Long-form live-streaming TTS requires context-dependent prosody and paragraph-level coherence. However, many existing instruction-based TTS systems use global or uniform conditions, providing limited explicit control over clause-level relative prosodic changes. We introduce Speaker-Relative Inline Prosody Control, where each Pitch, Energy, or Speed instruction targets a clause relative to the preceding clause from the same speaker, while Pause uses an absolute duration interval. In codec-based TTS, Speed and Pause affect sequence length, whereas Pitch and Energy rely on residual codebooks. By analyzing Qwen3-TTS RVQ codebooks, we find that Energy concentrates in early residual codebooks, whereas Pitch accumulates across a deeper prefix. We therefore propose Scope- and Codebook-Aware Instruction Conditioning (SCIC), combining a Temporal Instruction Router for frame-level tag activation with Tag-Specific Codebook Weighting over residual codebooks. SCIC improves speaker-relative Pitch and Energy control over standard instruction fine-tuning using text-token tags. We further apply multi-reward GDPO post-training to jointly optimize control and quality, improving control accuracy while preserving CER and speaker similarity. In long-form synthesis, SCIC produces a more distinct paragraph-level expressive hierarchy than speaker-adapted SFT without instructions. Audio demos are available at: https://taoliveaigc.github.io/SCIC/
Figures & tables
Figure 1: SCIC in Qwen3-TTS. The Main Talker predicts q0 and provides ht . For each Pitch/Energy tag, its tag embedding is scaled by the Temporal Instruction Router output and the tag-specific codebook weight to form st,k , which conditions MTP prediction of q1 – q15 .
Figure 2: Median speaker-relative change magnitudes after direction normalization. Down values are sign-flipped so positive values indicate the requested direction; black lines and diamonds denote frozen training-reference magnitudes.
Table 1: RVQ transfer (%). Left: selected single-codebook results with speaker-bootstrap 95% confidence intervals. Right: cumulative transfer over residual-codebook prefixes.
Variant
Pitch
Energy
MAE ↓
SCIC-SFT
90.14
96.39
295 ms
w/ Hard Router
88.41
96.44
375 ms
w/o Tag-Specific Codebook Weighting
89.43
94.90
292 ms
Table 3: Ablations: control-direction accuracy (%) and Router Boundary MAE.
Instruction-based text-to-speech (ITTS) systems enable natural-language control of expressive speech generation, but often offer limited transparency and fine-grained control over individual text units. Character-level controllable TTS systems provide explicit acoustic control, yet typically rely on user-specified acoustic attributes. To bridge this gap, we propose InstCharVoice, a unified framework that grounds natural-language instructions in character-level acoustic control. We first construct grounded instruction annotations on the WordVoice-5A-zh corpus using Qwen3-Omni. With this supervision, we train an autoregressive model to identify instruction-relevant characters and predict their acoustic attributes before generating the corresponding speech tokens. Keyword prediction and grounding-aware loss weighting help the model focus on instruction-relevant characters and attributes. Experiments show improved instruction following and keyword-level acoustic control over representative ITTS systems, with competitive speech naturalness and explicit character-level controllability. Audio samples are available at https://xxh333.github.io/instcharvoice-demo/.
Sihang Nie, Xueru Li, Xiaofen Xing +4
South China University of Technology · Huya Inc. · The Hongkong Polytechnic University
While existing text-to-speech (TTS) models exhibit high expressiveness, fine-grained control over composite instructions remains challenging due to the structural mismatch between discrete textual intents and continuous acoustic realizations. Inspired by human cognitive decoupling, we introduce AgentSteerTTS, a multi-agent closed-loop framework designed for intent-faithful expressive control of composite instructions. First, in our framework, an adversarial disentanglement agent mitigates speaker-emotion leakage by learning separable identity and emotion-prosody subspaces with leakage-suppressing regularization. Next, a Dual-Stream Anchoring Controller grounds abstract intents using a large-scale acoustic prototype library: a Retrieval Agent selects expressive anchors, while a Synthesis Agent fuses them into continuous control vectors via gated attention. Finally, a Fast-Slow Feedback Agent refines output intensity through latent gradient correction and resolves semantic-acoustic mismatches using high-level perceptual critique. Experiments on a composite-instruction benchmark and public test sets show that AgentSteerTTS yields consistent and significant improvements to the baselines, demonstrating the effectiveness of the proposed method.
Bin Kang, Shaoguo Wen, Yang Fan +6
University of Chinese Academy of Sciences · Shenzhen Loop Area Institute · Tencent Turinglab +2
Long-form text-to-speech (TTS) enables multi-turn conversations with consistent prosody and higher quality voice cloning from longer reference audio. Recent open-weights autoregressive TTS models such as Qwen3-TTS and VoxCPM2 attain state-of-the-art word error rate (WER) and speaker similarity (SIM) on short-form prompts but significantly deteriorate when used with long-form prompts. We propose Localized Attention-Constrained Inference (LACI), an inference-only method to detect TTS errors in near real-time, roll back to the error onset and regenerate with temporary guardrails, adding negligible computational overhead. Using LACI, we improve worst-of-N WER across 10 RNG seeds for Qwen3-TTS-0.6B from 35.2% to 3.4% on prompts longer than 1500 words, even surpassing its short-form reliability of 5.4% on prompts with fewer than 500 words. To demonstrate the efficacy of LACI on voice cloning reliability, we propose a sliding-window version of the SIM metric that we call wSIM. wSIM exposes several novel failure patterns that are not captured by SIM. LACI improves worst-of-N wSIM from 0.01 to 0.47 on 120 seconds of reference audio while reducing the rate of catastrophic generations with WER above 30% from 26% to below 1%
Rongxiang Wang, Berkin Durmus, Aysegul Orhon +2
Argmax, Inc. · University of Virginia · Bilkent University