Confucius4‑R2T2, a stable‑prefix streaming ASR model, is introduced to ensure that once a token is emitted it is never revised, thereby reducing wasted work and incorrect actions in real‑time transcription. The model is derived by post‑training Qwen3‑ASR‑1.7B on 1,500 hours of Chinese and English audio while keeping its audio encoder frozen, and it leverages a new concept called Longest Stable Prefix (LSP) to convert a full‑context model into a stable‑prefix streaming one without architectural changes. Evaluation shows that R2T2 achieves near full‑context quality with low latency and stable output, and its benefits generalize beyond the languages used in training.
arXiv:2609.14991v1 Announce Type: new Abstract: Open Thai automatic speech recognition (ASR) is dominated by offline, Whisper-based models that read the whole utterance before transcribing, ruling ou...
Open Thai automatic speech recognition (ASR) is dominated by offline, Whisper-based models that read the whole utterance before transcribing, ruling out low-latency uses such as live captioning and vo...
Recent advances in zero-shot text-to-speech (TTS) have substantially improved speech quality and voice cloning fidelity. However, many zero-shot TTS systems still depend on audio prompt transcripts at inference time.
The paper introduces AWED, a word‑level emission‑delay metric, and demonstrates that post‑training a delayed‑streaming model with a joint reward (GRPO) improves both accuracy and latency. Using a single operating point (τ=6 frames), the method outperforms both its supervised baseline and the Voxtral Realtime backbone across all lookahead budgets, reducing WER by up to 30.8% at 80 ms delay and lowering median AWED from 1.17 s to 1.04 s.
arXiv:2607. 20086v1 Announce Type: cross Abstract: State-space sequence models are attractive for streaming speech because they maintain compact recurrent state, but scan-style training kernels can have unfavorable constants for short audio tasks.
X2Streaming-ASR introduces a method for streaming automatic speech recognition that separates the decision of when to commit a transcript from what to commit. The approach uses a three‑stage training process: first establishing streaming capability, then warm‑starting a commit policy with automatically probed trajectories, and finally refining the policy with character‑level, segment‑assigned group‑relative rewards for accuracy and latency. On AISHELL‑1/2/3 and WenetSpeech datasets, the system achieves mean character‑level commit latencies of 27–84 ms, far lower than baseline systems, while also attaining the best streaming character error rates on AISHELL‑1 and AISHELL‑3.
The Qwen-Audio-3.0-ASR Technical Report introduces a Mixture-of-Experts large language model-based automatic speech recognition system that addresses real‑world production challenges such as regional dialects, dynamic entities, hotwords, long‑range context, and disfluent speech. Built on the Qwen backbone and trained on tens of millions of hours of speech data, it supports transcription in 30 languages and 16 Chinese dialects, and offers industry‑domain entity recognition, hierarchical hotword customization, single‑pass polishing, and long‑audio contextual modeling. A streaming variant, Qwen-Audio-3.0-ASR-Streaming, is also presented for low‑latency applications, with evaluations showing state‑of‑the‑art performance against leading commercial systems.
The paper presents a method for improving simultaneous speech translation by adapting a full‑utterance speech language model with prefix supervision derived from its own complete and partial waveform translations, eliminating the need for transcripts or human translations. Experiments on FLEURS and CoVoST2 across three language directions show that prefix training enhances quality–latency trade‑offs, especially when combined with multi‑turn append‑only decoding, and that a confidence threshold effectively controls the inference‑time quality–latency balance. The study also explores the impact of synthesis margin on translation quality and calibration, finding a non‑monotonic relationship with latency.
arXiv:2606. 07080v1 Announce Type: cross Abstract: We present dots.
TontaubeV1 is a streaming text‑to‑speech model that preserves natural prosody while running on a single consumer GPU. It encodes speech with a hierarchical DualCodec representation at 12.5 Hz, separating a semantic stream from successive acoustic refinements, and uses Qwen3‑derived transformers to predict the semantic stream, utterance duration, and acoustic refinements. The system supports up to one minute of reference audio for voice conditioning, streams with a 200 ms latency to first audio, and achieves real‑time factors of 0.08 (single input) and 0.02 (eight concurrent inputs).
TontaubeV1 is a streaming text‑to‑speech model that maintains natural prosody while running on a single consumer GPU. It encodes speech with a hierarchical DualCodec representation at 12.5 Hz, separating a semantic stream from successive acoustic refinements, and uses Qwen3‑derived transformers to predict the semantic stream and add refinements. The model supports up to one minute of reference audio for voice conditioning, streams with a 200 ms latency to first audio, and achieves real‑time factors of 0.08 (single input) and 0.02 (eight concurrent inputs).
The paper introduces phoneme‑guided initialization for large language model‑based speech recognition. By pre‑training an audio encoder on speech‑to‑phoneme conversion and an LLM on phoneme‑to‑grapheme conversion, then fine‑tuning jointly on ASR, the method leverages phoneme‑mediated processing within an end‑to‑end framework. Experiments on Japanese, Chinese, Tatar, and Urdu demonstrate that this approach matches or surpasses both cascaded S2P‑P2G baselines and end‑to‑end models without phoneme initialization.