T5Gemma-TTS Technical Report

Autoregressive neural codec language models have shown strong zero-shot voice cloning ability, but decoder-only architectures treat input text as a prefix that competes with the growing audio sequence for positional capacity, weakening text conditioning over long utterances. We present T5Gemma-TTS, an encoder-decoder codec language model that maintains persistent text conditioning by routing bidirectional text representations through cross-attention at every decoder layer. Built on the T5Gemma pretrained encoder-decoder backbone (2B encoder + 2B decoder; 4B parameters), it inherits rich linguistic knowledge without phoneme conversion and processes text directly at the subword level. To improve duration control, we introduce Progress-Monitoring Rotary Position Embedding (PM-RoPE) in all 26 cross-attention layers, injecting normalized progress signals that help the decoder track target speech length. Trained on 170,000 hours of multilingual speech in English, Chinese, and Japanese, T5Gemma-TTS achieves a statistically significant speaker-similarity gain on Japanese over XTTSv2 (0.677 vs. 0.622; non-overlapping 95% confidence intervals) and the highest numerical Korean speaker similarity (0.747) despite Korean not being included in training, although this margin over XTTSv2 (0.741) is not statistically conclusive. It also attains the lowest numerical Japanese character error rate among five baselines (0.126), though this ranking should be interpreted cautiously because of partial confidence-interval overlap with Kokoro. English results on LibriSpeech should be viewed as an upper-bound estimate because LibriHeavy is a superset of LibriSpeech. Using the same checkpoint, disabling PM-RoPE at inference causes near-complete synthesis failure: CER degrades from 0.129 to 0.982 and duration accuracy drops from 79% to 46%. Code and weights are available at https://github.com/Aratako/T5Gemma-TTS.

TRACE: Training-Free Partial Audio Deepfake Detection via Embedding Trajectory Analysis of Speech Foundation Models

Partial audio deepfakes, where synthesized segments are spliced into genuine recordings, are particularly deceptive because most of the audio remains authentic. Existing detectors are supervised: they require frame-level annotations, overfit to specific synthesis pipelines, and must be retrained as new generative models emerge. We argue that this supervision is unnecessary. We hypothesize that speech foundation models implicitly encode a forensic signal: genuine speech forms smooth, slowly varying embedding trajectories, while splice boundaries introduce abrupt disruptions in frame-level transitions. Building on this, we propose TRACE (Training-free Representation-based Audio Countermeasure via Embedding dynamics), a training-free framework that detects partial audio deepfakes by analyzing the first-order dynamics of frozen speech foundation model representations without any training, labeled data, or architectural modification. We evaluate TRACE on four benchmarks that span two languages using six speech foundation models. In PartialSpoof, TRACE achieves 8.08% EER, competitive with fine-tuned supervised baselines. In LlamaPartialSpoof, the most challenging benchmark featuring LLM-driven commercial synthesis, TRACE surpasses a supervised baseline outright (24.12% vs. 24.49% EER) without any target-domain data. These results show that temporal dynamics in speech foundation models provide an effective, generalize signal for training-free audio forensics.

Cinematic Audio Source Separation Using Visual Cues

Cinematic Audio Source Separation (CASS) aims to decompose mixed film audio into speech, music, and sound effects, enabling applications like dubbing and remastering. Existing CASS approaches are audio-only, overlooking the inherent audio-visual nature of films, where sounds often align with visual cues. We present the first framework for audio-visual CASS (AV-CASS), leveraging visual context to enhance separation quality. Our method formulates CASS as a conditional generative modeling problem using conditional flow matching, enabling multimodal audio source separation. To address the lack of cinematic datasets with isolated sound tracks, we introduce a training data synthesis pipeline that pairs in-the-wild audio and video streams (e.g., facial videos for speech, scene videos for effects) and design a dedicated visual encoder for this dual-stream setup. Trained entirely on synthetic data, our model generalizes effectively to real-world cinematic content and achieves strong performance on synthetic, real-world, and audio-only CASS benchmarks. Code and demo are available at \url{https://cass-flowmatching.github.io}.

How Open is Open TTS? A Practical Evaluation of Open Source TTS Tools for Romanian

Open-source text-to-speech (TTS) frameworks have emerged as highly adaptable platforms for developing speech synthesis systems across a wide range of languages. However, their applicability is not uniform -- particularly when the target language is under-resourced or when computational resources are constrained. In this study, we systematically assess the feasibility of building novel TTS models using four widely adopted open-source architectures: FastPitch, VITS, Grad-TTS, and Matcha-TTS. Our evaluation spans multiple dimensions, including qualitative aspects such as ease of installation, dataset preparation, and hardware requirements, as well as quantitative assessments of synthesis quality for Romanian. We employ both objective metrics and subjective listening tests to evaluate intelligibility, speaker similarity, and naturalness of the generated speech. The results reveal significant challenges in tool chain setup, data preprocessing, and computational efficiency, which can hinder adoption in low-resource contexts. By grounding the analysis in reproducible protocols and accessible evaluation criteria, this work aims to inform best practices and promote more inclusive, language-diverse TTS development. All information needed to reproduce this study (i.e. code and data) are available in our git repository: https://gitlab.com/opentts_ragman/OpenTTS

InterDyad: Interactive Dyadic Speech-to-Video Generation by Querying Intermediate Visual Guidance

Despite progress in speech-to-video synthesis, existing methods often struggle to capture cross-individual dependencies and provide fine-grained control over reactive behaviors in dyadic settings. To address these challenges, we propose InterDyad, a framework that enables naturalistic interactive dynamics synthesis via querying structural motion guidance. Specifically, we first design an Interactivity Injector that achieves video reenactment based on identity-agnostic motion priors extracted from reference videos. Building upon this, we introduce a MetaQuery-based modality alignment mechanism to bridge the gap between conversational audio and these motion priors. By leveraging a Multimodal Large Language Model (MLLM), our framework is able to distill linguistic intent from audio to dictate the precise timing and appropriateness of reactions. To further improve lip-sync quality under extreme head poses, we propose Role-aware Dyadic Gaussian Guidance (RoDG) for enhanced lip-synchronization and spatial consistency. Finally, we introduce a dedicated evaluation suite with novelly designed metrics to quantify dyadic interaction. Comprehensive experiments demonstrate that InterDyad significantly outperforms state-of-the-art methods in producing natural and contextually grounded two-person interactions. Please refer to our project page for demo videos: https://interdyad.github.io/.

Once-for-All Channel Mixers (HYPERTINYPW): Generative Compression for TinyML

Deploying neural networks on microcontrollers is constrained by kilobytes of flash and SRAM, where 1x1 pointwise (PW) mixers often dominate memory even after INT8 quantization across vision, audio, and wearable sensing. We present HYPER-TINYPW, a compression-as-generation approach that replaces most stored PW weights with generated weights: a shared micro-MLP synthesizes PW kernels once at load time from tiny per-layer codes, caches them, and executes them with standard integer operators. This preserves commodity MCU runtimes and adds only a one-off synthesis cost; steady-state latency and energy match INT8 separable CNN baselines. Enforcing a shared latent basis across layers removes cross-layer redundancy, while keeping PW1 in INT8 stabilizes early, morphology-sensitive mixing. We contribute (i) TinyML-faithful packed-byte accounting covering generator, heads/factorization, codes, kept PW1, and backbone; (ii) a unified evaluation with validation-tuned t* and bootstrap confidence intervals; and (iii) a deployability analysis covering integer-only inference and boot versus lazy synthesis. On three ECG benchmarks (Apnea-ECG, PTB-XL, MIT-BIH), HYPER-TINYPW shifts the macro-F1 versus flash Pareto frontier: at about 225 kB it matches a roughly 1.4 MB CNN while being 6.31x smaller (84.15% fewer bytes), retaining at least 95% of large-model macro-F1. Under 32-64 kB budgets it sustains balanced detection where compact baselines degrade. The mechanism applies broadly to other 1D biosignals, on-device speech, and embedded sensing tasks where per-layer redundancy dominates, indicating a wider role for compression-as-generation in resource-constrained ML systems. Beyond ECG, HYPER-TINYPW transfers to TinyML audio: on Speech Commands it reaches 96.2% test accuracy (98.2% best validation), supporting broader applicability to embedded sensing workloads where repeated linear mixers dominate memory.

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.

Audio Avatar Fingerprinting: An Approach for Authorized Use of Voice Cloning in the Era of Synthetic Audio

With the advancements in AI speech synthesis, it is easier than ever before to generate realistic audio in a target voice. One only needs a few seconds of reference audio from the target, quite literally putting words in the target person's mouth. This imposes a new set of forensics-related challenges on speech-based authentication systems, videoconferencing, and audio-visual broadcasting platforms, where we want to detect synthetic speech. At the same time, leveraging AI speech synthesis can enhance the different modes of communication through features such as low-bandwidth communication and audio enhancements - leading to ever-increasing legitimate use-cases of synthetic audio. In this case, we want to verify if the synthesized voice is actually spoken by the user. This will require a mechanism to verify whether a given synthetic audio is driven by an authorized identity, or not. We term this task audio avatar fingerprinting. As a step towards audio forensics in these new and emerging situations, we analyze and extend an off-the-shelf speaker verification model developed outside of forensics context for the task of fake speech detection and audio avatar fingerprinting, the first experimentation of its kind. Furthermore, we observe that no existing dataset allows for the novel task of verifying the authorized use of synthetic audio - a limitation which we address by introducing a new speech forensics dataset for this novel task.

Gesture2Speech: How Far Can Hand Movements Shape Expressive Speech?

Human communication seamlessly integrates speech and bodily motion, where hand gestures naturally complement vocal prosody to express intent, emotion, and emphasis. While recent text-to-speech (TTS) systems have begun incorporating multimodal cues such as facial expressions or lip movements, the role of hand gestures in shaping prosody remains largely underexplored. We propose a novel multimodal TTS framework, Gesture2Speech, that leverages visual gesture cues to modulate prosody in synthesized speech. Motivated by the observation that confident and expressive speakers coordinate gestures with vocal prosody, we introduce a multimodal Mixture-of-Experts (MoE) architecture that dynamically fuses linguistic content and gesture features within a dedicated style extraction module. The fused representation conditions an LLM-based speech decoder, enabling prosodic modulation that is temporally aligned with hand movements. We further design a gesture-speech alignment loss that explicitly models their temporal correspondence to ensure fine-grained synchrony between gestures and prosodic contours. Evaluations on the PATS dataset show that Gesture2Speech outperforms state-of-the-art baselines in both speech naturalness and gesture-speech synchrony. To the best of our knowledge, this is the first work to utilize hand gesture cues for prosody control in neural speech synthesis. Demo samples are available at https://research.sri-media-analysis.com/aaai26-beeu-gesture2speech/

On the Emotion Understanding of Synthesized Speech

Emotion is a core paralinguistic feature in voice interaction. It is widely believed that emotion understanding models learn fundamental representations that transfer to synthesized speech, making emotion understanding results a plausible reward or evaluation metric for assessing emotional expressiveness in speech synthesis. In this work, we critically examine this assumption by systematically evaluating Speech Emotion Recognition (SER) on synthesized speech across datasets, discriminative and generative SER models, and diverse synthesis models. We find that current SER models can not generalize to synthesized speech, largely because speech token prediction during synthesis induces a representation mismatch between synthesized and human speech. Moreover, generative Speech Language Models (SLMs) tend to infer emotion from textual semantics while ignoring paralinguistic cues. Overall, our findings suggest that existing SER models often exploit non-robust shortcuts rather than capturing fundamental features, and paralinguistic understanding in SLMs remains challenging.