---
title: "BatonVoice — paper by Wang et al. (2026) — summary and ideas"
description: "Current LLM-based text-to-speech systems mostly use the LLM as a backbone for generating speech tokens. 38 references."
authors: ["Yue Wang", "Ruotian Ma", "Xingyu Chen", "Zhengliang Shi", "Morunliu Yang", "Wanshun Chen", "Huang Liu", "Jiadi Yao", "Xin He", "Qu Yang", "Qingxuan Jiang", "Fanghua Ye", "Juntao Li", "Zhaopeng Tu", "Xiaolong Li", "Liefeng Bo", "Min Zhang"]
year: 2026
venue: "Proceedings of the 64th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)"
source: "https://aclanthology.org/2026.acl-long.2165/"
doi: "10.18653/v1/2026.acl-long.2165"
licence: "CC BY 4.0 (ACL Anthology)"
licence_url: "https://creativecommons.org/licenses/by/4.0/"
summary_licence: "Elamur Cosmic License"
canonical: "https://research.elamur.ai/free/en/papers/wang-2026-batonvoice-operationalist-framework-enhancing-controllable/"
---
[Elamur Research](https://md.research.elamur.ai/free/en/) · [this page in HTML](https://research.elamur.ai/free/en/papers/wang-2026-batonvoice-operationalist-framework-enhancing-controllable/)

[home](https://md.research.elamur.ai/free/en/) / [papers](https://md.research.elamur.ai/free/en/papers/) / [2026](https://md.research.elamur.ai/free/en/papers/2026/page/23/) / BatonVoice: An Operationalist Framework for Enhancing Controllable Speech Synthesis with Linguistic Intelligence from LLMs

# BatonVoice: An Operationalist Framework for Enhancing Controllable Speech Synthesis with Linguistic Intelligence from LLMs

**Authors:** [Yue Wang](https://md.research.elamur.ai/free/en/authors/yue-wang/), [Ruotian Ma](https://md.research.elamur.ai/free/en/authors/ruotian-ma/), [Xingyu Chen](https://md.research.elamur.ai/free/en/authors/xingyu-chen/), [Zhengliang Shi](https://md.research.elamur.ai/free/en/authors/zhengliang-shi/), Morunliu Yang, Wanshun Chen, Huang Liu, Jiadi Yao, Xin He, [Qu Yang](https://md.research.elamur.ai/free/en/authors/qu-yang/), [Qingxuan Jiang](https://md.research.elamur.ai/free/en/authors/qingxuan-jiang/), [Fanghua Ye](https://md.research.elamur.ai/free/en/authors/fanghua-ye/), [Juntao Li](https://md.research.elamur.ai/free/en/authors/juntao-li/), [Zhaopeng Tu](https://md.research.elamur.ai/free/en/authors/zhaopeng-tu/), [Xiaolong Li](https://md.research.elamur.ai/free/en/authors/xiaolong-li/), Liefeng Bo and [Min Zhang](https://md.research.elamur.ai/free/en/authors/min-zhang/) · **Year:** 2026 · **Venue:** [Proceedings of the 64th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)](https://md.research.elamur.ai/free/en/venues/acl/) ([ACL 2026](https://md.research.elamur.ai/free/en/venues/acl/2026/)) · **Publisher:** [Association for Computational Linguistics](https://md.research.elamur.ai/free/en/publishers/association-for-computational-linguistics/) · **Licence:** [CC BY 4.0 (ACL Anthology)](https://creativecommons.org/licenses/by/4.0/)

**Links:** [Publisher's page](https://aclanthology.org/2026.acl-long.2165/) · [PDF](https://aclanthology.org/2026.acl-long.2165.pdf) · [DOI 10.18653/v1/2026.acl-long.2165](https://doi.org/10.18653/v1/2026.acl-long.2165) · [OpenAlex](https://openalex.org/W7166804570)

**Citations:** 0 citations (OpenAlex, 2026-10-14)

## Summary

Current LLM-based text-to-speech systems mostly use the LLM as a backbone for generating speech tokens. Controllable synthesis then depends on costly, inconsistent human annotation of speech with style labels, and the model's instruction-following ability goes largely unused. Inspired by operationalism, the authors split the task in two. In BatonVoice, a capable text LLM (the "conductor", Gemini 2.5 Pro at inference) turns a user instruction into an explicit textual vocal plan. This plan is a JSON list of word-level acoustic features such as pitch, energy and spectral centroid. A dedicated model, BatonTTS (the "orchestra"), then renders speech from the transcript plus this plan. BatonTTS is trained in three stages on automatically extracted features, so it needs no manually written instructions. The system beats strong open- and closed-source baselines on emotional speech synthesis, and output quality improves as the conductor LLM gets stronger. It also transfers feature control zero-shot to Chinese, which was absent from feature-control training. The work suggests that expressing other modalities as text can tap LLM reasoning more effectively.

## Key points

- **Vocal plan features:** pitch (mean and slope), energy (RMS and slope) and timbre (spectral centroid). These are extracted per segment after word-level alignment and written out as JSON-like text.
- **Architecture:** BatonTTS uses Qwen3-1.7B or Qwen2.5-0.5B as the LLM backbone, combined with the frozen CosyVoice2 speech decoder (speech token encoder, conditional flow matching, HiFi-GAN vocoder).
- **Training pipeline:**
  - pre-training on text–speech-token pairs;
  - SFT on sequences of transcript, features and speech tokens;
  - preference optimization to sharpen feature control and reduce failure modes.
- **Emotion accuracy:** the 1.7B system reaches 57.6% with Gemini 2.5 Pro as conductor, versus 29.8% when the 1.7B model plans for itself. This supports the claim that stronger linguistic intelligence yields better synthesis.
- **Limitations:**
  - mostly English-focused;
  - extra latency from the two-stage pipeline;
  - high training cost;
  - risk of harmful output.

  The code is released on GitHub (Tencent/digitalhuman, BatonVoice).

*Our summary of the paper, in our own words — not its text. The work belongs to its authors; read it at the links above.*

## Institutions

- [Tencent](https://md.research.elamur.ai/free/en/companies/tencent/) — Company, Shenzhen, China · 134 works · 19,743 citations
- [Soochow University](https://md.research.elamur.ai/free/en/universities/soochow-university/) — University, Suzhou, China · 63 works · 4,507 citations

📚 Shelves (5): [🖥️ Computer Science](https://md.research.elamur.ai/free/en/shelves/computer-science/) · 16,667 works · [🤖 Artificial Intelligence](https://md.research.elamur.ai/free/en/shelves/computer-science/artificial-intelligence/) · 16,194 works · [🎭 Arts and Humanities](https://md.research.elamur.ai/free/en/shelves/arts-and-humanities/) · 16,039 works · [🗣️ Language and Linguistics](https://md.research.elamur.ai/free/en/shelves/arts-and-humanities/language-and-linguistics/) · 15,891 works · [🔊 Signal Processing](https://md.research.elamur.ai/free/en/shelves/computer-science/signal-processing/) · 452 works

🧭 Topics (2): [🧠 Language models & pretraining](https://md.research.elamur.ai/free/en/topics/language-models/) · 3,320 works · [🎙️ Speech & spoken language](https://md.research.elamur.ai/free/en/topics/speech/) · 480 works

## Cite

Cite as: Yue Wang, Ruotian Ma, Xingyu Chen, Zhengliang Shi, Morunliu Yang, Wanshun Chen, Huang Liu, Jiadi Yao, Xin He, Qu Yang, Qingxuan Jiang, Fanghua Ye, Juntao Li, Zhaopeng Tu, Xiaolong Li, Liefeng Bo and Min Zhang. 2026. BatonVoice: An Operationalist Framework for Enhancing Controllable Speech Synthesis with Linguistic Intelligence from LLMs. Proceedings of the 64th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers).

The paper itself, not this page. BibTeX from the library's record of the work.

```bibtex
@inproceedings{wang2026batonvoice,
  title = {{BatonVoice: An Operationalist Framework for Enhancing Controllable Speech Synthesis with Linguistic Intelligence from LLMs}},
  author = {Yue Wang, Ruotian Ma, Xingyu Chen, Zhengliang Shi, Morunliu Yang, Wanshun Chen, Huang Liu, Jiadi Yao, Xin He, Qu Yang, Qingxuan Jiang, Fanghua Ye, Juntao Li, Zhaopeng Tu, Xiaolong Li, Liefeng Bo and Min Zhang},
  booktitle = {Proceedings of the 64th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)},
  year = {2026},
  doi = {10.18653/v1/2026.acl-long.2165},
  url = {https://doi.org/10.18653/v1/2026.acl-long.2165}
}
```

- [← previous Baselines for Detection and Classification of Discourse Presentation in English Narrative](https://md.research.elamur.ai/free/en/papers/polo-2026-baselines-detection-classification-discourse-presentation/)
- [next → Bayesian Social Deduction with Graph-Informed Language Models](https://md.research.elamur.ai/free/en/papers/rahimirad-2026-bayesian-social-deduction-graph-informed/)
- [↑ back to papers of 2026](https://md.research.elamur.ai/free/en/papers/2026/page/23/)

## References

38 references, 7 of them with a card of ours. ↗ leads to the work, «find» searches for one we have no id for yet.

1. Philip Anastassiou, Jiawei Chen, Jitong Chen, Yuanzhe Chen, Zhuo Chen, Ziyi Chen, Jian Cong, Lelai Deng, Chuang Ding, Lu Gao, and 1 others. 2024. [Seed-tts: A family of high-quality versatile speech generation models](https://md.research.elamur.ai/free/en/papers/anastassiou-2024-seed-tts-family-high/). arXiv preprint arXiv:2406.02430.
2. Joon Son Chung, Arsha Nagrani, and Andrew Zisserman. 2018. Voxceleb2: Deep speaker recognition. In Proceedings of the Annual Conference of the International Speech Communication Association, INTER-SPEECH, volume 2018, pages 1086–1090. [find](https://scholar.google.com/scholar?q=Joon+Son+Chung%2C+Arsha+Nagrani%2C+and+Andrew+Zisserman%2E+2018%2E+Voxceleb2%3A+Deep+speaker+recognition%2E+In+Proceedings+of+the+Annual+Conference+of+the+International+Speech+Communication+Association%2C+INTER%2DSPE)
3. Chaorui Deng, Deyao Zhu, Kunchang Li, Chenhui Gou, Feng Li, Zeyu Wang, Shu Zhong, Weihao Yu, Xiaonan Nie, Ziang Song, and 1 others. 2025. Emerging properties in unified multimodal pretraining. arXiv preprint arXiv:2505.14683. — [arXiv:2505.14683](https://arxiv.org/abs/2505.14683) [find](https://scholar.google.com/scholar?q=Chaorui+Deng%2C+Deyao+Zhu%2C+Kunchang+Li%2C+Chenhui+Gou%2C+Feng+Li%2C+Zeyu+Wang%2C+Shu+Zhong%2C+Weihao+Yu%2C+Xiaonan+Nie%2C+Ziang+Song%2C+and+1+others%2E+2025%2E+Emerging+properties+in+unified+multimodal+pretraining%2E+arXiv+p)
4. Karel D’Oosterlinck, Winnie Xu, Chris Develder, Thomas Demeester, Amanpreet Singh, Christopher Potts, Douwe Kiela, and Shikib Mehri. 2025. [Anchored preference optimization and contrastive revisions: Addressing underspecification in alignment](https://md.research.elamur.ai/free/en/papers/oosterlinck-2025-anchored-preference-optimization-contrastive-revisions/). Transactions of the Association for Computational Linguis…
5. Zhihao Du, Yuxuan Wang, Qian Chen, Xian Shi, Xiang Lv, Tianyu Zhao, Zhifu Gao, Yexin Yang, Changfeng Gao, Hui Wang, Fan Yu, Huadai Liu, Zhengyan Sheng, Yue Gu, Chong Deng, Wen Wang, Shiliang Zhang, Zhijie Yan, and Jingren Zhou. 2024a. Cosyvoice 2: Scalable streaming speech synthesis with large langu… [find](https://scholar.google.com/scholar?q=Zhihao+Du%2C+Yuxuan+Wang%2C+Qian+Chen%2C+Xian+Shi%2C+Xiang+Lv%2C+Tianyu+Zhao%2C+Zhifu+Gao%2C+Yexin+Yang%2C+Changfeng+Gao%2C+Hui+Wang%2C+Fan+Yu%2C+Huadai+Liu%2C+Zhengyan+Sheng%2C+Yue+Gu%2C+Chong+Deng%2C+Wen+Wang%2C+Shiliang+Zhang%2C+Zh)
6. Zhihao Du, Yuxuan Wang, Qian Chen, Xian Shi, Xiang Lv, Tianyu Zhao, Zhifu Gao, Yexin Yang, Changfeng Gao, Hui Wang, and 1 others. 2024b. Cosyvoice 2: Scalable streaming speech synthesis with large language models. arXiv preprint arXiv:2412.10117. [find](https://scholar.google.com/scholar?q=Zhihao+Du%2C+Yuxuan+Wang%2C+Qian+Chen%2C+Xian+Shi%2C+Xiang+Lv%2C+Tianyu+Zhao%2C+Zhifu+Gao%2C+Yexin+Yang%2C+Changfeng+Gao%2C+Hui+Wang%2C+and+1+others%2E+2024b%2E+Cosyvoice+2%3A+Scalable+streaming+speech+synthesis+with+large+lan)
7. Zhifang Guo, Yichong Leng, Yihan Wu, Sheng Zhao, and Xu Tan. 2023. Prompttts: Controllable text-to-speech with text descriptions. In ICASSP 2023-2023 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP), pages 1–5. IEEE. [find](https://scholar.google.com/scholar?q=Zhifang+Guo%2C+Yichong+Leng%2C+Yihan+Wu%2C+Sheng+Zhao%2C+and+Xu+Tan%2E+2023%2E+Prompttts%3A+Controllable+text%2Dto%2Dspeech+with+text+descriptions%2E+In+ICASSP+2023%2D2023+IEEE+International+Conference+on+Acoustics%2C+Speech)
8. Wenyi Hong, Wenmeng Yu, Xiaotao Gu, Guo Wang, Guobing Gan, Haomiao Tang, Jiale Cheng, Ji Qi, Junhui Ji, Lihang Pan, and 1 others. 2025. Glm-4.1 v-thinking: Towards versatile multimodal reasoning with scalable reinforcement learning. arXiv e-prints, pages arXiv–2507. [find](https://scholar.google.com/scholar?q=Wenyi+Hong%2C+Wenmeng+Yu%2C+Xiaotao+Gu%2C+Guo+Wang%2C+Guobing+Gan%2C+Haomiao+Tang%2C+Jiale+Cheng%2C+Ji+Qi%2C+Junhui+Ji%2C+Lihang+Pan%2C+and+1+others%2E+2025%2E+Glm%2D4%2E1+v%2Dthinking%3A+Towards+versatile+multimodal+reasoning+with+)
9. Rongjie Huang, Ruofan Hu, Yongqi Wang, Zehan Wang, Xize Cheng, Ziyue Jiang, Zhenhui Ye, Dongchao Yang, Luping Liu, Peng Gao, and 1 others. 2024. Instructspeech: Following speech editing instructions via large language models. In Forty-first International Conference on Machine Learning. [find](https://scholar.google.com/scholar?q=Rongjie+Huang%2C+Ruofan+Hu%2C+Yongqi+Wang%2C+Zehan+Wang%2C+Xize+Cheng%2C+Ziyue+Jiang%2C+Zhenhui+Ye%2C+Dongchao+Yang%2C+Luping+Liu%2C+Peng+Gao%2C+and+1+others%2E+2024%2E+Instructspeech%3A+Following+speech+editing+instructions+v)
10. Wenxuan Huang, Shuang Chen, Zheyong Xie, Shaosheng Cao, Shixiang Tang, Yufan Shen, Qingyu Yin, Wenbo Hu, Xiaoman Wang, Yuntian Tang, Junbo Qiao, Yue Guo, Yao Hu, Zhenfei Yin, Philip Torr, Yu Cheng, Wanli Ouyang, and Shaohui Lin. 2025a. Interleaving reasoning for better text-to-image generation. Prep… — [arXiv:2509.06945](https://arxiv.org/abs/2509.06945) [find](https://scholar.google.com/scholar?q=Wenxuan+Huang%2C+Shuang+Chen%2C+Zheyong+Xie%2C+Shaosheng+Cao%2C+Shixiang+Tang%2C+Yufan+Shen%2C+Qingyu+Yin%2C+Wenbo+Hu%2C+Xiaoman+Wang%2C+Yuntian+Tang%2C+Junbo+Qiao%2C+Yue+Guo%2C+Yao+Hu%2C+Zhenfei+Yin%2C+Philip+Torr%2C+Yu+Cheng%2C+Wa)
11. Wenxuan Huang, Bohan Jia, Zijie Zhai, Shaosheng Cao, Zheyu Ye, Fei Zhao, Zhe Xu, Yao Hu, and Shaohui Lin. 2025b. Vision-r1: Incentivizing reasoning capability in multimodal large language models. Preprint, arXiv:2503.06749. [find](https://scholar.google.com/scholar?q=Wenxuan+Huang%2C+Bohan+Jia%2C+Zijie+Zhai%2C+Shaosheng+Cao%2C+Zheyu+Ye%2C+Fei+Zhao%2C+Zhe+Xu%2C+Yao+Hu%2C+and+Shaohui+Lin%2E+2025b%2E+Vision%2Dr1%3A+Incentivizing+reasoning+capability+in+multimodal+large+language+models%2E+Prep)
12. Pegah Jandaghi, Xianghai Sheng, Xinyi Bai, Jay Pujara, and Hakim Sidahmed. 2024. Faithful persona-based conversational dataset generation with large language models. In Proceedings of the 6th Workshop on NLP for Conversational AI (NLP4ConvAI 2024), pages 114–139, Bangkok, Thailand. Association for C… [find](https://scholar.google.com/scholar?q=Pegah+Jandaghi%2C+Xianghai+Sheng%2C+Xinyi+Bai%2C+Jay+Pujara%2C+and+Hakim+Sidahmed%2E+2024%2E+Faithful+persona%2Dbased+conversational+dataset+generation+with+large+language+models%2E+In+Proceedings+of+the+6th+Workshop)
13. Shengpeng Ji, Qian Chen, Wen Wang, Jialong Zuo, Minghui Fang, Ziyue Jiang, Hai Huang, Zehan Wang, Xize Cheng, Siqi Zheng, and Zhou Zhao. 2025. [Controlspeech: Towards simultaneous and independent zero-shot speaker cloning and zero-shot language style control](https://md.research.elamur.ai/free/en/papers/ji-2025-controlspeech-simultaneous-independent-zero-shot/). Preprint, arXiv:2406.01205.
14. Dongzhi Jiang, Ziyu Guo, Renrui Zhang, Zhuofan Zong, Hao Li, Le Zhuo, Shilin Yan, Pheng-Ann Heng, and Hongsheng Li. 2025. T2i-r1: Reinforcing image generation with collaborative semantic-level and token-level cot. arXiv preprint arXiv:2505.00703. — [arXiv:2505.00703](https://arxiv.org/abs/2505.00703) [find](https://scholar.google.com/scholar?q=Dongzhi+Jiang%2C+Ziyu+Guo%2C+Renrui+Zhang%2C+Zhuofan+Zong%2C+Hao+Li%2C+Le+Zhuo%2C+Shilin+Yan%2C+Pheng%2DAnn+Heng%2C+and+Hongsheng+Li%2E+2025%2E+T2i%2Dr1%3A+Reinforcing+image+generation+with+collaborative+semantic%2Dlevel+and+tok)
15. Ziyue Jiang, Jinglin Liu, Yi Ren, Jinzheng He, Zhenhui Ye, Shengpeng Ji, Qian Yang, Chen Zhang, Pengfei Wei, Chunfeng Wang, Xiang Yin, Zejun Ma, and Zhou Zhao. 2024. Mega-tts 2: Boosting prompting mechanisms for zero-shot speech synthesis. Preprint, arXiv:2307.07218. — [arXiv:2307.07218](https://arxiv.org/abs/2307.07218) [find](https://scholar.google.com/scholar?q=Ziyue+Jiang%2C+Jinglin+Liu%2C+Yi+Ren%2C+Jinzheng+He%2C+Zhenhui+Ye%2C+Shengpeng+Ji%2C+Qian+Yang%2C+Chen+Zhang%2C+Pengfei+Wei%2C+Chunfeng+Wang%2C+Xiang+Yin%2C+Zejun+Ma%2C+and+Zhou+Zhao%2E+2024%2E+Mega%2Dtts+2%3A+Boosting+prompting+mec)
16. Yinghao Aaron Li, Xilin Jiang, Cong Han, and Nima Mesgarani. 2024. Styletts-zs: Efficient high-quality zero-shot text-to-speech synthesis with distilled time-varying style diffusion. Preprint, arXiv:2409.10058. [find](https://scholar.google.com/scholar?q=Yinghao+Aaron+Li%2C+Xilin+Jiang%2C+Cong+Han%2C+and+Nima+Mesgarani%2E+2024%2E+Styletts%2Dzs%3A+Efficient+high%2Dquality+zero%2Dshot+text%2Dto%2Dspeech+synthesis+with+distilled+time%2Dvarying+style+diffusion%2E+Preprint%2C+arXiv%3A2)
17. Jiaqi Liao, Zhengyuan Yang, Linjie Li, Dianqi Li, Kevin Lin, Yu Cheng, and Lijuan Wang. 2025. Imagegen-cot: Enhancing text-to-image in-context learning with chain-of-thought reasoning. arXiv preprint arXiv:2503.19312. — [arXiv:2503.19312](https://arxiv.org/abs/2503.19312) [find](https://scholar.google.com/scholar?q=Jiaqi+Liao%2C+Zhengyuan+Yang%2C+Linjie+Li%2C+Dianqi+Li%2C+Kevin+Lin%2C+Yu+Cheng%2C+and+Lijuan+Wang%2E+2025%2E+Imagegen%2Dcot%3A+Enhancing+text%2Dto%2Dimage+in%2Dcontext+learning+with+chain%2Dof%2Dthought+reasoning%2E+arXiv+preprint+)
18. Run Luo, Lu Wang, Wanwei He, and Xiaobo Xia. 2025. Gui-r1 : A generalist r1-style vision-language action model for gui agents. Preprint, arXiv:2504.10458. — [arXiv:2504.10458](https://arxiv.org/abs/2504.10458) [find](https://scholar.google.com/scholar?q=Run+Luo%2C+Lu+Wang%2C+Wanwei+He%2C+and+Xiaobo+Xia%2E+2025%2E+Gui%2Dr1+%3A+A+generalist+r1%2Dstyle+vision%2Dlanguage+action+model+for+gui+agents%2E+Preprint%2C+arXiv%3A2504%2E10458%2E+%E2%80%94+%5BarXiv%3A2504%2E10458%5D%28https%3A%2F%2Farxiv%2Eorg%2Fabs%2F25)
19. Arsha Nagraniy, Joon Son Chungy, and Andrew Zisserman. 2017. Voxceleb: A large-scale speaker identification dataset. In Proceedings of the Annual Conference of the International Speech Communication Association, INTERSPEECH, volume 2017, pages 2616–2620. [find](https://scholar.google.com/scholar?q=Arsha+Nagraniy%2C+Joon+Son+Chungy%2C+and+Andrew+Zisserman%2E+2017%2E+Voxceleb%3A+A+large%2Dscale+speaker+identification+dataset%2E+In+Proceedings+of+the+Annual+Conference+of+the+International+Speech+Communication+A)
20. Tu Anh Nguyen, Wei-Ning Hsu, Antony D’Avirro, Bowen Shi, Itai Gat, Maryam Fazel-Zarandi, Tal Remez, Jade Copet, Gabriel Synnaeve, Michael Hassid, and 1 others. 2023. Expresso: A benchmark and analysis of discrete expressive speech resynthesis. In INTERSPEECH. [find](https://scholar.google.com/scholar?q=Tu+Anh+Nguyen%2C+Wei%2DNing+Hsu%2C+Antony+D%E2%80%99Avirro%2C+Bowen+Shi%2C+Itai+Gat%2C+Maryam+Fazel%2DZarandi%2C+Tal+Remez%2C+Jade+Copet%2C+Gabriel+Synnaeve%2C+Michael+Hassid%2C+and+1+others%2E+2023%2E+Expresso%3A+A+benchmark+and+analysis)
21. Yong Ren, Jiangyan Yi, Jianhua Tao, Haiyang Sun, Zhengqi Wen, Hao Gu, Le Xu, and Ye Bai. 2026. Ov-instructtts: Towards open-vocabulary instruct text-to-speech. arXiv preprint arXiv:2601.01459. — [arXiv:2601.01459](https://arxiv.org/abs/2601.01459) [find](https://scholar.google.com/scholar?q=Yong+Ren%2C+Jiangyan+Yi%2C+Jianhua+Tao%2C+Haiyang+Sun%2C+Zhengqi+Wen%2C+Hao+Gu%2C+Le+Xu%2C+and+Ye+Bai%2E+2026%2E+Ov%2Dinstructtts%3A+Towards+open%2Dvocabulary+instruct+text%2Dto%2Dspeech%2E+arXiv+preprint+arXiv%3A2601%2E01459%2E+%E2%80%94+%5BarXi)
22. Julius Richter, Yi-Chiao Wu, Steven Krenn, Simon Welker, Bunlong Lay, Shinji Watanabe, Alexander Richard, and Timo Gerkmann. 2024. Ears: An anechoic fullband speech dataset benchmarked for speech enhancement and dereverberation. In Proc. [find](https://scholar.google.com/scholar?q=Julius+Richter%2C+Yi%2DChiao+Wu%2C+Steven+Krenn%2C+Simon+Welker%2C+Bunlong+Lay%2C+Shinji+Watanabe%2C+Alexander+Richard%2C+and+Timo+Gerkmann%2E+2024%2E+Ears%3A+An+anechoic+fullband+speech+dataset+benchmarked+for+speech+enha)
23. Interspeech 2024, pages 4873–4877. [find](https://scholar.google.com/scholar?q=Interspeech+2024%2C+pages+4873%E2%80%934877%2E)
24. Maarten Sap, Hannah Rashkin, Derek Chen, Ronan Le Bras, and Yejin Choi. 2019. Social IQa: Commonsense reasoning about social interactions. In Proceedings of the 2019 Conference on Empirical Methods in Natural Language Processing and the 9th International Joint Conference on Natural Language Processi… [find](https://scholar.google.com/scholar?q=Maarten+Sap%2C+Hannah+Rashkin%2C+Derek+Chen%2C+Ronan+Le+Bras%2C+and+Yejin+Choi%2E+2019%2E+Social+IQa%3A+Commonsense+reasoning+about+social+interactions%2E+In+Proceedings+of+the+2019+Conference+on+Empirical+Methods+in)
25. Elvis Saravia, Hsien-Chi Toby Liu, Yen-Hao Huang, Junlin Wu, and Yi-Shin Chen. 2018. [Carer: Contextualized affect representations for emotion recognition](https://md.research.elamur.ai/free/en/papers/saravia-2018-carer-contextualized-affect-representations-emotion/). In Proceedings of the 2018 conference on empirical methods in natural language processing, pages 3687–3697. and 1 others. 2025a. Capspeech: Enabl…
26. Xinsheng Wang, Mingqi Jiang, Ziyang Ma, Ziyu Zhang, Songxiang Liu, Linqin Li, Zheng Liang, Qixi Zheng, Rui Wang, Xiaoqin Feng, and 1 others. 2025b. Spark-tts: An efficient llm-based text-to-speech model with single-stream decoupled speech tokens. arXiv preprint arXiv:2503.01710. — [arXiv:2503.01710](https://arxiv.org/abs/2503.01710) [find](https://scholar.google.com/scholar?q=Xinsheng+Wang%2C+Mingqi+Jiang%2C+Ziyang+Ma%2C+Ziyu+Zhang%2C+Songxiang+Liu%2C+Linqin+Li%2C+Zheng+Liang%2C+Qixi+Zheng%2C+Rui+Wang%2C+Xiaoqin+Feng%2C+and+1+others%2E+2025b%2E+Spark%2Dtts%3A+An+efficient+llm%2Dbased+text%2Dto%2Dspeech+mod)
27. Jin Xu, Zhifang Guo, Hangrui Hu, Yunfei Chu, Xiong Wang, Jinzheng He, Yuxuan Wang, Xian Shi, Ting He, Xinfa Zhu, and 1 others. 2025. [Qwen3-omni technical report](https://md.research.elamur.ai/free/en/papers/xu-2025-qwen3-omni-technical-report/). arXiv preprint arXiv:2509.17765.
28. Dongchao Yang, Songxiang Liu, Rongjie Huang, Chao Weng, and Helen Meng. 2024. Instructtts: Modelling expressive tts in discrete latent space with natural language style prompt. IEEE/ACM Transactions on Audio, Speech, and Language Processing, 32:2913– 2925. [find](https://scholar.google.com/scholar?q=Dongchao+Yang%2C+Songxiang+Liu%2C+Rongjie+Huang%2C+Chao+Weng%2C+and+Helen+Meng%2E+2024%2E+Instructtts%3A+Modelling+expressive+tts+in+discrete+latent+space+with+natural+language+style+prompt%2E+IEEE%2FACM+Transactions+o)
29. Guanrou Yang, Chen Yang, Qian Chen, Ziyang Ma, Wenxi Chen, Wen Wang, Tianrui Wang, Yifan Yang, Zhikang Niu, Wenrui Liu, and 1 others. 2025. Emovoice: Llm-based emotional text-to-speech model with freestyle text prompting. arXiv preprint arXiv:2504.12867. — [arXiv:2504.12867](https://arxiv.org/abs/2504.12867) [find](https://scholar.google.com/scholar?q=Guanrou+Yang%2C+Chen+Yang%2C+Qian+Chen%2C+Ziyang+Ma%2C+Wenxi+Chen%2C+Wen+Wang%2C+Tianrui+Wang%2C+Yifan+Yang%2C+Zhikang+Niu%2C+Wenrui+Liu%2C+and+1+others%2E+2025%2E+Emovoice%3A+Llm%2Dbased+emotional+text%2Dto%2Dspeech+model+with+free)
30. Aohan Zeng, Zhengxiao Du, Mingdao Liu, Kedong Wang, Shengmin Jiang, Lei Zhao, Yuxiao Dong, and Jie Tang. 2024. [Glm-4-voice: Towards intelligent and human-like end-to-end spoken chatbot](https://md.research.elamur.ai/free/en/papers/zeng-2024-glm-4-voice-intelligent/). arXiv preprint arXiv:2412.02612.
31. Haozhan Shen, Peng Liu, Jingcheng Li, Chunxin Fang, Yibo Ma, Jiajia Liao, Qiaoli Shen, Zilun Zhang, Kangjia Zhao, Qianqian Zhang, Ruochen Xu, and Tiancheng Zhao. 2025. Vlm-r1: A stable and generalizable r1-style large vision-language model. Preprint, arXiv:2504.07615. [find](https://scholar.google.com/scholar?q=Haozhan+Shen%2C+Peng+Liu%2C+Jingcheng+Li%2C+Chunxin+Fang%2C+Yibo+Ma%2C+Jiajia+Liao%2C+Qiaoli+Shen%2C+Zilun+Zhang%2C+Kangjia+Zhao%2C+Qianqian+Zhang%2C+Ruochen+Xu%2C+and+Tiancheng+Zhao%2E+2025%2E+Vlm%2Dr1%3A+A+stable+and+generalizab)
32. Bowen Zhang, Congchao Guo, Geng Yang, Hang Yu, Haozhe Zhang, Heidi Lei, Jialong Mai, Junjie Yan, Kaiyue Yang, Mingqi Yang, and 1 others. 2025a. Minimax-speech: Intrinsic zero-shot text-to-speech with a learnable speaker encoder. arXiv preprint arXiv:2505.07916. — [arXiv:2505.07916](https://arxiv.org/abs/2505.07916) [find](https://scholar.google.com/scholar?q=Bowen+Zhang%2C+Congchao+Guo%2C+Geng+Yang%2C+Hang+Yu%2C+Haozhe+Zhang%2C+Heidi+Lei%2C+Jialong+Mai%2C+Junjie+Yan%2C+Kaiyue+Yang%2C+Mingqi+Yang%2C+and+1+others%2E+2025a%2E+Minimax%2Dspeech%3A+Intrinsic+zero%2Dshot+text%2Dto%2Dspeech+with+)
33. Zhaochen Su, Linjie Li, Mingyang Song, Yunzhuo Hao, Zhengyuan Yang, Jun Zhang, Guanjie Chen, Jiawei Gu, Juntao Li, Xiaoye Qu, and 1 others. 2025. Openthinkimg: Learning to think with images via visual tool reinforcement learning. arXiv preprint arXiv:2505.08617. — [arXiv:2505.08617](https://arxiv.org/abs/2505.08617) [find](https://scholar.google.com/scholar?q=Zhaochen+Su%2C+Linjie+Li%2C+Mingyang+Song%2C+Yunzhuo+Hao%2C+Zhengyuan+Yang%2C+Jun+Zhang%2C+Guanjie+Chen%2C+Jiawei+Gu%2C+Juntao+Li%2C+Xiaoye+Qu%2C+and+1+others%2E+2025%2E+Openthinkimg%3A+Learning+to+think+with+images+via+visual)
34. Xintong Zhang, Zhi Gao, Bofei Zhang, Pengxiang Li, Xiaowen Zhang, Yang Liu, Tao Yuan, Yuwei Wu, Yunde Jia, Song-Chun Zhu, and 1 others. 2025b. Chain-of-focus: Adaptive visual search and zooming for multimodal reasoning via rl. arXiv preprint arXiv:2505.15436. — [arXiv:2505.15436](https://arxiv.org/abs/2505.15436) [find](https://scholar.google.com/scholar?q=Xintong+Zhang%2C+Zhi+Gao%2C+Bofei+Zhang%2C+Pengxiang+Li%2C+Xiaowen+Zhang%2C+Yang+Liu%2C+Tao+Yuan%2C+Yuwei+Wu%2C+Yunde+Jia%2C+Song%2DChun+Zhu%2C+and+1+others%2E+2025b%2E+Chain%2Dof%2Dfocus%3A+Adaptive+visual+search+and+zooming+for+mu)
35. Christophe Veaux, Junichi Yamagishi, Kirsten MacDonald, and 1 others. 2017. Cstr vctk corpus: English multi-speaker corpus for cstr voice cloning toolkit. University of Edinburgh. The Centre for Speech Technology Research (CSTR), 6:15. [find](https://scholar.google.com/scholar?q=Christophe+Veaux%2C+Junichi+Yamagishi%2C+Kirsten+MacDonald%2C+and+1+others%2E+2017%2E+Cstr+vctk+corpus%3A+English+multi%2Dspeaker+corpus+for+cstr+voice+cloning+toolkit%2E+University+of+Edinburgh%2E+The+Centre+for+Speec)
36. Ziwei Zheng, Michael Yang, Jack Hong, Chenxiao Zhao, Guohai Xu, Le Yang, Chao Shen, and Xing Yu. 2025. Deepeyes: Incentivizing" thinking with images" via reinforcement learning. arXiv preprint arXiv:2505.14362. — [arXiv:2505.14362](https://arxiv.org/abs/2505.14362) [find](https://scholar.google.com/scholar?q=Ziwei+Zheng%2C+Michael+Yang%2C+Jack+Hong%2C+Chenxiao+Zhao%2C+Guohai+Xu%2C+Le+Yang%2C+Chao+Shen%2C+and+Xing+Yu%2E+2025%2E+Deepeyes%3A+Incentivizing%22+thinking+with+images%22+via+reinforcement+learning%2E+arXiv+preprint+arXiv%3A2)
37. Chengyi Wang, Sanyuan Chen, Yu Wu, Ziqiang Zhang, Long Zhou, Shujie Liu, Zhuo Chen, Yanqing Liu, Huaming Wang, Jinyu Li, Lei He, Sheng Zhao, and Furu Wei. 2023. [Neural codec language models are zero-shot text to speech synthesizers](https://md.research.elamur.ai/free/en/papers/wang-2023-neural-codec-language-models-zero/). Preprint, arXiv:2301.02111.
38. Helin Wang, Jiarui Hai, Dading Chong, Karan Thakkar, Tiantian Feng, Dongchao Yang, Junhyeok Lee, Laureano Moro Velazquez, Jesus Villalba, Zengyi Qin, [find](https://scholar.google.com/scholar?q=Helin+Wang%2C+Jiarui+Hai%2C+Dading+Chong%2C+Karan+Thakkar%2C+Tiantian+Feng%2C+Dongchao+Yang%2C+Junhyeok+Lee%2C+Laureano+Moro+Velazquez%2C+Jesus+Villalba%2C+Zengyi+Qin%2C)

*Updated 1 November 2026*