arXiv subjects

Jiatong Shi

Publications and source records attributed to Jiatong Shi.

At least 19 recordsLinked to original sources

arXiv2026

Bagpiper: Solving Open-Ended Audio Tasks via Rich Captions

Current audio foundation models typically rely on rigid, task-specific supervision (e.g., speech recognition), addressing isolated factors of audio rather than the whole. In contrast, human processes audio holistically, seamlessly bridging raw audio waveform with abstract cognitive concepts (e.g., all perception details of audio events) to execute complex tasks. Grounded in this philosophy, we introduce Bagpiper, an 8B audio foundation model that interprets physical audio via rich captions, i.e., comprehensive natural language descriptions that encapsulate the critical cognitive concepts inherent in the audio. By pre-training on a massive corpus of 600B tokens, the model establishes a robust bidirectional mapping between raw audio and this high-level conceptual space. During fine-tuning, Bagpiper adopts a caption-then-process workflow, simulating an intermediate cognitive reasoning step to solve diverse tasks without knowing prior task-specific practice. Experimentally, Bagpiper achieves universal generation that can uniformly generate speech, sound effects, music, and their arbitrary combinations. It also maintains comparable performance with the 7B Qwen-2.5-Omni for audio understanding. To the best of our knowledge, Bagpiper is among the first works that achieve open-ended audio understanding and generation on speech, sound, and music. Model, data, and code will be released at Bagpiper Home Page.

cs.CL↗

arXiv2026

MMGenre: Benchmarking Singing Voice Synthesis across Multiple Musical Genres

Singing voice synthesis (SVS) has progressed rapidly, yet its ability to generalize across diverse musical genres remains underexplored. Existing benchmarks are heavily biased toward pop music, limiting systematic analysis of genre-dependent behavior. We introduce MMGenre, a benchmark for multi-genre SVS diagnosis, supported by an automatic pipeline for constructing genre-aligned music scores. MMGenre spans 10 major genres and 26 subgenres, enabling comprehensive analysis of genre-aware synthesis. Extensive evaluation of representative SVS models reveals limited genre discrimination: synthesized vocals across genres exhibit highly similar acoustic characteristics and weak separability. While zero-shot genre adaptation yields only marginal improvements, lightweight genre-specific continued training leads to substantial gains. MMGenre provides a standardized framework for multi-genre SVS evaluation and exposes critical challenges in achieving genre-aware singing voice synthesis.

cs.SD↗

arXiv2026

ESPnet3: Infrastructure for Scalable Speech and Audio Research in the Foundation Model Era

Recent speech research involves increasingly large datasets, complex models, and diverse experimental workflows. However, existing frameworks require substantial engineering effort to support such experiments. We present ESPnet3, a speech and audio research framework built on a modular system architecture with configuration-driven dataset composition and unified Python-based workflows. ESPnet3 introduces a DataOrganizer abstraction for flexible dataset integration and dataset sharding for memory-efficient large-scale training, while allowing recipe-specific logic through lightweight stage overrides. In OWSM pre-training experiments, ESPnet3 reduces per-epoch training time by \emph{21.1 minutes} compared to ESPnet2 and achieves \emph{>80\% GPU utilization} in multi-node training. Fine-tuning experiments show that new models and datasets can be integrated with around \emph{46 lines of additional code}. ESPnet3 will be publicly released with model checkpoints and training logs.

eess.AS↗

arXiv2026

ANCHOR: Autoregressive Non-intrusive Chunk-Ordered Refinement for Joint Multi-Resolution Speech Quality Modeling

While speech quality is typically assessed on complete utterances, streaming and generative systems require incremental estimation from partial audio. Existing predictors assume full context, degrading on prefix-constrained inputs. Extending ARECHO, we propose ANCHOR, reformulating incremental assessment as a multi-resolution autoregressive task. It models chunk- and utterance-level quality within a single decoder using dual-resolution tokens and a resolution-aware hierarchy for coarse-to-fine refinement. Experiments show substantial robustness under partial input, including a 48% PLCMOS error reduction on 2-second prefixes. Convergence analysis reveals a 4-6 s effective perceptual context horizon. A stress test further isolates structured extrapolation biases under localized corruption. Results demonstrate that hierarchical supervision improves incremental prediction and elucidates how perceptual quality accumulates over time.

eess.AS↗

arXiv2026

An Extensive Analysis of the Singing Voice Conversion Challenge 2025 Evaluation Results

We present a thorough analysis of the findings of the latest iteration of the Singing Voice Conversion Challenge, a scientific event aiming to compare and understand different voice conversion systems in a controlled environment. Compared to previous iterations which solely focused on converting the singer identity, this year we also focused on converting the singing style of the singer. To create a controlled environment and thorough evaluations, we developed a new challenge database, introduced two tasks, open-sourced baselines, and conducted large-scale crowd-sourced listening tests and objective evaluations. The challenge was run for two months and in total we evaluated 33 different systems. The results of the large-scale crowd-sourced listening test showed that top systems had comparable singer identity scores to ground truth samples. However, modeling the singing style and consequently achieving high naturalness still remains a challenge in this task, primarily due to the difficulty in modeling dynamic information in breathy, glissando, and vibrato singing styles. Further analyses of the challenge also discuss the limitations of both the traditional similarity test and the dynamic preference test in evaluating singing style similarity. Moreover, calculating Spearman's rank correlation coefficient shows that dependent objective metrics such as chroma-alignment and non-match metrics such as speaker embeddings are the most correlated to subjective scores, but are still not at a level where it could be considered as a true replacement for subjective scores.

cs.SD↗

arXiv2026

Full-Duplex-Bench-v2: A Multi-Turn Evaluation Framework for Duplex Dialogue Systems with an Automated Examiner

While full-duplex speech agents enable natural, low-latency interaction by speaking and listening simultaneously, their consistency and task performance in multi-turn settings remain underexplored. We introduce Full-Duplex-Bench-v2 (FDB-v2), a streaming framework that integrates with an automated examiner that enforces staged goals under two pacing setups (Fast vs. Slow). FDB-v2 covers four task families: daily, correction, entity tracking, and safety. We report turn-taking fluency, multi-turn instruction following, and task-specific competence. The framework is extensible, supporting both commercial APIs and open source models. When we test full-duplex systems with FDB-v2, they often get confused when people talk at the same time, struggle to handle corrections smoothly, and sometimes lose track of who or what is being talked about. Through an open-sourced, standardized streaming protocol and a task set, FDB-v2 makes it easy to extend to new task families, allowing the community to tailor and accelerate evaluation of multi-turn full-duplex systems.

eess.AS↗

arXiv2026

Reasoning Beyond Majority Vote: An Explainable SpeechLM Framework for Speech Emotion Recognition

Speech Emotion Recognition (SER) is typically trained and evaluated on majority-voted labels, which simplifies benchmarking but masks subjectivity and provides little transparency into why predictions are made. This neglects valid minority annotations and limits interpretability. We propose an explainable Speech Language Model (SpeechLM) framework that frames SER as a generative reasoning task. Given an utterance, the model first produces a transcript, then outputs both an emotion label and a concise natural-language rationale grounded in lexical and acoustic cues. Rationales are generated by a reasoning-capable teacher LLM and used as intermediate supervision, combined with majority labels during fine-tuning. Unlike prior work primarily focused on boosting classification accuracy, we aim to enhance explainability while preserving competitive performance. To this end, we complement majority-label metrics with annotator-aware scoring that credits matches with any annotator label. On MSP-Podcast v1.12, our model maintains improvements over zero-shot SpeechLM baselines, and produces rationales that human evaluators find plausible and well grounded. This demonstrates that incorporating rationale supervision offers a practical path toward interpretable SER without sacrificing predictive quality.

eess.AS↗

arXiv2026

SingMOS-Pro: An Comprehensive Benchmark for Singing Quality Assessment

Singing voice generation progresses rapidly, yet evaluating singing quality remains a critical challenge. Human subjective assessment, typically in the form of listening tests, is costly and time consuming, while existing objective metrics capture only limited perceptual aspects. In this work, we introduce SingMOS-Pro, a dataset for automatic singing quality assessment. Building on our preview version SingMOS, which provides only overall ratings, SingMOS-Pro extends the annotations of the additional data to include lyrics, melody, and overall quality, offering broader coverage and greater diversity. The dataset contains 7,981 singing clips generated by 41 models across 12 datasets, spanning from early systems to recent state-of-the-art approaches. Each clip is rated by at least five experienced annotators to ensure reliability and consistency. Furthermore, we investigate strategies for effectively utilizing MOS data annotated under heterogeneous standards and benchmark several widely used evaluation methods from related tasks on SingMOS-Pro, establishing strong baselines and practical references for future research. The dataset is publicly available at https://huggingface.co/datasets/TangRain/SingMOS-Pro.

cs.SD↗

arXiv2026

Optimizing Conversational Quality in Spoken Dialogue Systems with Reinforcement Learning from AI Feedback

Reinforcement learning from human or AI feedback (RLHF/RLAIF) for speech-in/speech-out dialogue systems (SDS) remains underexplored, with prior work largely limited to single semantic rewards applied at the utterance level. Such setups overlook the multi-dimensional and multi-modal nature of conversational quality, which encompasses semantic coherence, audio naturalness, speaker consistency, emotion alignment, and turn-taking behavior. Moreover, they are fundamentally mismatched with duplex spoken dialogue systems that generate responses incrementally, where agents must make decisions based on partial utterances. We address these limitations with the first multi-reward RLAIF framework for SDS, combining semantic, audio-quality, and emotion-consistency rewards. To align utterance-level preferences with incremental, blockwise decoding in duplex models, we apply turn-level preference sampling and aggregate per-block log-probabilities within a single DPO objective. We present the first systematic study of preference learning for improving SDS quality in both multi-turn Chain-of-Thought and blockwise duplex models, and release a multi-reward DPO dataset to support reproducible research. Experiments show that single-reward RLAIF selectively improves its targeted metric, while joint multi-reward training yields consistent gains across semantic quality and audio naturalness. These results highlight the importance of holistic, multi-reward alignment for practical conversational SDS.

cs.CL↗

arXiv2026

Do Neural Codecs Generalize? A Controlled Study Across Unseen Languages and Non-Speech Tasks

This paper investigates three crucial yet underexplored aspects of the generalization capabilities of neural audio codecs (NACs): (i) whether NACs can generalize to unseen languages during pre-training, (ii) whether speech-only pre-trained NACs can effectively generalize to non-speech applications such as environmental sounds, music, and animal vocalizations, and (iii) whether incorporating non-speech data during pre-training can improve performance on both speech and non-speech tasks. Existing studies typically rely on off-the-shelf NACs for comparison, which limits insight due to variations in implementation. In this work, we train NACs from scratch using strictly controlled configurations and carefully curated pre-training data to enable fair comparisons. We conduct a comprehensive evaluation of NAC performance on both signal reconstruction quality and downstream applications using 11 metrics. Our results show that NACs can generalize to unseen languages during pre-training, speech-only pre-trained NACs exhibit degraded performance on non-speech tasks, and incorporating non-speech data during pre-training improves performance on non-speech tasks while maintaining comparable performance on speech tasks.

cs.SD↗

arXiv2026

IKFST: IOO and KOO Algorithms for Accelerated and Precise WFST-based End-to-End Automatic Speech Recognition

End-to-end automatic speech recognition has become the dominant paradigm in both academia and industry. To enhance recognition performance, the Weighted Finite-State Transducer (WFST) is widely adopted to integrate acoustic and language models through static graph composition, providing robust decoding and effective error correction. However, WFST decoding relies on a frame-by-frame autoregressive search over CTC posterior probabilities, which severely limits inference efficiency. Motivated by establishing a more principled compatibility between WFST decoding and CTC modeling, we systematically study the two fundamental components of CTC outputs, namely blank and non-blank frames, and identify a key insight: blank frames primarily encode positional information, while non-blank frames carry semantic content. Building on this observation, we introduce Keep-Only-One and Insert-Only-One, two decoding algorithms that explicitly exploit the structural roles of blank and non-blank frames to achieve significantly faster WFST-based inference without compromising recognition accuracy. Experiments on large-scale in-house, AISHELL-1, and LibriSpeech datasets demonstrate state-of-the-art recognition accuracy with substantially reduced decoding latency, enabling truly efficient and high-performance WFST decoding in modern speech recognition systems.

cs.SD↗

arXiv2025

SingingSDS: A Singing-Capable Spoken Dialogue System for Conversational Roleplay Applications

With recent advances in automatic speech recognition (ASR), large language models (LLMs), and text-to-speech (TTS) technologies, spoken dialogue systems (SDS) have become widely accessible. However, most existing SDS are limited to conventional spoken responses. We present SingingSDS, a cascaded SDS that responds through singing rather than speaking, fostering more affective, memorable, and pleasurable interactions in character-based roleplay and interactive entertainment scenarios. SingingSDS employs a modular ASR-LLM-SVS pipeline and supports a wide range of configurations across character personas, ASR and LLM backends, SVS models, melody sources, and voice profiles, tailored to different needs in terms of latency, quality, and musical style. SingingSDS is available as a plug-and-play web demo, featuring modular, open-source code that supports customization and extension. Demo: https://huggingface.co/spaces/espnet/SingingSDS. Code: https://github.com/SingingSDS/SingingSDS.

cs.SD↗

arXiv2025

Robust Training of Singing Voice Synthesis Using Prior and Posterior Uncertainty

Singing voice synthesis (SVS) has seen remarkable advancements in recent years. However, compared to speech and general audio data, publicly available singing datasets remain limited. In practice, this data scarcity often leads to performance degradation in long-tail scenarios, such as imbalanced pitch distributions or rare singing styles. To mitigate these challenges, we propose uncertainty-based optimization to improve the training process of end-to-end SVS models. First, we introduce differentiable data augmentation in the adversarial training, which operates in a sample-wise manner to increase the prior uncertainty. Second, we incorporate a frame-level uncertainty prediction module that estimates the posterior uncertainty, enabling the model to allocate more learning capacity to low-confidence segments. Empirical results on the Opencpop and Ofuton-P, across Chinese and Japanese, demonstrate that our approach improves performance in various perspectives.

cs.SD↗

arXiv2025

Adapting Speech Language Model to Singing Voice Synthesis

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.

cs.SD↗

arXiv2025

Balancing Speech Understanding and Generation Using Continual Pre-training for Codec-based Speech LLM

Recent advances in speech language models (LLMs) have extended textual LLMs to the speech domain, but balancing speech understanding and generation remains challenging, especially with codec-based representations. We propose a continual pre-training (CPT) framework that adapts a textual LLM to handle codec-discretized speech, mitigating modality mismatch and preserving linguistic reasoning. Our unified model supports both understanding and generation, achieving strong results across ASR, TTS, S2T-Trans, and S2S-Trans. Notably, we present the first end-to-end, single-pass S2S-Trans system using only neural codec tokens, without intermediate transcriptions, translations, or semantic tokens. CPT proves essential for cross-modal alignment and task generalization, making it a powerful tool for building robust, unified speech LLMs.

eess.AS↗

arXiv2025

PURE Codec: Progressive Unfolding of Residual Entropy for Speech Codec Learning

Neural speech codecs have achieved strong performance in low-bitrate compression, but residual vector quantization (RVQ) often suffers from unstable training and ineffective decomposition, limiting reconstruction quality and efficiency. We propose PURE Codec (Progressive Unfolding of Residual Entropy), a novel framework that guides multi-stage quantization using a pre-trained speech enhancement model. The first quantization stage reconstructs low-entropy, denoised speech embeddings, while subsequent stages encode residual high-entropy components. This design improves training stability significantly. Experiments demonstrate that PURE consistently outperforms conventional RVQ-based codecs in reconstruction and downstream speech language model-based text-to-speech, particularly under noisy training conditions.

cs.SD↗

arXiv2025

CartoonSing: Unifying Human and Nonhuman Timbres in Singing Generation

Singing voice synthesis (SVS) and singing voice conversion (SVC) have achieved remarkable progress in generating natural-sounding human singing. However, existing systems are restricted to human timbres and have limited ability to synthesize voices outside the human range, which are increasingly demanded in creative applications such as video games, movies, and virtual characters. We introduce Non-Human Singing Generation (NHSG), covering non-human singing voice synthesis (NHSVS) and non-human singing voice conversion (NHSVC), as a novel machine learning task for generating musically coherent singing with non-human timbral characteristics. NHSG is particularly challenging due to the scarcity of non-human singing data, the lack of symbolic alignment, and the wide timbral gap between human and non-human voices. To address these challenges, we propose CartoonSing, a unified framework that integrates singing voice synthesis and conversion while bridging human and non-human singing generation. CartoonSing employs a two-stage pipeline: a score representation encoder trained with annotated human singing and a timbre-aware vocoder that reconstructs waveforms for both human and non-human audio. Experiments demonstrate that CartoonSing successfully generates non-human singing voices, generalizes to novel timbres, and extends conventional SVS and SVC toward creative, non-human singing generation.

cs.SD↗

arXiv2025

BSCodec: A Band-Split Neural Codec for High-Quality Universal Audio Reconstruction

Neural audio codecs have recently enabled high-fidelity reconstruction at high compression rates, especially for speech. However, speech and non-speech audio exhibit fundamentally different spectral characteristics: speech energy concentrates in narrow bands around pitch harmonics (80-400 Hz), while non-speech audio requires faithful reproduction across the full spectrum, particularly preserving higher frequencies that define timbre and texture. This poses a challenge: speech-optimized neural codecs suffer degradation on music or sound. Treating the full spectrum holistically is suboptimal: frequency bands have vastly different information density and perceptual importance by content type, yet full-band approaches apply uniform capacity across frequencies without accounting for these acoustic structures. To address this gap, we propose BSCodec (Band-Split Codec), a novel neural audio codec architecture that splits the spectral dimension into separate bands and compresses each band independently. Experimental results demonstrate that BSCodec achieves superior reconstruction over baselines across sound and music, while maintaining competitive quality in the speech domain, when trained on the same combined dataset of speech, music and sound. Downstream benchmark tasks further confirm that BSCodec shows strong potential for use in downstream applications.

eess.AS↗