wq2012 / VoiceIdentityBook

《声纹技术:从核心算法到工程实践》
https://item.jd.com/12970526.html
154 stars 19 forks source link

书中3.5.2数据预处理(p104)中提到的使用语音识别模型来做音频数据的筛选,可有推荐?还有其它语音质量评估工具的建议么? #5

Closed PiBingbin closed 3 years ago

wq2012 commented 3 years ago

任何语音识别模型都可以。你可以去搜索开源ASR模型。甚至条件允许的话你用商业级别的语音识别云服务都可以。

出了ASR之外,也可以用WADA SNR[1]去预估SNR,把SNR太低的去掉。

[1] Chanwoo Kim and Richard M Stern, “Robust signal-to-noise ratio estimation based on waveform amplitude distribution analysis,” in Ninth Annual Conference of the International Speech Communication Association, 2008.

PiBingbin commented 3 years ago

任何语音识别模型都可以。你可以去搜索开源ASR模型。甚至条件允许的话你用商业级别的语音识别云服务都可以。

出了ASR之外,也可以用WADA SNR[1]去预估SNR,把SNR太低的去掉。

[1] Chanwoo Kim and Richard M Stern, “Robust signal-to-noise ratio estimation based on waveform amplitude distribution analysis,” in Ninth Annual Conference of the International Speech Communication Association, 2008.

非常感谢大佬的解答!