , , , , , , , , |
|---|---|
| Formato: | Preprint |
| Publicado: |
2025
|
| Materias: | |
| Acceso en línea: | |
| Etiquetas: |
Agregar Etiqueta
Sin Etiquetas, Sea el primero en etiquetar este registro!
|
| _version_ | 1866909981464330240 |
|---|---|
| author | Schuhmann, Christoph Kaczmarczyk, Robert Rabby, Gollam Friedrich, Felix Kraus, Maurice Nadi, Kourosh Nguyen, Huu Kersting, Kristian Auer, Sören |
| author_facet | Schuhmann, Christoph Kaczmarczyk, Robert Rabby, Gollam Friedrich, Felix Kraus, Maurice Nadi, Kourosh Nguyen, Huu Kersting, Kristian Auer, Sören |
| contents | Speech emotion recognition (SER) systems are constrained by existing datasets that typically cover only 6-10 basic emotions, lack scale and diversity, and face ethical challenges when collecting sensitive emotional states. We introduce EMONET-VOICE, a comprehensive resource addressing these limitations through two components: (1) EmoNet-Voice Big, a 5,000-hour multilingual pre-training dataset spanning 40 fine-grained emotion categories across 11 voices and 4 languages, and (2) EmoNet-Voice Bench, a rigorously validated benchmark of 4,7k samples with unanimous expert consensus on emotion presence and intensity levels. Using state-of-the-art synthetic voice generation, our privacy-preserving approach enables ethical inclusion of sensitive emotions (e.g., pain, shame) while maintaining controlled experimental conditions. Each sample underwent validation by three psychology experts. We demonstrate that our Empathic Insight models trained on our synthetic data achieve strong real-world dataset generalization, as tested on EmoDB and RAVDESS. Furthermore, our comprehensive evaluation reveals that while high-arousal emotions (e.g., anger: 95% accuracy) are readily detected, the benchmark successfully exposes the difficulty of distinguishing perceptually similar emotions (e.g., sadness vs. distress: 63% discrimination), providing quantifiable metrics for advancing nuanced emotion AI. EMONET-VOICE establishes a new paradigm for large-scale, ethically-sourced, fine-grained SER research. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2506_09827 |
| institution | arXiv |
| publishDate | 2025 |
| record_format | arxiv |
| spellingShingle | EmoNet-Voice: A Fine-Grained, Expert-Verified Benchmark for Speech Emotion Detection Schuhmann, Christoph Kaczmarczyk, Robert Rabby, Gollam Friedrich, Felix Kraus, Maurice Nadi, Kourosh Nguyen, Huu Kersting, Kristian Auer, Sören Computation and Language Artificial Intelligence Speech emotion recognition (SER) systems are constrained by existing datasets that typically cover only 6-10 basic emotions, lack scale and diversity, and face ethical challenges when collecting sensitive emotional states. We introduce EMONET-VOICE, a comprehensive resource addressing these limitations through two components: (1) EmoNet-Voice Big, a 5,000-hour multilingual pre-training dataset spanning 40 fine-grained emotion categories across 11 voices and 4 languages, and (2) EmoNet-Voice Bench, a rigorously validated benchmark of 4,7k samples with unanimous expert consensus on emotion presence and intensity levels. Using state-of-the-art synthetic voice generation, our privacy-preserving approach enables ethical inclusion of sensitive emotions (e.g., pain, shame) while maintaining controlled experimental conditions. Each sample underwent validation by three psychology experts. We demonstrate that our Empathic Insight models trained on our synthetic data achieve strong real-world dataset generalization, as tested on EmoDB and RAVDESS. Furthermore, our comprehensive evaluation reveals that while high-arousal emotions (e.g., anger: 95% accuracy) are readily detected, the benchmark successfully exposes the difficulty of distinguishing perceptually similar emotions (e.g., sadness vs. distress: 63% discrimination), providing quantifiable metrics for advancing nuanced emotion AI. EMONET-VOICE establishes a new paradigm for large-scale, ethically-sourced, fine-grained SER research. |
| title | EmoNet-Voice: A Fine-Grained, Expert-Verified Benchmark for Speech Emotion Detection |
| topic | Computation and Language Artificial Intelligence |
| url | https://arxiv.org/abs/2506.09827 |