arXiv subjects

David Harwath

Publications and source records attributed to David Harwath.

At least 19 recordsLinked to original sources

arXiv2026

Learning Timeline-Preserving Speech-to-Speech Translation

Speech-to-speech translation (S2ST) increasingly preserves what is said and how it sounds, but not necessarily when speech and silence occur. For dubbing and other temporally grounded applications, matching total duration is not enough: speech and pauses must also occur at corresponding moments. We formulate timeline-preserving S2ST, which aims to preserve the source endpoint and salient internal pauses on a shared absolute timeline. We present TransDubber, an end-to-end autoregressive system that learns this temporal structure directly from aligned target waveforms, without inference-time frame-level timing control or post-hoc temporal correction. To provide this supervision, we construct 642K exactly duration-matched English-Chinese speech pairs by translating and synthesizing target speech on source-defined temporal scaffolds. Training combines a progressive supervised curriculum with verifier-guided reinforcement learning to balance translation quality and temporal alignment. On a 4,353-utterance test set spanning read and spontaneous speech, TransDubber increases the proportion of translations within 2% of the source duration from 25.6% to 61.5% and pause-alignment F1 from 0.118 to 0.550 compared with UniSS, while maintaining competitive translation quality. Our analysis shows that aligned supervision establishes timeline following, while reinforcement learning refines temporal precision and partially offsets the translation-quality cost of strict synchronization. Together, these results demonstrate the effectiveness of our data and training framework for timeline-preserving S2ST. Audio samples, code, datasets, and pretrained models are available at https://zhishengzheng.com/transdubber/.

eess.AS↗

arXiv2026

Discovering Interpretable Acoustic Features in Autoregressive Voice-Cloning TTS Models

Voice-cloning TTS models synthesize an unseen speaker's voice from brief speech prompts, but it remains unclear which acoustic properties they represent and use. We introduce a feature-based interpretability framework for autoregressive voice-cloning TTS models, analyzing features from sparse autoencoders (SAEs), PCA, and random directions via sparse probing, steering, and quantitative auto-interpretability. We focus on Qwen3-TTS and extend probing and steering to $5$ additional models to test generalization across architectures. These models encode concepts like pitch, energy, noise, and accent, but sparse prediction and steering are effective only for some concepts. Strong probing does not always translate to steering or vice-versa. SAEs are useful for sparse probing of human-labelled concepts and novel-concept discovery, while PCA is competitive for dense probing and steering. Auto-interpretability finds $684$ automatically validated SAE features in Qwen3-TTS covering speaker clusters, speaking style, and recording conditions beyond known concepts. Overall, we perform a comprehensive interpretability study of autoregressive voice-cloning TTS models.

eess.AS↗

arXiv2026

Partially Edited Audio: Generation, Defense, and Open Problems

Recent advances in generative audio models facilitate increasingly flexible and selective editing of linguistic content and paralinguistic, extralinguistic, and other attributes. This paradigm, named partially edited audio, differs from full audio generation as only portions of the real recording are edited. This technique can serve both beneficial and harmful applications. We therefore present a concise survey from both generation and defense perspectives, focusing on audio, as a contribution to the ICASSP 2027 special session entitled "Partially Edited Media: Perspectives from Synthesis and Defense." We first define partially edited audio and argue that what counts as ``fake'' or harmful is an application-dependent judgment, not a fixed property of the edit. We then survey audio editing by editing target, approaches, and evaluation, and then defense methods by detection, localization, diarization, and interpretability, summarizing representative benchmarks and toolkits. Finally, we discuss future directions and open problems.

cs.SD↗

arXiv2026

Encoder-Free Multimodal LLMs Learn Implicit Encoders

We empirically analyze two recently proposed encoder-free multimodal LLMs, Gemma-4 12B and Inkling-Small. Unlike previous multimodal LLMs, these models do not use modality-specific encoders to extract tokenized embeddings representing audio and image input; instead, they directly feed windows of samples, pixels, or spectral features to the Transformer language model. We analyze whether the early layers of these models learn to function similarly to dedicated audio and visual encoders. We find that these early layers do in fact act as ``implicit encoders,'' encoding audio and visual inputs in a way that is largely invariant to their cross-modal context. Furthermore, when performing multimodal reasoning tasks, we show that independently encoding the modalities throughout these early model layers achieves the same accuracy as jointly encoding them from the first layer onwards. Finally, we use layer-wise probing to analyze the salient information captured by the implicit encoders and find their behavior to be similar to several popular encoder models.

cs.CL↗

arXiv2026

Hear, Localize, and Reason: Spatially Aware Scene Understanding for Audio-Visual LLMs

Audio-visual large language models (AV-LLMs) have made strong progress in multimodal understanding, yet most treat audio as monaural semantic content, limiting their ability to reason about where sounds originate and how they relate to the visual scene. Recent spatial audio-visual studies address parts of this problem, but often focus on specific abilities, such as spatial correspondence or direction and distance reasoning. In this work, we introduce HLR-AVSceneQA, a comprehensive benchmark with a hear-localize-reason hierarchy that evaluates whether models can recognize what is heard, localize sound sources in egocentric and allocentric views, and use these localized sources to reason about how they relate to the visual scene. We further introduce HLR-LLM, which augments a strong audio-visual foundation model with a binaural spatial audio branch alongside its monaural semantic audio pathway. We train HLR-LLM with a curriculum that incorporates final-stage chain-of-thought supervision for multi-step spatial reasoning. Experiments show that existing AV-LLMs remain limited in spatial grounding and relational reasoning, while HLR-LLM substantially improves spatial audio-visual scene understanding without compromising semantic audio-visual perception.

cs.CV↗

arXiv2026

Phone Segmentation and Recognition through Phonological Activation Mapping

Phone segmentation and recognition are inherently related tasks, yet modern approaches typically model them separately. We argue that phonetic structure is already latent in the representations of self-supervised speech models (S3Ms), so one only needs to steer them to solve both tasks. We leverage S3M-based Phonological Activation Mapping (SPAM), which maps each S3M representation frame to a vector of phonological feature activations, such as voicing and nasality. On top of SPAM, we introduce two simple but effective lightweight, gradient-descent-free prediction heads: a recognition head and a segmentation head. Our method requires less than a minute of phonetic transcriptions, and the recognition head generalizes to phones unseen during training. Across a diverse range of datasets, our approach attains strong segmentation performance.

eess.AS↗

arXiv2026

RUBATO: Decoupled Clocks Allow Full-Duplex Models to Think as Fast as They Can

Recent Speech Full-Duplex (FDX) models are capable of engaging in naturalistic conversation by listening and speaking at the same time. Building on this foundation, recent architectures introduce internal thinking streams to enhance their intelligence. To guarantee these streams process successfully in real time, existing methods enforce a fixed decoding rate, which assumes a hardware profile capable of a specific latency. We argue this static approach is severely suboptimal. Because internal thinking is intrinsically asynchronous, an ideal FDX model should sustain real-time conversational flow while allowing each stream to decode at its own hardware-adaptive rate. Thinking streams must process as fast as the hardware allows to deliver accurate answers faster. To solve this, we propose RUBATO, a novel multistream positional encoding that assigns independent clocks to entirely decouple the streams. This structural isolation empowers the model to zero-shot interpolate across any inference decoding speed. It drastically simplifies dataset construction by requiring training on only two extreme decoding speeds while smoothly adapting to any intermediate rate during deployment. RUBATO demonstrates consistent robustness across five spoken reasoning tasks. On Spoken GSM8K, a baseline model collapses to 1.7% accuracy at interpolated speeds. Conversely, RUBATO-Full sustains 48.7% to 51.1% accuracy across all rates, leveraging faster hardware to achieve a 6.5x shorter time-to-answer with a constant time-to-first-response. This advantage holds on BIG-Bench Audio and amplifies significantly during complex full-duplex interruptions. Specifically, when a user interjects while the model is actively speaking and reasoning, RUBATO-Full achieves 40.0% to 45.1% accuracy, whereas the baseline fails completely.

eess.AS↗

arXiv2026

ASSET: Adaptive Speech Segmentation via Reinforcement Learning

Speech large language models (LLMs) are often built using pretrained acoustic encoders as an audio interface to an LLM decoder. However, LLMs scale poorly with input sequence length, while acoustic encoders tend to produce substantially longer sequences than typical textual inputs. This motivates learning to compress encoder outputs while preserving task-oriented information. To this end, we introduce ASSET, a reinforcement learning (RL)-based framework that learns to adaptively compress encoder outputs by first partitioning them into variable-duration segments and then pooling each segment into a continuous audio token. We use group-relative policy optimization (GRPO) to train a segmenter with task performance feedback and a token-rate penalty. In automatic speech recognition (ASR) experiments, ASSET achieves performance comparable to a transcript-assisted character-boundary oracle with approximately 25% fewer audio tokens on English LibriSpeech, while improving on a phone-boundary oracle at a similar token rate on Mandarin AISHELL. Exploratory experiments on non-ASR tasks further illustrate how learned segmentation changes when the framework is adapted to tasks beyond transcription.

cs.CL↗

arXiv2026

[b] = [d] - [t] + [p]: Self-supervised Speech Models Discover Phonological Vector Arithmetic

Self-supervised speech models (S3Ms) are known to encode rich phonetic information, yet how this information is structured remains underexplored. We conduct a comprehensive study across 96 languages to analyze the underlying structure of S3M representations, with particular attention to phonological vectors. We first show that there exist linear directions within the model's representation space that correspond to phonological features. We further demonstrate that the scale of these phonological vectors correlate to the degree of acoustic realization of their corresponding phonological features in a continuous manner. For example, the difference between [d] and [t] yields a voicing vector: adding this vector to [p] produces [b], while scaling it results in a continuum of voicing. Together, these findings indicate that S3Ms encode speech using phonologically interpretable and compositional vectors, demonstrating phonological vector arithmetic. All code and interactive demos are available at https://github.com/juice500ml/phonetic-arithmetic .

eess.AS↗

arXiv2026

Voices as Handles: Reasoning about Speaker Identity with Frozen Text LLMs

Multi-user voice agents must track who said what across dialogue sessions. Text LLMs are attractive backbones for such agents, but transcripts alone do not expose acoustic speaker identity, leaving the model without a persistent reference for linking information to speakers across sessions. We address this gap by introducing Speaker Handles, soft-token representations that expose acoustic speaker identity to a frozen text LLM for cross-session speaker-dependent reasoning. A three-stage curriculum trains a lightweight projector, with fewer than 0.1% of the backbone's parameters, to map speaker embeddings into these handles. Establishing whether the resulting handles truly support cross-session speaker-dependent reasoning is challenging with existing benchmarks because textual cues can partially reveal fact ownership. We therefore present SpeakerBind, a controlled shared-agent benchmark in which overlapping facts across users require correct cross-session speaker attribution. Speaker Handles achieve 97.40-98.36% accuracy on VoxCeleb1 and 70.40% on SpeakerBind, close to the 71.88% topline. These results show that the proposed Speaker Handles provide an efficient way to integrate acoustic speaker identity into frozen text LLMs for speaker-content reasoning.

eess.AS↗

arXiv2026

SPADE: A Multilingual Dataset for Speech Partial Deepfake Detection and Localization

Recent improvements in voice-cloning speech generation systems raise concerns about misuse by malicious actors to impersonate others and spread misinformation. Detecting such tampering is difficult, since deepfakes in the wild may be created by different generative models in a wide range of languages. Furthermore, the speech audio may also only be partially modified, presenting a different and potentially more challenging task than detecting fully-synthetic speech waveforms. To enable further research in this direction, we propose a multilingual dataset for detection and localization of partially edited speech samples. Our dataset includes speech in 12 languages, generated by up to five systems per language, and includes both a training set as well as an evaluation benchmark. To showcase the utility of our proposed dataset, we train localization models of existing architectures and study generalization across three axes: across different languages, across different speech synthesis and editing systems, and across different acoustic environments. Our results show that localization models almost always generalize poorly to speech edited by systems not seen during training. On the other hand, generalization to edited speech in unseen languages still degrades performance but to a lesser extent. We also augment our testing sets with noise to evaluate generalization across acoustic environments, and find that performance of localization models degrade significantly when tested on different acoustic conditions. All together, our results imply that existing deepfake speech detection methods are insufficient for reliably detecting edit-based speech deepfakes in various scenarios unseen during training. SPADE is publicly available on HuggingFace.

eess.AS↗

arXiv2026

Quantifying Consonant Contributions to Word Intelligibility via Acoustic Masking

Consonants contribute unequally to whether a word is understood. Given the limited time available for therapy, ranking consonants by contribution to intelligibility helps prioritize intervention targets in motor speech disorders. However, measuring this contribution relies on perceptual studies that are difficult to scale. This paper presents a scalable method that measures consonant contribution using acoustic masking. We silence one consonant at a time in an isolated word and test whether an automatic speech recognition (ASR) model still recognizes the word. We define a consonant's contribution score as the proportion of its masked instances for which the word becomes misrecognized, which we refer to as the mask-induced misrecognition rate (MMR). We relate MMR to two linguistic factors previously reported to correlate with consonant contribution, namely phoneme frequency and functional load. We apply this analysis across four languages, English, Spanish, German, and Czech, using three ASR architectures, MMS (encoder-only), Whisper (encoder-decoder), and Qwen3-ASR (LLM-based). Using partial Spearman correlations, we find that phoneme frequency correlates negatively with MMR while functional load correlates positively. In other words, more frequent consonants are less disruptive when masked, whereas consonants carrying more lexical contrast are more disruptive. Further cross-language analysis shows that consonant rankings agree only partially across languages, indicating that consonant contribution is language-dependent.

cs.CL↗

arXiv2026

MP-Bench: Evaluating Voice Agents as a Multiparty Conversation Participant

Conversational voice agents have advanced significantly, offering increasingly natural human-machine interactions through both cascaded and end-to-end architectures. However, while recent benchmarks extensively evaluate dyadic interactions and passive audio comprehension, they largely overlook a prevalent real-world scenario: multi-party conversations. Evaluating agents in these settings is fundamentally more challenging than in dyadic interactions due to the exponentially greater conversational complexity. For voice agents to integrate seamlessly into human group dynamics, they must not only generate contextually appropriate responses but also demonstrate a nuanced understanding of open turn-taking. To address this gap, we introduce Multiparty Bench (MP-Bench), the first benchmark specifically designed to objectively evaluate conversational speech systems as active participants within multi-party contexts. MP-Bench assesses agent behavior along two primary dimensions: turn-taking awareness and response appropriateness. Additionally, we incorporate comprehension-based question-answering tasks as a complementary evaluation. By benchmarking 12 voice agents, we find that real-time voice agents stay at or below 22% on multiparty comprehension and remain near chance on multiparty turn-taking, exposing an open challenge for real-time voice agents under multiparty scenario.

eess.AS↗

arXiv2026

RetroThinker: Enabling Retrospective Thinking in Speech LLMs

Speech large language models (SpeechLLMs) offer reduced latency and retain paralinguistic nuances that are typically lost in cascaded automatic speech recognition (ASR) and text-based LM architectures. However, they continue to lag behind text-only LLMs on complex reasoning tasks, while real-time spoken interaction imposes strict latency constraints. Although prior works employ Chain-of-Thought (CoT) and concurrent reasoning to enhance reasoning capabilities without inducing prohibitive delays, an inherent accuracy-latency trade-off persists. In this paper, we investigate whether a streaming SpeechLLM can dynamically revise its reasoning traces on the fly. We introduce RetroThinker, a multi-stage post-training framework that equips the Moshi model to self-verify and forward-correct CoT steps during inference. RetroThinker combines supervised fine-tuning (SFT) on curated retrospective thinking data with length-based direct preference optimization (DPO) to optimize retrospective during early reasoning (i.e., reasoning concurrently while the user speaks). Evaluated on the GSM8K benchmark, RetroThinker significantly improves the accuracy-latency trade-off over non-retrospective baselines, achieving an 11% absolute accuracy gain at a comparable latency.

eess.AS↗

arXiv2026

Unsupervised Speech Recognition at the Syllable Level

Training speech recognizers with unpaired speech and text -- known as unsupervised speech recognition (UASR) -- is a crucial step toward extending ASR to low-resource languages in the long-tail distribution and enabling multimodal learning from non-parallel data. However, existing approaches based on phones often rely on costly resources such as grapheme-to-phoneme converters (G2Ps) and struggle to generalize to languages with ambiguous phoneme boundaries due to training instability. In this paper, we address both challenges by introducing a syllable-level UASR framework based on masked language modeling, which avoids the need for G2P and the instability of GAN-based methods. Our approach achieves up to a 40\% relative reduction in character error rate (CER) on LibriSpeech and generalizes effectively to low-resource languages that have remained particularly difficult for prior methods. Code is publicly available\footnote{https://github.com/cactuswiththoughts/SylCipher}.

eess.AS↗

arXiv2026

BAT: Learning to Reason about Spatial Sounds with Large Language Models

Spatial sound reasoning is a fundamental human skill, enabling us to navigate and interpret our surroundings based on sound. In this paper we present BAT, which combines the spatial sound perception ability of a binaural acoustic scene analysis model with the natural language reasoning capabilities of a large language model (LLM) to replicate this innate ability. To address the lack of existing datasets of in-the-wild spatial sounds, we synthesized a binaural audio dataset using AudioSet and SoundSpaces 2.0. Next, we developed SpatialSoundQA, a spatial sound-based question-answering dataset, offering a range of QA tasks that train BAT in various aspects of spatial sound perception and reasoning. The acoustic front end encoder of BAT is a novel spatial audio encoder named Spatial Audio Spectrogram Transformer, or Spatial-AST, which by itself achieves strong performance across sound event detection, spatial localization, and distance estimation. By integrating Spatial-AST with LLaMA-2 7B model, BAT transcends standard Sound Event Localization and Detection (SELD) tasks, enabling the model to reason about the relationships between the sounds in its environment. Our experiments demonstrate BAT's superior performance on both spatial sound perception and reasoning, showcasing the immense potential of LLMs in navigating and interpreting complex spatial audio environments.

eess.AS↗

arXiv2026

CtrlSpeech: Coarse-to-Fine Control for Expressive Speech Synthesis

Recent Text-To-Speech (TTS) systems have achieved strong naturalness and zero-shot voice cloning performance, but fine-grained control of expressive speech at the word or phoneme level remains challenging. We propose CtrlSpeech, a controllable, expressive TTS framework with coarse-to-fine control. Built on the DiTAR architecture, CtrlSpeech combines global speaker conditioning with phone-aligned pitch, loudness, and duration signals, enabling localized prosodic control while preserving the target speaker's timbre. This design allows users to adjust expressive attributes at a fine temporal granularity, making speech refinement more flexible and controllable. Experimental results show that CtrlSpeech achieves competitive zero-shot TTS performance and improves controllability over expressive attributes, demonstrating its effectiveness for flexible and practical expressive speech synthesis.

eess.AS↗

arXiv2026

Adapting Self-Supervised Speech Representations for Cross-lingual Dysarthria Detection in Parkinson's Disease

The limited availability of dysarthric speech data makes cross-lingual detection an important but challenging problem. A key difficulty is that speech representations often encode language-dependent structure that can confound dysarthria detection. We propose a representation-level language shift (LS) that aligns source-language self-supervised speech representations with the target-language distribution using centroid-based vector adaptation estimated from healthy-control speech. We evaluate the approach on oral DDK recordings from Parkinson's disease speech datasets in Czech, German, and Spanish under both cross-lingual and multilingual settings. LS substantially improves sensitivity and F1 in cross-lingual settings, while yielding smaller but consistent gains in multilingual settings. Representation analysis further shows that LS reduces language identity in the embedding space, supporting the interpretation that LS removes language-dependent structure.

cs.CL↗