IMaSC - ICFOSS Malayalam Speech Corpus

IMaSC是由国际自由及开源软件中心资助创建的马拉雅拉姆语文本与语音语料库，包含约50小时的录音，由8位发音人录制，总计34,473对文本-音频数据。该数据集通过从马拉雅拉姆语维基百科中提取文本，并由语言学家协助精心制作，旨在为马拉雅拉姆语的文本到语音合成系统提供丰富的训练资源。IMaSC的应用领域主要集中在提升马拉雅拉姆语的语音合成质量，解决该语言在语音技术领域的资源匮乏问题。

IMaSC is a Malayalam text and speech corpus funded and developed by the International Free and Open Source Software Center. It contains approximately 50 hours of recorded audio, recorded by 8 speakers, with a total of 34,473 text-audio pairs. This corpus was constructed by extracting text from Malayalam Wikipedia and carefully curated with the assistance of linguists, aiming to provide a rich training resource for Malayalam text-to-speech synthesis systems. The main application areas of IMaSC focus on improving the quality of Malayalam speech synthesis and addressing the problem of scarce resources for Malayalam in the field of speech technology.