Saved in:
Bibliographic Details
Main Authors: Tan, Chao-Hong, Chen, Qian, Wang, Wen, Deng, Chong, Zhang, Qinglin, Cheng, Luyao, Yu, Hai, Zhang, Xin, Lv, Xiang, Zhao, Tianyu, Zhang, Chong, Ma, Yukun, Chen, Yafeng, Wang, Hui, Liu, Jiaqing, Li, Xiangang, Ye, Jieping
Format: Preprint
Published: 2025
Subjects:
Online Access:https://arxiv.org/abs/2506.09349
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866917164782452736
author Tan, Chao-Hong
Chen, Qian
Wang, Wen
Deng, Chong
Zhang, Qinglin
Cheng, Luyao
Yu, Hai
Zhang, Xin
Lv, Xiang
Zhao, Tianyu
Zhang, Chong
Ma, Yukun
Chen, Yafeng
Wang, Hui
Liu, Jiaqing
Li, Xiangang
Ye, Jieping
author_facet Tan, Chao-Hong
Chen, Qian
Wang, Wen
Deng, Chong
Zhang, Qinglin
Cheng, Luyao
Yu, Hai
Zhang, Xin
Lv, Xiang
Zhao, Tianyu
Zhang, Chong
Ma, Yukun
Chen, Yafeng
Wang, Hui
Liu, Jiaqing
Li, Xiangang
Ye, Jieping
contents Recent studies on end-to-end (E2E) speech generation with large language models (LLMs) have attracted significant community attention, with multiple works extending text-based LLMs to generate discrete speech tokens. Existing E2E approaches primarily fall into two categories: (1) Methods that generate discrete speech tokens independently without incorporating them into the LLM's autoregressive process, resulting in text generation being unaware of concurrent speech synthesis. (2) Models that generate interleaved or parallel speech-text tokens through joint autoregressive modeling, enabling mutual modality awareness during generation. This paper presents DrVoice, a parallel speech-text voice conversation model based on joint autoregressive modeling, featuring dual-resolution speech representations. Notably, while current methods utilize mainly 12.5Hz input audio representation, our proposed dual-resolution mechanism reduces the input frequency for the LLM to 5Hz, significantly reducing computational cost and alleviating the frequency discrepancy between speech and text tokens and in turn better exploiting LLMs' capabilities. Experimental results demonstrate that DrVoice-7B establishes new state-of-the-art (SOTA) on prominent speech benchmarks including OpenAudioBench, VoiceBench, UltraEval-Audio and Big Bench Audio, making it a leading open-source speech foundation model in ~7B models.
format Preprint
id arxiv_https___arxiv_org_abs_2506_09349
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle DrVoice: Parallel Speech-Text Voice Conversation Model via Dual-Resolution Speech Representations
Tan, Chao-Hong
Chen, Qian
Wang, Wen
Deng, Chong
Zhang, Qinglin
Cheng, Luyao
Yu, Hai
Zhang, Xin
Lv, Xiang
Zhao, Tianyu
Zhang, Chong
Ma, Yukun
Chen, Yafeng
Wang, Hui
Liu, Jiaqing
Li, Xiangang
Ye, Jieping
Computation and Language
Recent studies on end-to-end (E2E) speech generation with large language models (LLMs) have attracted significant community attention, with multiple works extending text-based LLMs to generate discrete speech tokens. Existing E2E approaches primarily fall into two categories: (1) Methods that generate discrete speech tokens independently without incorporating them into the LLM's autoregressive process, resulting in text generation being unaware of concurrent speech synthesis. (2) Models that generate interleaved or parallel speech-text tokens through joint autoregressive modeling, enabling mutual modality awareness during generation. This paper presents DrVoice, a parallel speech-text voice conversation model based on joint autoregressive modeling, featuring dual-resolution speech representations. Notably, while current methods utilize mainly 12.5Hz input audio representation, our proposed dual-resolution mechanism reduces the input frequency for the LLM to 5Hz, significantly reducing computational cost and alleviating the frequency discrepancy between speech and text tokens and in turn better exploiting LLMs' capabilities. Experimental results demonstrate that DrVoice-7B establishes new state-of-the-art (SOTA) on prominent speech benchmarks including OpenAudioBench, VoiceBench, UltraEval-Audio and Big Bench Audio, making it a leading open-source speech foundation model in ~7B models.
title DrVoice: Parallel Speech-Text Voice Conversation Model via Dual-Resolution Speech Representations
topic Computation and Language
url https://arxiv.org/abs/2506.09349