mlsanityData to 14 Oct 2026

MagpieTTS-LF: Inference-Time Long-Form Speech Generation Without Training on Long-Form data

arXiv 2606.18485 ↗First posted 16 Jun 2026cs.SD cs.AI eess.AScited by 1

Not on the map. Why?

Abstract (arXiv)

Neural Text-to-Speech (TTS) systems achieve remarkable quality on short utterances but long-form speech generation shows prosodic drift, speaker inconsistencies and sentence boundary artifacts. Existing approaches either compress sequences, increase context length or naively concatenate independently synthesized chunks. We present an inference-time approach called MagpieTTS-LF that enables MagpieTTS to produce coherent long-form speech without model retraining. Our method introduces three key innovations: (1) soft attention priors to guide monotonic alignment while preserving past and future context; (2) a stateful inference algorithm that maintains context across sentence chunks, ensuring prosodic continuity; (3) history-aware text encoding that uses past text for discourse-level prosodic planning. Experiments on long texts show significant improvements in long-range intelligibility, prosodic coherence, speaker consistency, and boundary naturalness compared to other baselines.

Key facts. Cited by 1 paper in the map (data through 14 Oct 2026). Builds on 22 of its 22 references that are in the map. First posted on arXiv 16 Jun 2026.

Builds on

22 of 22 references.

References · 22

  1. 2022Robust Speech Recognition via Large-Scale Weak Supervisioncited by 4,057
  2. 2021WavLM: Large-Scale Self-Supervised Pre-Training for Full Stack Speech Processingcited by 1,657
  3. 2020MLS: A Large-Scale Multilingual Dataset for Speech Researchcited by 476
  4. 2022UTMOS: UTokyo-SaruLab System for VoiceMOS Challenge 2022cited by 390
  5. 2023Voicebox: Text-Guided Multilingual Universal Speech Generation at Scalecited by 379
  6. 2024CosyVoice: A Scalable Multilingual Zero-shot Text-to-speech Synthesizer based on Supervised Semantic Tokenscited by 334
  7. 2024CosyVoice 2: Scalable Streaming Speech Synthesis with Large Language Modelscited by 288
  8. 2023NaturalSpeech 2: Latent Diffusion Models are Natural and Zero-Shot Speech and Singing Synthesizerscited by 235
  9. 2024NaturalSpeech 3: Zero-Shot Speech Synthesis with Factorized Codec and Diffusion Modelscited by 226
  10. 2024XTTS: a Massively Multilingual Zero-Shot Text-to-Speech Modelcited by 179
  11. 2024VALL-E 2: Neural Codec Language Models are Human Parity Zero-Shot Text to Speech Synthesizerscited by 122
  12. 2026Qwen3-TTS Technical Reportcited by 118
10 more
  1. 2023Better speech synthesis through scalingcited by 63
  2. 2021TitaNet: Neural Model for speaker representation with 1D Depth-wise separable convolutions and global contextcited by 62
  3. 2024VALL-E R: Robust and Efficient Zero-Shot Text-to-Speech Synthesis via Monotonic Alignmentcited by 38
  4. 2025VibeVoice Technical Reportcited by 29
  5. 2025Koel-TTS: Enhancing LLM based Speech Generation with Preference Alignment and Classifier Free Guidancecited by 27
  6. 2024Improving Robustness of LLM-based Speech Synthesis by Learning Monotonic Alignmentcited by 22
  7. 2025NanoCodec: Towards High-Quality Ultra Fast Speech LLM Inferencecited by 15
  8. 2022Structured State Space Decoder for Speech Recognition and Synthesiscited by 11
  9. 2023HiGNN-TTS: Hierarchical Prosody Modeling with Graph Neural Networks for Expressive Long-form TTScited by 1
  10. 2025Long-Context Speech Synthesis with Context-Aware Memorycited by 1

Built on by

Cited by year (2026 to 14 Oct)

Citing papers per year: 2026: 1.

1 paper.

  1. 2026X2Streaming-TTS: Causal Token-Level Text-to-Speech from Streaming Text with Speech-State Inheritancecited by 0