遇见数据集

SynDataLab/omnivoice-pesq-stoi-filtered-tr

收藏
Hugging Face2026-04-10 更新2026-04-12 收录
官方服务:

资源简介:

--- language: - tr license: cc-by-4.0 task_categories: - text-to-speech tags: - omnivoice - synthetic - tts - filtered - pesq - stoi --- # OmniVoice PESQ/STOI Filtered - Turkish Synthetic Turkish speech dataset generated with [OmniVoice](https://github.com/k2-fsa/OmniVoice), filtered by **PESQ and STOI**. - **Filter:** PESQ >= 1.5 AND STOI >= 0.5 (estimated via SQUIM, no reference needed) - **Sample rate:** 24 kHz - **Original:** 200 files - **After filtering:** 163 files (82% kept)

--- 语言: - 土耳其语(tr) 许可证:CC BY 4.0 任务类别: - 文本转语音(text-to-speech) 标签: - OmniVoice - 合成 - TTS - 已过滤 - PESQ - STOI --- # OmniVoice PESQ/STOI 过滤版——土耳其语数据集 本数据集为基于OmniVoice生成的土耳其语合成语音数据集,经**PESQ(Perceptual Evaluation of Speech Quality,语音感知质量评估)**与**STOI(Short-Time Objective Intelligibility,短时客观可懂度)**指标过滤得到。 - **过滤规则**:PESQ≥1.5 且 STOI≥0.5(通过SQUIM模型估算,无需参考音频) - **采样率**:24 kHz - **原始样本量**:200条音频文件 - **过滤后样本量**:163条音频文件(保留率82%)

提供机构:
SynDataLab
二维码
社区交流群
二维码
科研交流群
商业服务