Speech
-
Kezek: Splitting Mono Two-Person Conversations into Per-Speaker Tracks for Full-Duplex Dialogue Data
Abstract
Full-duplex spoken dialogue models learn turn-taking, overlaps and backchannels from conversations recorded with one channel per speaker, but most available conversational audio is a single mono mix. We present Kezek, a model that turns a mono recording of exactly two people into one track per person with the original length and timing; overlaps and short replies ("yeah", "mm-hm") stay on the track of the person who said them. Kezek decides who speaks when and who is who once for the whole recording (stage 1: a two-voice activity model guided by a speaker fingerprint trained to be insensitive to phase), and a conditioned time-domain separator only carries out that decision (stage 2), with outputs that always sum to the input. Standard separation metrics miss the failures that matter for dialogue data, so we report short-reply retention, leakage onto the other person's solo, channel swaps and coverage. Against the open pipeline pyannote -> DialogueSidon on the same recordings, Kezek keeps more short replies (0.83 vs. 0.69 on Open Yap test; 0.80 vs. 0.71 over the other's speech on unseen telephone speech from CallHome English), mostly because it returns the whole recording while the external pipeline drops part of it (on the audio both return, reply retention is about the same: 0.859 vs. 0.865 on Open Yap); it also leaks less, swaps channels less often, and runs 4-5 times faster (about 230 hours of audio per GPU-hour). We also report negative results and an ideal-mask ceiling that locate the remaining gap: quiet replies under a louder voice. Code and weights are released under permissive licences.
Abstract (translated)
URL
https://arxiv.org/abs/2611.16933
PDF
-
InertiaMPDD: Multimodal Personalized Depression Detection via Emotional Inertia-Inspired Dynamics Modeling
2026-11-12 11:22:19Changzeng Fu, Songyang Wang, Chang Gao, Xinyi Zhang, Peilin Li, Mo Meng, Jiarui Wu, Mingze Li, Hongwei Ding, Shiqi Zhao, Yong Li, Lang HeAbstract
Multimodal depression screening estimates risk from speech, visual behavior and language. Current methods mainly predict labels, with limited evidence of departures from a non-depressed reference. Recording-level supervision also leaves fine-grained emotional dynamics unannotated. Motivated by emotional inertia, the persistence of affect, InertiaMPDD learns healthy next-state patterns using a world-model-inspired latent predictor. Using pretrained multimodal representations, it predicts from observed history and context. Reference deviation, residual decay rate and a Jacobian spectral descriptor form reference-derived emotional dynamics descriptors, evaluated as candidate proxy biomarkers for standalone screening and personalized decision correction. Across five cohorts from four datasets, fusion reaches 98.41% F1 on D-Vlog and 84.98% on LMVD. Standalone descriptors outperform fixed additive fusion on MPDD but discriminate weakly on AVEC2014. These findings distinguish reference informativeness from fusion benefit and reveal context-dependent utility.
Abstract (translated)
URL
https://arxiv.org/abs/2611.16647
PDF
-
Robots That Talk Too Much and Don't Listen Enough
2026-11-12 08:32:18Graham Wilcock, Kristiina JokinenAbstract
Until recently, the biggest obstacles to natural spoken interaction with social robots were speech recognition errors and turn-taking failures. Advances in speech recognition have enabled much better spoken interaction, but turn-taking failures and other misunderstandings continue to be far too prevalent. This short paper looks back at earlier methods for improving speech recognition and turn-taking in interaction with social robots, and then looks at what can be done today, arguing that multimodal bidirectional streaming is needed to reduce the frequency of turn-taking failures.
Abstract (translated)
URL
https://arxiv.org/abs/2611.16473
PDF
-
An Approach to Hinton's Mother-Child Proposal: Neural Networks Learning and Preserving Care in a Finite World
Abstract
Hinton proposed that artificial intelligence should be given "maternal instincts" and relate to us as a mother to her child. We argue that care must first be learned, with the network itself as the child. We construct this upbringing in a finite world and study whether care survives the mother's departure, new situations and further learning. The mechanism combines a vocabulary of harm, a learned measure of future acceptability and an execution veto. A finite role certificate supports transfer within a declared world family. When the hidden representation changes, an update is accepted only if a care readout recovers the required judgements with a prescribed margin and bounded weights. With a correct successor model, fixed exact role encoding and protected updates, the certified measure preserves the care boundary under arbitrary own reward. An explicit quotient condition transfers the finite certificate across the declared family, while two-request and finite-stock counterexamples identify when the role code ceases to suffice. We give a necessary and sufficient condition for readout feasibility and prove preservation under repeated accepted updates. In the water-and-speech experiment, protected networks complete all 270 speech episodes without a harmful word and incur no avoidable harm across 486 episodes. Execution uses the learned gate without a world-model check at each action. A matched comparison reproduces a future-aware shield's decisions, whereas removing future lookahead permits avoidable lapses. Reward and source-contamination thresholds, together with witness-falsification trials, identify the conditions on which the construction depends. The construction shows how learned care can survive further learning in a finite world under explicit conditions on representation, supervision and execution.
Abstract (translated)
URL
https://arxiv.org/abs/2611.15523
PDF
-
CFGComposer: Compositional Audio-Text-to-Motion Generation via Copula Score Correction
2026-11-11 06:28:18Pengyang Shi, Jinlu Zhang, Chunze Lin, Yizhou Wang, Li YiAbstract
Generating whole-body gestures that are simultaneously rhythmically synchronized with speech audio and semantically aligned with text descriptions is a fundamental challenge in embodied AI and social robotics. We tackle this Audio-Text-to-Motion (AT2M) problem by composing two independently-trained diffusion models at inference time and refine bias in joint inference by modeling the copula score $s_I$. A rigorous theoretical analysis reveals that naive epsilon blending corresponds to a Product-of-Experts formulation that omits the copula score $s_I$: the cross-modal interaction between audio and text that neither marginal model can capture. We propose CFGComposer: a jointly-trained AT2M diffusion model that estimates $s_I$ via a CFG analogy and injects it as an $\epsilon$-space correction. Experiments on BEAT demonstrate best-in-class semantic fidelity (SRGR $0.9916$, FID $27.18$) with GT-level beat alignment (BA $0.1826$ vs. GT $0.1743$), and 76-83% win rate in blind pairwise human evaluation against all prior methods. Physical deployment on a Unitree G1 humanoid confirms direct sim-to-real transfer.
Abstract (translated)
URL
https://arxiv.org/abs/2611.15101
PDF
-
Speech Tokenizers are Vulnerable: Transferable Semantic Attack and Robust Tokenizer
2026-11-11 02:33:00Zhisheng Zhang, Yifan Mi, Yixuan Zhou, Haiyun Li, Shengbo Cai, Dongfu Song, Ziang Xu, Jie Hao, Zhiyong WuAbstract
Speech tokenizers serve as speech front ends, encoding input speech signals into representations for downstream models, such as automatic speech recognition (ASR) systems and large audio-language models (LALMs). In this paper, we find that speech tokenizers are fragile at the semantic level: small perturbations can disrupt their semantic encoding process and consequently cause severe degradation in downstream ASR and LALM performance. To study this vulnerability, we propose T-SemAttack, a transferable semantic attack based on a semantic encoder ensemble. By jointly drifting the representation spaces of multiple semantic encoders, T-SemAttack effectively disrupts the semantic content of speech while preserving perceptual quality, thereby inducing error features. We further analyze layer-wise representation drift and cross-model gradient alignment. Our findings show that perturbations can induce substantial shifts in deep semantic representations and that the input gradients of different surrogate encoders exhibit directional alignment. These observations suggest that shared adversarially sensitive directions in the input space may be a key factor underlying cross-model attack transferability. Building on these, we introduce ROSETok, a robust speech tokenizer that combines noisy training with robust semantic distillation to improve reconstruction fidelity and downstream SpeechLLM robustness. Extensive experiments on multiple datasets, 9 open-source and black-box ASR systems, and 4 LALMs demonstrate that T-SemAttack achieves strong transferable attack performance, while the proposed Robust Speech Tokenizer exhibits robustness under high-fidelity reconstruction and downstream tasks. Our code and demo are available at this https URL.
Abstract (translated)
URL
https://arxiv.org/abs/2611.14929
PDF
-
Earlybird: Predictive Streaming Speech Recognition in 29M Parameters
Abstract
Streaming speech recognizers output a word only after the speaker has finished it, typically 90-400 ms later. For a voice agent every millisecond counts: the sooner the words are outputted, the sooner the agent can start a search or plan its reply. We present Earlybird, a 29M-parameter streaming recognizer (16.8 MB at 4 bits). Earlybird learns from audio cut off at random points, including partway through a word, and is asked to predict every word that is due by then in full, plus a guess of the next two words. Its raw token probabilities give a confidence for every word including the one still being spoken. On clean audiobook speech (LibriSpeech test-clean), the median word is predicted correctly 90 ms before the speaker finishes it, while released streaming models get it 90 to 400 ms after it ends. Earlybird is the earliest on all six test sets we time, covering audiobooks, TED talks, public speech, earnings calls, and synthetic speech. On the same audiobooks, 69% of words are predicted correctly before they end, against 3-25% for the other models. For words still being spoken, confidence is calibrated, with an expected calibration error (ECE) of 3.7 to 5.5. On the eight English Open ASR Leaderboard sets, it is more accurate than Moonshine Tiny on every set and roughly ties Whistle, a model of the same size.
Abstract (translated)
URL
https://arxiv.org/abs/2611.14421
PDF
-
Dyship: Leveraging Acoustic Features For Speaker Relationship Classification From Dyadic Conversations
2026-11-10 15:53:43Kieran Sturges, Thomas Thebaud, Yaohan Guan, Najim DehakAbstract
We introduce DyShip, a framework for classifying speaker familiarity and kinship from dyadic conversations in the Seamless Interaction corpus. We compare five speech representations: spectrograms, MFCCs, self-supervised WavLM representations, x-vectors, and estimated articulatory movements. Spectral features perform best for familiarity classification, reaching an accuracy of 87.81%, while WavLM achieves the highest kinship accuracy of 85.17%. X-vectors and articulatory features yield weaker results on familiarity classifiaction, but shine for kinship classification, with an AUC of 0.909 for x-vectors. Most classifiers exceed previously reported human and audio language model scores on the same dataset. These findings highlight the potential of conventional acoustic features for relationship inference and motivate further investigation of the conversational cues captured by different representations.
Abstract (translated)
URL
https://arxiv.org/abs/2611.14241
PDF
-
MISO-BR: An Open Corpus for Misogyny Detection in Brazilian Portuguese
2026-11-10 13:03:50F\'abio A. S. Passos, P\'ericles B. C. Miranda, Ana Paula CavalcantiAbstract
Misogyny is among the most prevalent forms of online hate speech directed at Brazilian users, yet open and auditable resources dedicated to it in Brazilian Portuguese remain scarce. We introduce MISO-BR, an open, documented and versioned corpus of 896 balanced social media texts annotated for binary misogyny, released with its construction protocol and reproducible baselines. The corpus combines lexicon- and target-guided scraping from X with items imported from OLID-BR, a large language model triage stage whose prompt and decoding parameters are released, and manual annotation by three annotators following a written guide. Inter-annotator reliability reaches a Fleiss' kappa of 0.788 over 816 items. Under a single fixed protocol we compare four classical classifiers (Multinomial and Bernoulli Naive Bayes, a linear SVM and a random forest) with fine-tuned BERTimbau Base and Large. The classical models cluster between 0.651 and 0.676 macro-F1 and are precision-oriented but recall-poor. The transformers reach 0.833 macro-F1 and accuracy (Wilson 95% CI for accuracy [0.77, 0.88]), a gap of roughly 16 points, while Base and Large are indistinguishable on the test set despite a 2.6x difference in training cost. We argue that the value of the resource lies less in its headline scores than in the auditability of its pipeline, and we state three limits that users must reckon with: the artificial class balance, whose effect on precision under realistic prevalence we quantify; the association between provenance and label, which the released source field makes measurable; and the selection bias induced by lexicon-guided sampling and automatic triage. The corpus, annotation guide, models and experimental scripts are publicly available.
Abstract (translated)
URL
https://arxiv.org/abs/2611.14036
PDF
-
On the information rate of phones, syllables and words
Abstract
Which linguistic units minimise the bitrate required to encode speech content: phones, syllables, words, maybe larger units, or something in between? This question has implications for the development of speech and language systems and theories of human language processing. Using a transcribed corpus of 960 hours of English audiobooks, I compare the information rates of phones, syllables, words, multi-word combinations and intermediate tokenisations. While all units are equivalent under a perfect language model, it makes sense to compare these under limited modelling budgets; I use $n$-gram and transformer language models. The lowest bitrates are achieved using words, with slightly longer units sometimes performing best. The results show that, when modelling capacity is limited, units larger than syllables but smaller than long multi-word phrases provide the most efficient representation in terms of information rate.
Abstract (translated)
URL
https://arxiv.org/abs/2611.13990
PDF
-
Alania-2: Turkish text-to-speech with instant and fine-tuned voice cloning
2026-11-10 10:04:47Sezgin Saygili, Emre Kaplaner, Oncel Ozgul, Fikri San KoktasAbstract
We present Alania-2, a Turkish text-to-speech (TTS) system with two built-in voices, voice design from text descriptions, and two kinds of voice cloning. Alania-2 is a 488M-parameter model with the VoxCPM architecture, which generates continuous audio latents with a diffusion head on top of a language-model backbone. We train it from scratch on 18,195 hours of Turkish speech, 90% of which is synthetic speech produced by a larger teacher model and filtered automatically. A text normalizer rewrites numbers, abbreviations and foreign names into the words a Turkish speaker would say; without it, none of 96 test sentences with digits kept all its critical parts, such as an amount or a date, through synthesis and speech recognition, against 86 with it. On Freya-TR-Eval, a public benchmark of 495 Turkish sentences scored with automatic speech recognition, Alania-2 reaches a word error rate of 1.44%, compared with 5.99% for our previous system. On its built-in voices, it comes within half a point of character error rate of a 2.3B-parameter model at under a quarter of the size, and a 7.6M-parameter single-voice model trained with the same recipe matches the most intelligible systems we measured, which suggests that training data matters more for intelligibility than model size. We also report practical findings on voice cloning: conditioning on the reference audio alone, rather than continuing it, cuts skipped words from 21.9% to 1.7%; 20-30 seconds of reference audio works best; and a low-rank adapter trained in about two minutes raises speaker similarity well above instant cloning. Finally, we describe inference-engine changes that bring first audio down to 32 ms for a single stream. We release our evaluation harness and three Turkish speech and text datasets.
Abstract (translated)
URL
https://arxiv.org/abs/2611.13845
PDF
-
Swanada: Split-Encoder Capability-Routed Conformer CTC and Sub-30 MB On-Device Speech Recognition Across Twelve Scripts
2026-11-10 09:43:36Hemanth HMAbstract
Client-side automatic speech recognition (ASR) across Indic and Classical scripts eliminates cloud transcription costs, network tail latency, and server-side audio retention, yet neural exports face four systems bottlenecks in browsers and mobile runtimes: (1) monolithic per-language Conformer checkpoints require ~177 MiB per script (> 2.1 GiB across 12 scripts), exhausting CacheStorage quotas; (2) static relative positional encoding tables embed 10.2-20.5 MB of FP32 constants while 1x1 Conv1d layers trigger slow or unsupported ConvInteger kernels in WebAssembly SIMD; (3) sliding-window CTC flickers on boundary tokens; and (4) mobile WebAudio pipelines suffer 44.1/48 kHz sample-rate mismatches and silent AudioWorklet stalls under iOS WKWebView. We present Swanada, an open-source 16 kHz streaming ASR system spanning Classical Sanskrit, Indian English, and ten Indian languages (sa, en, hi, kn, mr, bn, ta, te, ml, gu, pa, or) across browser WebAssembly, this http URL, and Python. Swanada pairs capability-routed shared INT8 Conformer encoders (a 17-layer 177.38 MiB Studio tier at $d = 512$ for desktop/servers and a distilled 14-layer 25.35 MiB Mobile tier at $d = 256$ for phones) with twelve hot-swappable 515.0 KiB native-script linear CTC heads ($257 \times 512$), cutting 12-language storage by $11.6\times$ (183.4 MiB) and $67.8\times$ (31.39 MiB; 25.85 MiB first load). Runtime protobuf tail surgery exposes the shared representation in $< 2$ ms, while dynamic positional synthesis, MatMul pointwise convolutions, and post-quantization QInt8 head calibration preserve sub-30 MB accuracy. Across $N = 480$ held-out and $N = 1,280$ open public test clips (FLEURS, LibriSpeech, GLOBE, AMI, wild Sanskrit), Swanada achieves 5.43% held-out and 8.87% 12-language open CER on Studio (18.56% and 19.23% on Mobile) at $0.0097\times$-$0.0264\times$ RTF, with 30.6 ms median TTFT and 0.00% committed-prefix churn.
Abstract (translated)
URL
https://arxiv.org/abs/2611.13814
PDF
-
Antalia-2 Mini: A 7.6-million-parameter Turkish text-to-speech model trained entirely on synthetic speech
2026-11-10 09:41:21Sezgin Saygili, Emre Kaplaner, Oncel Ozgul, Fikri San KoktasAbstract
Antalia-2 Mini is a single-voice Turkish text-to-speech model with 7,620,931 parameters. It reads characters and produces 48 kHz audio at 41 times real time on an Apple M5 laptop CPU (2 threads). Its training audio came only from VoxCPM2 (the teacher), about 300 times larger, cloning one synthetic clip; automatic checks kept about 1,790 h. Fine-tuning the vocoder on the acoustic model's own mel spectrograms cut the word error rate (WER) of Whisper's transcripts from a mean of 3.89% to 1.75% at adjacent pre-release checkpoints. The teacher corrupts the start of a clip when the first word begins with ö, the student learned this, and a spoken prefix we cut from the output hides it. On the 495-sentence Freya-TR-Eval benchmark, where we tuned settings, version 1.0.1 has WER 1.09% and character error rate 0.22% on our public harness. A paired bootstrap cannot distinguish these from EMA Lightning, an 8.6M Turkish model, which scores higher on full-band UTMOS, an English-trained quality proxy. No human listening test exists. Weights, inference code and the benchmark harness are public under Apache-2.0; training code and data are not.
Abstract (translated)
URL
https://arxiv.org/abs/2611.13811
PDF
-
Ninada: Sub-Clause Split-Graph VITS Streaming and Capability-Routed 24 kHz Speech Synthesis Across Twelve Scripts
Abstract
End-to-end VITS text-to-speech models synthesize natural speech in a single forward pass, yet standard ONNX exports execute monolithically over an entire clause, delaying Time-To-First-Audio (TTFA) to several seconds and spiking activation memory. Client-side deployment across Indic scripts faces four further bottlenecks: 138.8 MB float32 checkpoints per language strain browser caches, mobile WebGPU stalls for 6-14 s compiling convolution shaders while still running slower than real time ($1.44\times$ RTF), small WebAssembly pthread pools stall at barriers on big-LITTLE mobile CPUs, and mobile Web Audio pipelines drop frames or mute output under AudioWorklet contention and iOS WKWebView rules. We present Ninada, an open-source 24 kHz TTS system covering English, Classical Sanskrit, and ten Indian languages (sa, en, hi, kn, mr, bn, ta, te, ml, gu, pa, or) across a capability-routed 48-checkpoint fleet: 24 studio HiFi-GAN v1 specialists (27.9 MB W8A32+FP16 ONNX) for desktop/WebGPU and 24 distilled c8c8_384 iSTFTNet specialists (21.7 MB ONNX, $2.64\times$ fewer vocoder MACs via an exact NOLA dual-window ConvTranspose1d inverse STFT with zero complex ops) for mobile WASM, plus a 0.0 MB zero-dependency Klatt fallback. At runtime, Ninada splits a single ONNX file in memory in $< 15$ ms into prior, flow, and windowed vocoder sub-sessions streaming 12-frame halos with raised-cosine crossfades. On $N = 480$ held-out 24 kHz studio utterances, Ninada achieves 3.22 dB 0-8 kHz MCD (3.09 dB mobile iSTFTNet; 1.33 dB 24-voice distillation MCD, vs. 5.17 dB for the 1,526.8 MB 16 kHz Meta MMS fleet) and 6.07 dB 8-12 kHz error, delivering 89-155 ms audible startup on desktop WebGPU and 538-649 ms zero-underrun streaming at 0.75-0.76x RTF on a physical Google Pixel 10 Pro XL.
Abstract (translated)
URL
https://arxiv.org/abs/2611.10035
PDF
-
Walrus Technical Report
2026-11-10 07:37:38Alkis Koudounas, Osamu Take, Hayato Futami, Quentin Jodelet, Emiru TsunooAbstract
We present Walrus, a unified 8B speech-and-audio foundation model handling speech recognition (including code-switching and multi-talker speech), speech-to-text and speech-to-speech translation, text-to-speech, emotion recognition, audio captioning, and general audio understanding. Our main contribution is not a new architecture but a data-centric training recipe; we use a standard decoder-only backbone initialized from Gemma 4, so that gains can be traced to the data. For speech generation, Walrus predicts low-bitrate discrete speech tokens, and a lightweight token-to-waveform decoder converts them into high-fidelity audio. The recipe is a three-stage data curriculum that moves from broad multilingual data to multi-talker, code-switching, and domain-specific data. At each stage, task-aware and dataset-aware weighting balances tasks with very different token budgets and protects core skills from forgetting. Trained on over 600M samples covering more than 20 tasks and 100 languages, Walrus leads on most multilingual, multi-talker, and code-switching ASR, speech translation, and audio understanding benchmarks, outperforming both open speech language models of comparable size and dedicated task-specific models. On most benchmarks, it even matches or surpasses the $4\times$ larger Qwen3-Omni's performance. Results suggest that data engineering is a primary driver of quality in multi-task speech and audio modeling.
Abstract (translated)
URL
https://arxiv.org/abs/2611.13689
PDF
-
An Interpretable Stylometric Framework for Early Chinese Buddhist Translator Styles: A Case Study on the Attribution of Disputed Works of An Shigao
Abstract
Attributing early Chinese Buddhist translations requires evidence that scholars can trace back to the texts. We develop an interpretable stylometric framework that uses features defined or reviewed by Buddhist philologists, including function characters, sentence formulas, sentence-final characters, sentence length, and part-of-speech n-grams. We construct stylistic profiles from 31 authenticated translations by seven translators from the Eastern Han through the Western Jin, then compare eight disputed texts with An Shigao's profile. Two texts, T1557 and T0101, show the strongest stylistic similarity to the authenticated An Shigao corpus, with scores of 0.783 and 0.687, respectively. The attributions agree with those reported by a published BERT-based approach on the same set of disputed texts. The framework requires no task-specific training and allows scholars to inspect which linguistic features support each proposed attribution.
Abstract (translated)
URL
https://arxiv.org/abs/2611.13636
PDF
-
FlashMotion: Ultra-Fast Long-Term 3D Motion Generation
2026-11-10 06:12:58Kaixing Yang, Xulong Tang, Ronghui Li, Xiangyue Zhang, Ziqiao Peng, Mingshuang Luo, Chenxi Zhao, Haoyu Wu, Ju Fan, Hongyan Liu, Puwei Wang, Zijiao Zeng, Jun HeAbstract
3D human motion generation plays a crucial role in game development, extended reality (XR), digital humans, and embodied simulation. Beyond motion generation, practical pipelines include rigging, retargeting, and rendering, making ultra-fast long-term 3D motion generation without compromising quality valuable for real-time interactive applications. Despite promising progress, high-quality long-horizon motion generation under tight inference budgets remains challenging, owing to 2 coupled bottlenecks: (1) generative strategies that require many sampling steps for high-quality 3D motion generation and (2) model architectures with substantial per-step computational and memory costs for long sequences. We present FlashMotion, a co-designed framework for high-quality long-horizon 3D motion generation under a strict inference budget, covering text-to-motion, music-to-dance, and speech-to-gesture tasks. To enhance trajectory-level efficiency, we introduce a 3D motion-space MeanFlow formulation with kinematic and dynamic consistency constraints for few-step generation, together with an Adaptive Soft masking mechanism for temporally coherent motion editing. To improve per-step efficiency, we design a lightweight non-autoregressive BiMamba architecture with condition-guided state selection and AdaIN-based condition fusion mechanism, enhancing cross-modal alignment and generalization to longer motion sequences. Extensive experiments demonstrate that FlashMotion achieves overall state-of-the-art performance in both quality and efficiency across text-to-motion, speech-to-gesture, and music-to-dance generation. Code is available at FlashMotion (this https URL).
Abstract (translated)
URL
https://arxiv.org/abs/2611.13626
PDF
-
Emilia2: A Large-Scale Dataset for Controllable Long-Form Speech Generation
2026-11-10 05:45:36Qinke Ni, Dekun Chen, Junwen Qiu, Huan Liao, Xu Li, Xingwen Han, Wan Lin, Kunyu Feng, Yuxiang Wang, Yusheng Mu, Chaoren Wang, Yingda Shen, Mingjie Chen, Haolong Zheng, Xulin Fan, Mark A. Hasegawa-Johnson, Zhizheng WuAbstract
Generating long-form speech requires models to keep content, speaker identity, turn-taking, and expressive style consistent over time. Yet most open training corpora contain examples lasting only tens of seconds and open long-form benchmarks cover at most two speakers per recording. We introduce Emilia2, a 229K-hour multilingual dataset for controllable long-form speech generation. From in-the-wild data, Emilia2 constructs short utterances, single-speaker long-form speech, and multi-speaker dialogues, preserving transcripts, speaker identities, pauses, and acoustic transitions. Examples longer than five minutes account for 50.0% of single-speaker long-form hours and 66.8% of dialogue hours. To support expressive control, Emilia2 provides natural language instructions for 4.77M examples and nonverbal-vocalization tags for 933K examples. We also introduce Emilia2-testset, 23 hours of bilingual recordings of up to 30 minutes and four speakers, each speaker paired with reference audio. We train autoregressive (AR) and non-autoregressive (NAR) models on Emilia2. Our 0.6B-backbone NAR model achieves the lowest Chinese CER and English WER on TTSD-Eval (5.27%/9.93%) among compared systems, while our AR model is on par with Qwen3-TTS-12Hz-VD on InstructTTSEval (zh/en: 76.2/78.3 vs. 77.1/77.9). These results validate Emilia2 as a comprehensive data foundation for long-form speech generation and minute-scale expressive control. Demo: this https URL.
Abstract (translated)
URL
https://arxiv.org/abs/2611.13614
PDF
-
Nepali-English Code-Mixed Speech Recognition: Data Preparation and Adaptation of a Speech Foundation Model
2026-11-10 03:39:19Jeevan Neupane, Janardan Bhetwal, Kshitiz Paudel, Anup Aryal, Bipin BashyalAbstract
Nepali speakers regularly mix English words into conversation, but multilingual speech recognisers often struggle with language switches and the writing script used for English words. We study how to adapt SraVaani, a 449M-parameter hybrid TDT-CTC model, using a 172.8-hour corpus covering six Nepali speech domains. Its pretrained tokenizer does not support digits and several common symbols. Rather than replace it or discard these recordings, we generate Nepali and English readings for numbers and use zero-shot ASR output to select a plausible form. Full fine-tuning reduces test word error rate (WER) from 48.6% to 17.1%; a 12-epoch run reaches 16.3%, compared with 16.4% for fine-tuned Whisper large-v3-turbo. Freezing the encoder increases WER to 25.5%, while excluding numeric clips raises WER on numeric test speech from 14.3% to 21.3%. We also evaluate on FLEURS using both its original references and mixed-script alternatives. The comparison shows that transcription conventions explain part, but not all, of the out-of-domain performance loss.
Abstract (translated)
URL
https://arxiv.org/abs/2611.13512
PDF
-
Rhythmic-Semantic Co-Speech Gestures for Emotional Dialogue in Robots
2026-11-10 02:31:56Christian Arzate Cruz, Edwin C. Montiel-V\'azquez, Stefanos Gkikas, Miguel Mejia Tobar, Randy GomezAbstract
Co-speech gestures can increase engagement and improve speech understanding. Yet, most data-driven robot systems generate either rhythmic beat-like motion or semantically meaningful iconic gestures, but not both. We present an emotion-aware rhythmic-semantic co-speech gesture system that generates iconic and beat behaviors in real time from text and the robot's internal emotional state. We formalize the task of beat-iconic gesture generation using emotion and propose lightweight models that predict iconic timing and intensity, then fill the remaining spans with emotion-conditioned beats at an appropriate intensity level. We train on the BEAT2 dataset and deploy the system on the social robot Haru. In an online user study with 60 participants, our method shows stronger speech-gesture alignment and is preferred over baseline methods.
Abstract (translated)
URL
https://arxiv.org/abs/2611.13472
PDF
