# Pronouncing Hard Words: Phoneme Adaptation of a Pretrained Conversational Text to Speech Model

arXiv: https://arxiv.org/abs/2610.24164 · HTML: https://mlsanity.com/paper/2610.24164/

## Key facts

- No citing papers in the map yet (data through 19 Oct 2026).
- Placed in Audio language models and spoken dialogue, one of 11 directions in Speech and audio language models.
- Builds on 25 of its 31 references that are in the map.
- First posted on arXiv 16 Oct 2026.

## Placed in

- Speech and audio language models / [Audio language models and spoken dialogue](https://mlsanity.com/speech/audio-language-models-and-spoken-dialogue/)

## Builds on (most cited first, 25 of 25)

- [DeepSeekMath: Pushing the Limits of Mathematical Reasoning in Open Language Models](https://mlsanity.com/paper/2402.03300/) (2024)
- [Natural TTS Synthesis by Conditioning WaveNet on Mel Spectrogram Predictions](https://mlsanity.com/paper/1712.05884/) (2017)
- [FastSpeech 2: Fast and High-Quality End-to-End Text to Speech](https://mlsanity.com/paper/2006.04558/) (2020)
- [Conditional Variational Autoencoder with Adversarial Learning for End-to-End Text-to-Speech](https://mlsanity.com/paper/2106.06103/) (2021)
- [Seed-TTS: A Family of High-Quality Versatile Speech Generation Models](https://mlsanity.com/paper/2406.02430/) (2024)
- [YourTTS: Towards Zero-Shot Multi-Speaker TTS and Zero-Shot Voice Conversion for everyone](https://mlsanity.com/paper/2112.02418/) (2021)
- [CosyVoice 3: Towards In-the-wild Speech Generation via Scaling-up and Post-training](https://mlsanity.com/paper/2505.17589/) (2025)
- [FastPitch: Parallel Text-to-speech with Pitch Prediction](https://mlsanity.com/paper/2006.06873/) (2020)
- [Qwen3-TTS Technical Report](https://mlsanity.com/paper/2601.15621/) (2026)
- [The T05 System for The VoiceMOS Challenge 2024: Transfer Learning from Deep Image Classifier to Naturalness MOS Prediction of High-Quality Synthetic Speech](https://mlsanity.com/paper/2409.09305/) (2024)
- [EXPRESSO: A Benchmark and Analysis of Discrete Expressive Speech Resynthesis](https://mlsanity.com/paper/2308.05725/) (2023)
- [Simple and Effective Zero-shot Cross-lingual Phoneme Recognition](https://mlsanity.com/paper/2109.11680/) (2021)
- [PnG BERT: Augmented BERT on Phonemes and Graphemes for Neural TTS](https://mlsanity.com/paper/2103.15060/) (2021)
- [VoxCPM: Tokenizer-Free TTS for Context-Aware Speech Generation and True-to-Life Voice Cloning](https://mlsanity.com/paper/2509.24650/) (2025)
- [FireRedTTS-2: Towards Long Conversational Speech Generation for Podcast and Chatbot](https://mlsanity.com/paper/2509.02020/) (2025)
- [XPhoneBERT: A Pre-trained Multilingual Model for Phoneme Representations for Text-to-Speech](https://mlsanity.com/paper/2305.19709/) (2023)
- [MOSS-TTS Technical Report](https://mlsanity.com/paper/2603.18090/) (2026)
- [Mixed-Phoneme BERT: Improving BERT with Mixed Phoneme and Sup-Phoneme Representations for Text to Speech](https://mlsanity.com/paper/2203.17190/) (2022)
- [CoVoMix2: Advancing Zero-Shot Dialogue Generation with Fully Non-Autoregressive Flow Matching](https://mlsanity.com/paper/2506.00885/) (2025)
- [HiFiTTS-2: A Large-Scale High Bandwidth Speech Dataset](https://mlsanity.com/paper/2506.04152/) (2025)
- [Representation Mixing for TTS Synthesis](https://mlsanity.com/paper/1811.07240/) (2018)
- [Neural Lexicon Reader: Reduce Pronunciation Errors in End-to-end TTS by Leveraging External Textual Knowledge](https://mlsanity.com/paper/2110.09698/) (2021)
- [MOSS-TTSD: Text to Spoken Dialogue Generation](https://mlsanity.com/paper/2603.19739/) (2026)
- [Multilingual context-based pronunciation learning for Text-to-Speech](https://mlsanity.com/paper/2307.16709/) (2023)
- [UltraVoice: Scaling Fine-Grained Style-Controlled Speech Conversations for Spoken Dialogue Models](https://mlsanity.com/paper/2510.22588/) (2025)

## Abstract (arXiv)

Voice agents built as cascaded automatic speech recognition (ASR)->large language model (LLM)-> text-to-speech (TTS) pipelines increasingly rely on conversational TTS: streaming, context conditioned models that adapt delivery to dialogue history. Such models are natural and expressive but offer little explicit control over how rare, trust-critical drug names are pronounced, because these entities are too infrequent, and their vocabulary too fast-growing, to be covered by any amount of pre-, mid-, or post-training. We retrofit phoneme controllability onto a pretrained conversational TTS model without any pre/mid-training, using an ARPAbet vocabulary exten-sion, continued phoneme-augmented supervised fine-tuning (SFT), and a training-free inference-time acoustic-biasing step. On a 7,500 clip clinical drug-pronunciation evaluation, our model raises strict Gemini-judged correctness from 54.6% (no phoneme augmentation) to 90.7%, outperforming every conversational and stateless TTS baseline on the automatic metric. Because no training stage sees a drug name, the entire evaluation is out-of-domain, so the gains reflect learned pronunciation rather than memorization. In a blinded, randomized human study, adult listeners judge 84% of our clips correct, versus 82% for the strongest baseline, with overlapping confidence intervals on a stratified subset. Across all 250 matched clips, Gemini agrees with the human majority on 84.8% of binary labels (k{appa}=0.57).