# Adapting Speech Language Model to Singing Voice Synthesis

arXiv: https://arxiv.org/abs/2512.14657 · HTML: https://mlsanity.com/paper/2512.14657/

## Key facts

- No citing papers in the map yet (data through 19 Oct 2026).
- Placed in Vocoders and speech synthesis, one of 11 directions in Speech and audio language models.
- Builds on 22 of its 27 references that are in the map.
- First posted on arXiv 16 Dec 2025.

## Placed in

- Speech and audio language models / [Vocoders and speech synthesis](https://mlsanity.com/speech/vocoders-and-speech-synthesis/)

## Builds on (most cited first, 22 of 22)

- [Density estimation using Real NVP](https://mlsanity.com/paper/1605.08803/) (2016)
- [Glow: Generative Flow with Invertible 1x1 Convolutions](https://mlsanity.com/paper/1807.03039/) (2018)
- [HiFi-GAN: Generative Adversarial Networks for Efficient and High Fidelity Speech Synthesis](https://mlsanity.com/paper/2010.05646/) (2020)
- [Neural Codec Language Models are Zero-Shot Text to Speech Synthesizers](https://mlsanity.com/paper/2301.02111/) (2023)
- [Qwen2-Audio Technical Report](https://mlsanity.com/paper/2407.10759/) (2024)
- [Phi-4-Mini Technical Report: Compact yet Powerful Multimodal Language Models via Mixture-of-LoRAs](https://mlsanity.com/paper/2503.01743/) (2025)
- [Grad-TTS: A Diffusion Probabilistic Model for Text-to-Speech](https://mlsanity.com/paper/2105.06337/) (2021)
- [WaveGlow: A Flow-based Generative Network for Speech Synthesis](https://mlsanity.com/paper/1811.00002/) (2018)
- [Kimi-Audio Technical Report](https://mlsanity.com/paper/2504.18425/) (2025)
- [Speak, Read and Prompt: High-Fidelity Text-to-Speech with Minimal Supervision](https://mlsanity.com/paper/2302.03540/) (2023)
- [ESPnet-TTS: Unified, Reproducible, and Integratable Open Source End-to-End Text-to-Speech Toolkit](https://mlsanity.com/paper/1910.10909/) (2019)
- [ESPnet-Codec: Comprehensive Training and Evaluation of Neural Codecs for Audio, Music, and Speech](https://mlsanity.com/paper/2409.15897/) (2024)
- [XiaoiceSing: A High-Quality and Integrated Singing Voice Synthesis System](https://mlsanity.com/paper/2006.06261/) (2020)
- [MOS-Bench: Benchmarking Generalization Abilities of Subjective Speech Quality Assessment Models](https://mlsanity.com/paper/2411.03715/) (2024)
- [SingMOS: An extensive Open-Source Singing Voice Dataset for MOS Prediction](https://mlsanity.com/paper/2406.10911/) (2024)
- [ESPnet-SpeechLM: An Open Speech Language Model Toolkit](https://mlsanity.com/paper/2502.15218/) (2025)
- [Prompt-Singer: Controllable Singing-Voice-Synthesis with Natural Language Prompt](https://mlsanity.com/paper/2403.11780/) (2024)
- [OpusLM: A Family of Open Unified Speech Language Models](https://mlsanity.com/paper/2506.17611/) (2025)
- [Accompanied Singing Voice Synthesis with Fully Text-controlled Melody](https://mlsanity.com/paper/2407.02049/) (2024)
- [InspireMusic: Integrating Super Resolution and Large Language Model for High-Fidelity Long-Form Music Generation](https://mlsanity.com/paper/2503.00084/) (2025)
- [TokSing: Singing Voice Synthesis based on Discrete Tokens](https://mlsanity.com/paper/2406.08416/) (2024)
- [Improved Normalizing Flow-Based Speech Enhancement using an All-pole Gammatone Filterbank for Conditional Input Representation](https://mlsanity.com/paper/2210.11654/) (2022)

## Abstract (arXiv)

Speech Language Models (SLMs) have recently emerged as a unified paradigm for addressing a wide range of speech-related tasks, including text-to-speech (TTS), speech enhancement (SE), and automatic speech recognition (ASR). However, the generalization capability of large-scale pre-trained SLMs remains underexplored. In this work, we adapt a 1.7B parameter TTS pretrained SLM for singing voice synthesis (SVS), using only a 135-hour synthetic singing corpus, ACE-Opencpop. Building upon the ESPNet-SpeechLM, our recipe involves the following procedure: (1) tokenization of music score conditions and singing waveforms, (2) multi-stream language model token prediction, (3) conditional flow matching-based mel-spectrogram generation. (4) a mel-to-wave vocoder. Experimental results demonstrate that our adapted SLM generalizes well to SVS and achieves performance comparable to leading discrete token-based SVS models.