arXiv subjects

Xingchen Song

Publications and source records attributed to Xingchen Song.

At least 19 recordsLinked to original sources

arXiv2026

Semantic Refinement of Universal Audio Representations through Audio-Description Alignment

Universal audio representations must preserve acoustic detail while making high-level concepts accessible across speech, music, environmental sound, and downstream models of different capacities. We study semantic refinement of an acoustically pretrained encoder by adding audio-description alignment to a foundation of BEST-RQ, reconstruction, and CTC. We compare matched control, shuffled-description, and correctly paired trajectories to distinguish correct correspondence from an extra contrastive objective. Each endpoint is frozen and evaluated with a temporal-mean linear probe and a sequence-aware LLM readout, testing whether the refined information is directly accessible and remains useful to a stronger model. Across three paired seeds, correct alignment improves domain-balanced classification by 4.66 points with the linear probe and 2.59 points with the sequence-aware LLM, with positive changes in every domain. Correct pairing accounts for 87% of the linear-probe gain, while the LLM shows its clearest correspondence-specific benefit in captioning. Dense acoustic objectives provide complementary gains under both readouts. A separate 24-layer continuation remains competitive with leading public encoders under the shared evaluator, supporting the recipe beyond the controlled study.

eess.AS↗

arXiv2026

Qwen-Audio-3.0-Gen-Preview Technical Report

Existing single-domain and multi-task audio systems remain limited in directly organizing heterogeneous audio components, ambience, and multiple roles into long-form temporal scenes. We present Qwen-Audio-3.0-Gen-Preview, a unified non-autoregressive framework that uses a Diffusion Transformer (DiT) and a shared variational autoencoder (VAE) to generate the complete mixed waveform. Prompt enhancement converts free-form requests into structured temporal records that are rendered as textual conditions, while a two-stage data curriculum and semantic conditional views train the proposed model to use these conditions across standalone and mixed-scene audio. A shared continuous VAE compresses 48kHz stereo waveforms into 25Hz latent sequences and incorporates semantic supervision, providing one representation for heterogeneous audio. On the public reference-conditioned benchmark, speaker similarity is the proposed model's clearest strength across all three subsets. Across the multi-speaker and rich-timeline benchmarks, its clearest comparative strengths are cross-turn consistency in both languages and temporal localization, respectively. On AudioCaps, its advantages are concentrated in evaluations using large audio-language models and AudioBox. These results demonstrate the potential of unified generation for temporally structured audio without task-specific branches.

eess.AS↗

arXiv2026

Harness TTS: Towards Context-Aware Expressive Speech Synthesis with Harness Layer

Expressive speech synthesis for voice assistants requires flexible style control that adapts to explicit requests and broader interaction context. We propose Harness TTS, a lightweight control layer that wraps around a TTS engine to externalize and govern its expressive behavior. It reformulates style control as closed-set prompt-tool routing: offline, a compact registry of stylistic prompt tools is constructed with structured metadata; online, an LLM planner selects the appropriate tool based on a priority-aware observation schema, and the TTS executor synthesizes speech using the corresponding prompt audio. We evaluate Harness TTS on both routing and synthesis tasks. In routing, Qwen3-4B achieves Top-1 accuracies of 74.3%, 43.0%, and 64.6% on explicit, implicit, and conflict subsets. For synthesis, experiments on CosyVoice3 and VoxCPM2 show that Harness TTS outperforms instruction-only control, achieving higher instruction-following win rates (margins of 23.1-35.6 points on CosyVoice3 and 13.8-20.0 points on VoxCPM2) and improving UTMOSv2 scores by 0.11-0.38. Moreover, the 4B planner delivers its first tool recommendation in under 50 ms in standard mode, introducing negligible latency for real-time interaction. These results demonstrate that equipping TTS engines with a dedicated Harness layer offers a practical, auditable, and context-aware solution for voice assistant expression control.

cs.SD↗

arXiv2026

F3-Tokenizer: Taming Audio Autoencoder Latents for Understanding and Generation

Continuous audio autoencoders reconstruct waveforms well but often produce latents with weak structure for understanding, while self-supervised audio encoders capture semantics but are not directly decodable. This mismatch complicates a single audio tokenizer that must support both understanding and generation. We adapt continuous autoencoder latents to this setting with two components: a noise-regularized autoencoder bottleneck and a latent-side representation encoder. The bottleneck uses channel normalization and stochastic perturbation instead of KL-based variational training, yielding scale-controlled continuous latents for reconstruction and autoregressive generation. The representation encoder is trained on frozen autoencoder latents with RQ-MTP and frozen-LLM supervision. The resulting tokenizer provides high-dimensional representations for understanding while preserving normalized continuous latents as generation targets

cs.SD↗

arXiv2026

TTS-PRISM: A Perceptual Reasoning and Interpretable Speech Model for Fine-Grained Diagnosis

While generative text-to-speech (TTS) models approach human-level quality, monolithic metrics fail to diagnose fine-grained acoustic artifacts or explain perceptual collapse. To address this, we propose TTS-PRISM, a multi-dimensional diagnostic framework for Mandarin. First, we establish a 12-dimensional schema spanning stability to advanced expressiveness. Second, we design a targeted synthesis pipeline with adversarial perturbations and expert anchors to build a high-quality diagnostic dataset. Third, schema-driven instruction tuning embeds explicit scoring criteria and reasoning into an efficient end-to-end model. Experiments on a 1,600-sample Gold Test Set show TTS-PRISM outperforms generalist models in human alignment. Profiling six TTS paradigms establishes intuitive diagnostic flags that reveal fine-grained capability differences. TTS-PRISM is open-source, with code and checkpoints at https://github.com/xiaomi-research/tts-prism.

cs.CL↗

arXiv2026

Iterate to Differentiate: Enhancing Discriminability and Reliability in Zero-Shot TTS Evaluation

Reliable evaluation of modern zero-shot text-to-speech (TTS) models remains challenging. Subjective tests are costly and hard to reproduce, while objective metrics often saturate, failing to distinguish SOTA systems. To address this, we propose Iterate to Differentiate (I2D), an evaluation framework that recursively synthesizes speech using the model's own outputs as references. Higher-quality models exhibit greater resilience to the distributional shift induced by iterative synthesis, resulting in slower performance degradation. I2D exploits this differential degradation to amplify performance gaps and reveal robustness. By aggregating objective metrics across iterations, I2D improves discriminability and alignment with human judgments, increasing system-level SRCC from 0.118 to 0.464 for UTMOSv2. Experiments on 11 models across Chinese, English, and emotion datasets demonstrate that I2D enables more reliable automated evaluation for zero-shot TTS.

cs.SD↗

arXiv2026

Borderless Long Speech Synthesis

Most existing text-to-speech (TTS) systems either synthesize speech sentence by sentence and stitch the results together, or drive synthesis from plain-text dialogues alone. Both approaches leave models with little understanding of global context or paralinguistic cues, making it hard to capture real-world phenomena such as multi-speaker interactions (interruptions, overlapping speech), evolving emotional arcs, and varied acoustic environments. We introduce the Borderless Long Speech Synthesis framework for agent-centric, borderless long audio synthesis. Rather than targeting a single narrow task, the system is designed as a unified capability set spanning VoiceDesigner, multi-speaker synthesis, Instruct TTS, and long-form text synthesis. On the data side, we propose a "Labeling over filtering/cleaning" strategy and design a top-down, multi-level annotation schema we call Global-Sentence-Token. On the model side, we adopt a backbone with a continuous tokenizer and add Chain-of-Thought (CoT) reasoning together with Dimension Dropout, both of which markedly improve instruction following under complex conditions. We further show that the system is Native Agentic by design: the hierarchical annotation doubles as a Structured Semantic Interface between the LLM Agent and the synthesis engine, creating a layered control protocol stack that spans from scene semantics down to phonetic detail. Text thereby becomes an information-complete, wide-band control channel, enabling a front-end LLM to convert inputs of any modality into structured generation commands, extending the paradigm from Text2Speech to borderless long speech synthesis.

cs.SD↗

arXiv2025

MiMo-Audio: Audio Language Models are Few-Shot Learners

Existing audio language models typically rely on task-specific fine-tuning to accomplish particular audio tasks. In contrast, humans are able to generalize to new audio tasks with only a few examples or simple instructions. GPT-3 has shown that scaling next-token prediction pretraining enables strong generalization capabilities in text, and we believe this paradigm is equally applicable to the audio domain. By scaling MiMo-Audio's pretraining data to over one hundred million of hours, we observe the emergence of few-shot learning capabilities across a diverse set of audio tasks. We develop a systematic evaluation of these capabilities and find that MiMo-Audio-7B-Base achieves SOTA performance on both speech intelligence and audio understanding benchmarks among open-source models. Beyond standard metrics, MiMo-Audio-7B-Base generalizes to tasks absent from its training data, such as voice conversion, style transfer, and speech editing. MiMo-Audio-7B-Base also demonstrates powerful speech continuation capabilities, capable of generating highly realistic talk shows, recitations, livestreaming and debates. At the post-training stage, we curate a diverse instruction-tuning corpus and introduce thinking mechanisms into both audio understanding and generation. MiMo-Audio-7B-Instruct achieves open-source SOTA on audio understanding benchmarks (MMSU, MMAU, MMAR, MMAU-Pro), spoken dialogue benchmarks (Big Bench Audio, MultiChallenge Audio) and instruct-TTS evaluations, approaching or surpassing closed-source models. Model checkpoints and full evaluation suite are available at https://github.com/XiaomiMiMo/MiMo-Audio.

cs.CL↗

arXiv2025

ZipVoice-Dialog: Non-Autoregressive Spoken Dialogue Generation with Flow Matching

Generating spoken dialogue is inherently more complex than monologue text-to-speech (TTS), as it demands both realistic turn-taking and the maintenance of distinct speaker timbres. While existing autoregressive (AR) models have made progress, they often suffer from high inference latency and stability issues. To overcome these limitations, we propose ZipVoice-Dialog, a non-autoregressive (NAR) zero-shot spoken dialogue generation model based on flow-matching. Observing that applying vanilla flow-matching to dialogue generation leads to poor speech intelligibility and turn-taking precision, we introduce two simple yet effective methods to adapt flow-matching architectures for dialogue generation: (1) a curriculum learning strategy to ensure robust speech-text alignment, and (2) speaker-turn embeddings to govern precise speaker turn-taking. Additionally, we introduce dedicated strategies to support stereo dialogue generation. Recognizing the lack of training datasets in this field, we curate and release OpenDialog, the first large-scale (6.8k hours) open-source spoken dialogue dataset derived from in-the-wild speech data. Moreover, for fair and rigorous evaluations, we established a benchmark to comprehensively evaluate dialogue generation models. Experiments demonstrate the effectiveness of the proposed methods and dataset, showing that ZipVoice-Dialog achieves superior performance in inference speed, intelligibility, speaker turn-taking accuracy, and speaker similarity. Our code, model checkpoints, and the OpenDialog dataset are publicly available at https://github.com/k2-fsa/ZipVoice.

eess.AS↗

arXiv2025

Adapting Whisper for Streaming Speech Recognition via Two-Pass Decoding

OpenAI Whisper is a family of robust Automatic Speech Recognition (ASR) models trained on 680,000 hours of audio. However, its encoder-decoder architecture, trained with a sequence-to-sequence objective, lacks native support for streaming ASR. In this paper, we fine-tune Whisper for streaming ASR using the WeNet toolkit by adopting a Unified Two-pass (U2) structure. We introduce an additional Connectionist Temporal Classification (CTC) decoder trained with causal attention masks to generate streaming partial transcripts, while the original Whisper decoder reranks these partial outputs. Our experiments on LibriSpeech and an earnings call dataset demonstrate that, with adequate fine-tuning data, Whisper can be adapted into a capable streaming ASR model. We also introduce a hybrid tokenizer approach, which uses a smaller token space for the CTC decoder while retaining Whisper's original token space for the attention decoder, resulting in improved data efficiency and generalization.

cs.SD↗

arXiv2025

MiMo-VL Technical Report

We open-source MiMo-VL-7B-SFT and MiMo-VL-7B-RL, two powerful vision-language models delivering state-of-the-art performance in both general visual understanding and multimodal reasoning. MiMo-VL-7B-RL outperforms Qwen2.5-VL-7B on 35 out of 40 evaluated tasks, and scores 59.4 on OlympiadBench, surpassing models with up to 78B parameters. For GUI grounding applications, it sets a new standard with 56.1 on OSWorld-G, even outperforming specialized models such as UI-TARS. Our training combines four-stage pre-training (2.4 trillion tokens) with Mixed On-policy Reinforcement Learning (MORL) integrating diverse reward signals. We identify the importance of incorporating high-quality reasoning data with long Chain-of-Thought into pre-training stages, and the benefits of mixed RL despite challenges in simultaneous multi-domain optimization. We also contribute a comprehensive evaluation suite covering 50+ tasks to promote reproducibility and advance the field. The model checkpoints and full evaluation suite are available at https://github.com/XiaomiMiMo/MiMo-VL.

cs.CL↗

arXiv2025

MiMo: Unlocking the Reasoning Potential of Language Model -- From Pretraining to Posttraining

We present MiMo-7B, a large language model born for reasoning tasks, with optimization across both pre-training and post-training stages. During pre-training, we enhance the data preprocessing pipeline and employ a three-stage data mixing strategy to strengthen the base model's reasoning potential. MiMo-7B-Base is pre-trained on 25 trillion tokens, with additional Multi-Token Prediction objective for enhanced performance and accelerated inference speed. During post-training, we curate a dataset of 130K verifiable mathematics and programming problems for reinforcement learning, integrating a test-difficulty-driven code-reward scheme to alleviate sparse-reward issues and employing strategic data resampling to stabilize training. Extensive evaluations show that MiMo-7B-Base possesses exceptional reasoning potential, outperforming even much larger 32B models. The final RL-tuned model, MiMo-7B-RL, achieves superior performance on mathematics, code and general reasoning tasks, surpassing the performance of OpenAI o1-mini. The model checkpoints are available at https://github.com/xiaomimimo/MiMo.

cs.CL↗

arXiv2024

TouchASP: Elastic Automatic Speech Perception that Everyone Can Touch

Large Automatic Speech Recognition (ASR) models demand a vast number of parameters, copious amounts of data, and significant computational resources during the training process. However, such models can merely be deployed on high-compute cloud platforms and are only capable of performing speech recognition tasks. This leads to high costs and restricted capabilities. In this report, we initially propose the elastic mixture of the expert (eMoE) model. This model can be trained just once and then be elastically scaled in accordance with deployment requirements. Secondly, we devise an unsupervised data creation and validation procedure and gather millions of hours of audio data from diverse domains for training. Using these two techniques, our system achieves elastic deployment capabilities while reducing the Character Error Rate (CER) on the SpeechIO testsets from 4.98\% to 2.45\%. Thirdly, our model is not only competent in Mandarin speech recognition but also proficient in multilingual, multi-dialect, emotion, gender, and sound event perception. We refer to this as Automatic Speech Perception (ASP), and the perception results are presented in the experimental section.

eess.AS↗

arXiv2024

TouchTTS: An Embarrassingly Simple TTS Framework that Everyone Can Touch

It is well known that LLM-based systems are data-hungry. Recent LLM-based TTS works typically employ complex data processing pipelines to obtain high-quality training data. These sophisticated pipelines require excellent models at each stage (e.g., speech denoising, speech enhancement, speaker diarization, and punctuation models), which themselves demand high-quality training data and are rarely open-sourced. Even with state-of-the-art models, issues persist, such as incomplete background noise removal and misalignment between punctuation and actual speech pauses. Moreover, the stringent filtering strategies often retain only 10-30\% of the original data, significantly impeding data scaling efforts. In this work, we leverage a noise-robust audio tokenizer (S3Tokenizer) to design a simplified yet effective TTS data processing pipeline that maintains data quality while substantially reducing data acquisition costs, achieving a data retention rate of over 50\%. Beyond data scaling challenges, LLM-based TTS systems also incur higher deployment costs compared to conventional approaches. Current systems typically use LLMs solely for text-to-token generation, while requiring separate models (e.g., flow matching models) for token-to-waveform generation, which cannot be directly executed by LLM inference engines, further complicating deployment. To address these challenges, we eliminate redundant modules in both LLM and flow components, replacing the flow model backbone with an LLM architecture. Building upon this simplified flow backbone, we propose a unified architecture for both streaming and non-streaming inference, significantly reducing deployment costs. Finally, we explore the feasibility of unifying TTS and ASR tasks using the same data for training, thanks to the simplified pipeline and the S3Tokenizer that reduces the quality requirements for TTS training data.

cs.SD↗

arXiv2024

HydraFormer: One Encoder For All Subsampling Rates

In automatic speech recognition, subsampling is essential for tackling diverse scenarios. However, the inadequacy of a single subsampling rate to address various real-world situations often necessitates training and deploying multiple models, consequently increasing associated costs. To address this issue, we propose HydraFormer, comprising HydraSub, a Conformer-based encoder, and a BiTransformer-based decoder. HydraSub encompasses multiple branches, each representing a distinct subsampling rate, allowing for the flexible selection of any branch during inference based on the specific use case. HydraFormer can efficiently manage different subsampling rates, significantly reducing training and deployment expenses. Experiments on AISHELL-1 and LibriSpeech datasets reveal that HydraFormer effectively adapts to various subsampling rates and languages while maintaining high recognition performance. Additionally, HydraFormer showcases exceptional stability, sustaining consistent performance under various initialization conditions, and exhibits robust transferability by learning from pretrained single subsampling rate automatic speech recognition models\footnote{Model code and scripts: https://github.com/HydraFormer/hydraformer}.

eess.AS↗

arXiv2024

U2++ MoE: Scaling 4.7x parameters with minimal impact on RTF

Scale has opened new frontiers in natural language processing, but at a high cost. In response, by learning to only activate a subset of parameters in training and inference, Mixture-of-Experts (MoE) have been proposed as an energy efficient path to even larger and more capable language models and this shift towards a new generation of foundation models is gaining momentum, particularly within the field of Automatic Speech Recognition (ASR). Recent works that incorporating MoE into ASR models have complex designs such as routing frames via supplementary embedding network, improving multilingual ability for the experts, and utilizing dedicated auxiliary losses for either expert load balancing or specific language handling. We found that delicate designs are not necessary, while an embarrassingly simple substitution of MoE layers for all Feed-Forward Network (FFN) layers is competent for the ASR task. To be more specific, we benchmark our proposed model on a large scale inner-source dataset (160k hours), the results show that we can scale our baseline Conformer (Dense-225M) to its MoE counterparts (MoE-1B) and achieve Dense-1B level Word Error Rate (WER) while maintaining a Dense-225M level Real Time Factor (RTF). Furthermore, by applying Unified 2-pass framework with bidirectional attention decoders (U2++), we achieve the streaming and non-streaming decoding modes in a single MoE based model, which we call U2++ MoE. We hope that our study can facilitate the research on scaling speech foundation models without sacrificing deployment efficiency.

cs.CL↗

arXiv2023

Spike-Triggered Contextual Biasing for End-to-End Mandarin Speech Recognition

The attention-based deep contextual biasing method has been demonstrated to effectively improve the recognition performance of end-to-end automatic speech recognition (ASR) systems on given contextual phrases. However, unlike shallow fusion methods that directly bias the posterior of the ASR model, deep biasing methods implicitly integrate contextual information, making it challenging to control the degree of bias. In this study, we introduce a spike-triggered deep biasing method that simultaneously supports both explicit and implicit bias. Moreover, both bias approaches exhibit significant improvements and can be cascaded with shallow fusion methods for better results. Furthermore, we propose a context sampling enhancement strategy and improve the contextual phrase filtering algorithm. Experiments on the public WenetSpeech Mandarin biased-word dataset show a 32.0% relative CER reduction compared to the baseline model, with an impressively 68.6% relative CER reduction on contextual phrases.

eess.AS↗

arXiv2023

LightGrad: Lightweight Diffusion Probabilistic Model for Text-to-Speech

Recent advances in neural text-to-speech (TTS) models bring thousands of TTS applications into daily life, where models are deployed in cloud to provide services for customs. Among these models are diffusion probabilistic models (DPMs), which can be stably trained and are more parameter-efficient compared with other generative models. As transmitting data between customs and the cloud introduces high latency and the risk of exposing private data, deploying TTS models on edge devices is preferred. When implementing DPMs onto edge devices, there are two practical problems. First, current DPMs are not lightweight enough for resource-constrained devices. Second, DPMs require many denoising steps in inference, which increases latency. In this work, we present LightGrad, a lightweight DPM for TTS. LightGrad is equipped with a lightweight U-Net diffusion decoder and a training-free fast sampling technique, reducing both model parameters and inference latency. Streaming inference is also implemented in LightGrad to reduce latency further. Compared with Grad-TTS, LightGrad achieves 62.2% reduction in paramters, 65.7% reduction in latency, while preserving comparable speech quality on both Chinese Mandarin and English in 4 denoising steps.

cs.SD↗