SELM: Speech Enhancement Using Discrete Tokens And Language Models

Abstract

Language models (LMs) have shown superior performances in various speech generation tasks recently, demonstrating their powerful ability for semantic context modeling. Given the intrinsic similarity between speech generation and speech enhancement, harnessing semantic information holds potential advantages for speech enhancement tasks. In light of this, we propose SELM, a novel paradigm for speech enhancement, which integrates discrete tokens and leverages language models. SELM comprises three stages: encoding, modeling, and decoding. We transform continuous waveform signals into discrete tokens using pre-trained self-supervised learning (SSL) models and a k-means tokenizer. Language models then capture comprehensive contextual information within these tokens. Finally, a detokenizer and HiFi-GAN restore them into enhanced speech. Experimental results demonstrate that SELM achieves comparable performance in objective metrics alongside superior results in subjective perception. Our demos

Stats

• citations30
• S2 citations—
• github stars0
• HF likes0
• heat score11.19
• arxiv keywang2023selm

Related papers

• TSELM: Target Speaker Extraction Using Discrete Tokens And Language Models (2024)0.00
• SELMA: A Speech-enabled Language Model For Virtual Assistant Interactions (2025)2.26
• Gense: Generative Speech Enhancement Via Language Models Using Hierarchical Modeling (2025)0.00
• Audiolm: A Language Modeling Approach To Audio Generation (2022)18.91
• Aligning Generative Speech Enhancement With Perceptual Feedback (2025)0.00
• Discrete Multimodal Transformers With A Pretrained Large Language Model For Mixed-supervision Speech Processing (2024)0.00
• LM-VC: Zero-shot Voice Conversion Via Speech Generation Based On Language Models (2023)0.00
• Speechlm: Enhanced Speech Pre-training With Unpaired Textual Data (2022)0.00