Broadcast-quality synthetic narration: a workflow for fine-grained text-to-speech intonation and emotion control

Kruszielski LF, Leite PHL, Fernandes MP, Pereira A, Biscainho LWP. Broadcast-quality synthetic narration: a workflow for fine-grained text-to-speech intonation and emotion control. In: 2025 AES International Conference on Artificial Intelligence and Machine Learning for Audio. Audio Engineering Society; 2025. Paper 31. Available from: https://aes.org/publications/elibrary-page/?id=23020

@inproceedings{Kruszielski2025_23020,
  author = {Kruszielski, Luiz Fernando and Leite, Pedro H.L. and Fernandes, Myllene P. and Pereira, Andre and Biscainho, Luiz W. P.},
  title = {{Broadcast-quality synthetic narration: a workflow for fine-grained text-to-speech intonation and emotion control}},
  booktitle = {2025 AES International Conference on Artificial Intelligence and Machine Learning for Audio},
  note = {Paper 31},
  year = {2025},
  month = sep,
  publisher = {Audio Engineering Society},
  url = {https://aes.org/publications/elibrary-page/?id=23020}
}

TY  - CPAPER
TI  - Broadcast-quality synthetic narration: a workflow for fine-grained text-to-speech intonation and emotion control
AU  - Kruszielski, Luiz Fernando
AU  - Leite, Pedro H.L.
AU  - Fernandes, Myllene P.
AU  - Pereira, Andre
AU  - Biscainho, Luiz W. P.
T2  - 2025 AES International Conference on Artificial Intelligence and Machine Learning for Audio
M1  - Paper 31
PY  - 2025
DA  - 2025/09/02
UR  - https://aes.org/publications/elibrary-page/?id=23020
PB  - Audio Engineering Society
LA  - en
AB  - The ability to generate high-quality synthetic speech with fine-grained control over intonation and emotion is crucial for broadcast applications such as narration and voice-over production. This paper presents a novel text-to-speech (TTS) workflow designed to generate expressive and user-controllable speech synthesis. By leveraging a one-hot encoded emotional embedding approach through application-aware, manually labeled emotion classes and a fine-tuned neural model, our method allows for precise control over speech tone and intention. We detail the data preparation pipeline, including manual annotation and automatic emotion classification, and describe the implementation of our speech synthesis model using VITS2. Additionally, we introduce an intuitive prosody control mechanism that enables word- and phoneme-level adjustments for pitch and duration. Subjective evaluation results indicate that our synthetic narration achieves parity with professionally recorded voice-overs, with some preference for synthetic speech due to its enhanced customization. This research contributes to the advancement of AI-driven voice production, enabling scalable, high-fidelity, and fine-grained controlled speech synthesis for media and entertainment industries.
ER  -