研究人员开发了 Praxy Voice，一种使用预训练的非印度语模型来改进印度语文本到语音 (TTS) 的方法。该方法结合了用于脚本罗马化的 Brahmic Unified Phoneme Space (BUPS)、用于文本令牌预测器的 LoRA 适配器以及语音提示恢复技术。该方法在无需新的声码器训练或商业 TTS 数据的情况下，实现了泰卢固语、泰米尔语和印地语的商业级音频输出。 AI
影响 能够通过最小的干预和无商业数据，利用现有模型创建高质量的印度语 TTS。
排序理由 详细介绍 TTS 合成新方法的学术论文。
AI 生成摘要 · Google Gemini · 来自 3 个来源。 我们如何撰写摘要 →
arXiv:2604.25441v1 Announce Type: cross Abstract: Commercial TTS systems produce near-native Indic audio, but the best open-source bases (Chatterbox, Indic Parler-TTS, IndicF5) trail them on measured phonological dimensions, and the most widely adopted multilingual base (Chatterb…
Commercial TTS systems produce near-native Indic audio, but the best open-source bases (Chatterbox, Indic Parler-TTS, IndicF5) trail them on measured phonological dimensions, and the most widely adopted multilingual base (Chatterbox, 23 languages) does not even tokenise Telugu or…
Commercial TTS systems produce near-native Indic audio, but the best open-source bases (Chatterbox, Indic Parler-TTS, IndicF5) trail them on measured phonological dimensions, and the most widely adopted multilingual base (Chatterbox, 23 languages) does not even tokenise Telugu or…