研究人员推出了CN-NewsTTS Bench，这是一个旨在评估中文文本转语音（TTS）系统发音准确性的新基准。该基准专门针对新闻内容中常见的复杂书写形式，如分数、连字符姓名以及中英文字母数字混合表达，这些形式可能导致TTS系统发音错误。该基准包含开发集和测试集、自动评估目标以及来自自动语音识别（ASR）集成系统的转录文本，初步结果显示表现最佳的系统准确率达到0.879，而其他系统则显著低于0.60。 AI
影响 该基准旨在提高中文新闻TTS的自然度和准确性，有望带来更好的语音助手和音频内容生成。
排序理由 该集群描述了一个用于评估TTS系统的新基准，属于研究范畴。
在 Hugging Face Daily Papers 阅读 →
AI 生成摘要 · Google Gemini · 来自 3 个来源。 我们如何撰写摘要 →
arXiv:2606.24714v1 Announce Type: new Abstract: Chinese news text contains dense written forms such as scores, hyphenated model names, ranges, unit symbols, percentages, English abbreviations, and mixed Chinese-Latin-digit names. These forms are frequent in real listening workflo…
Chinese news text contains dense written forms such as scores, hyphenated model names, ranges, unit symbols, percentages, English abbreviations, and mixed Chinese-Latin-digit names. These forms are frequent in real listening workflows, and a text-to-speech (TTS) system can preser…
Chinese news text contains dense written forms such as scores, hyphenated model names, ranges, unit symbols, percentages, English abbreviations, and mixed Chinese-Latin-digit names. These forms are frequent in real listening workflows, and a text-to-speech (TTS) system can preser…