OpenVoiceOS/tts-test-sentence
收藏资源简介:
tts-test-sentence数据集是一个用于多语言文本到语音(TTS)测试的资源,包含一个标准测试句子(称为彩虹句子)的跨语言翻译。该句子在英文中为:After the rain, a colorful rainbow appeared in the sky, and the children went outside to look at it.,已被翻译成多种语言和脚本(支持超过100种语言,如af、am、ar、az等,包括中文)。数据集每行提供语言代码(BCP-47)、语言名称、脚本(ISO-15924)、本地翻译文本和英文源句子。它旨在帮助开发者和研究人员快速检查多语言TTS语音是否能在各自语言中产生可理解的语音输出。数据集覆盖多种语言变体(例如塞尔维亚语有西里尔和拉丁脚本),但翻译可能基于AI生成或最佳努力,欢迎更正和扩展。目前尚未覆盖约1000种长尾MMS语言。
The tts-test-sentence dataset is a resource for multilingual text-to-speech (TTS) testing, containing translations of a canonical test sentence (known as the rainbow sentence) across languages and scripts. The English source sentence is: After the rain, a colorful rainbow appeared in the sky, and the children went outside to look at it. It has been translated into numerous languages (over 100 languages including af, am, ar, az, and zh). Each row in the dataset includes lang (BCP-47), language (name), script (ISO-15924), text (native translation), and source_en. It is designed to quickly sanity-check multilingual TTS voices to ensure they produce intelligible speech in their respective languages. The dataset covers multiple language variants (e.g., Serbian in Cyrillic and Latin scripts), but translations are best-effort and may be AI-authored; corrections and additional languages/scripts are welcomed. It does not yet cover approximately 1000 long-tail MMS languages.




