Research

Paper

AI LLM March 30, 2026

ParaSpeechCLAP: A Dual-Encoder Speech-Text Model for Rich Stylistic Language-Audio Pretraining

Authors

Anuj Diwan, Eunsol Choi, David Harwath

Abstract

We introduce ParaSpeechCLAP, a dual-encoder contrastive model that maps speech and text style captions into a common embedding space, supporting a wide range of intrinsic (speaker-level) and situational (utterance-level) descriptors (such as pitch, texture and emotion) far beyond the narrow set handled by existing models. We train specialized ParaSpeechCLAP-Intrinsic and ParaSpeechCLAP-Situational models alongside a unified ParaSpeechCLAP-Combined model, finding that specialization yields stronger performance on individual style dimensions while the unified model excels on compositional evaluation. We further show that ParaSpeechCLAP-Intrinsic benefits from an additional classification loss and class-balanced training. We demonstrate our models' performance on style caption retrieval, speech attribute classification and as an inference-time reward model that improves style-prompted TTS without additional training. ParaSpeechCLAP outperforms baselines on most metrics across all three applications. Our models and code are released at https://github.com/ajd12342/paraspeechclap .

Metadata

arXiv ID: 2603.28737
Provider: ARXIV
Primary Category: eess.AS
Published: 2026-03-30
Fetched: 2026-03-31 06:05

Related papers

Raw Data (Debug)
{
  "raw_xml": "<entry>\n    <id>http://arxiv.org/abs/2603.28737v1</id>\n    <title>ParaSpeechCLAP: A Dual-Encoder Speech-Text Model for Rich Stylistic Language-Audio Pretraining</title>\n    <updated>2026-03-30T17:50:07Z</updated>\n    <link href='https://arxiv.org/abs/2603.28737v1' rel='alternate' type='text/html'/>\n    <link href='https://arxiv.org/pdf/2603.28737v1' rel='related' title='pdf' type='application/pdf'/>\n    <summary>We introduce ParaSpeechCLAP, a dual-encoder contrastive model that maps speech and text style captions into a common embedding space, supporting a wide range of intrinsic (speaker-level) and situational (utterance-level) descriptors (such as pitch, texture and emotion) far beyond the narrow set handled by existing models. We train specialized ParaSpeechCLAP-Intrinsic and ParaSpeechCLAP-Situational models alongside a unified ParaSpeechCLAP-Combined model, finding that specialization yields stronger performance on individual style dimensions while the unified model excels on compositional evaluation. We further show that ParaSpeechCLAP-Intrinsic benefits from an additional classification loss and class-balanced training. We demonstrate our models' performance on style caption retrieval, speech attribute classification and as an inference-time reward model that improves style-prompted TTS without additional training. ParaSpeechCLAP outperforms baselines on most metrics across all three applications. Our models and code are released at https://github.com/ajd12342/paraspeechclap .</summary>\n    <category scheme='http://arxiv.org/schemas/atom' term='eess.AS'/>\n    <category scheme='http://arxiv.org/schemas/atom' term='cs.AI'/>\n    <category scheme='http://arxiv.org/schemas/atom' term='cs.CL'/>\n    <category scheme='http://arxiv.org/schemas/atom' term='cs.SD'/>\n    <published>2026-03-30T17:50:07Z</published>\n    <arxiv:comment>Under review</arxiv:comment>\n    <arxiv:primary_category term='eess.AS'/>\n    <author>\n      <name>Anuj Diwan</name>\n    </author>\n    <author>\n      <name>Eunsol Choi</name>\n    </author>\n    <author>\n      <name>David Harwath</name>\n    </author>\n  </entry>"
}