A Crowdsourced Open-Source Kazakh Speech Corpus and Initial Speech Recognition Baseline
This addresses the problem of low-resource language processing for Kazakh speakers and researchers, though it is incremental as it builds on existing methods for corpus creation.
The authors tackled the lack of Kazakh speech data by creating an open-source corpus with 332 hours of transcribed audio, achieving a 2.8% character error rate and 8.7% word error rate in initial speech recognition tests.
We present an open-source speech corpus for the Kazakh language. The Kazakh speech corpus (KSC) contains around 332 hours of transcribed audio comprising over 153,000 utterances spoken by participants from different regions and age groups, as well as both genders. It was carefully inspected by native Kazakh speakers to ensure high quality. The KSC is the largest publicly available database developed to advance various Kazakh speech and language processing applications. In this paper, we first describe the data collection and preprocessing procedures followed by a description of the database specifications. We also share our experience and challenges faced during the database construction, which might benefit other researchers planning to build a speech corpus for a low-resource language. To demonstrate the reliability of the database, we performed preliminary speech recognition experiments. The experimental results imply that the quality of audio and transcripts is promising (2.8% character error rate and 8.7% word error rate on the test set). To enable experiment reproducibility and ease the corpus usage, we also released an ESPnet recipe for our speech recognition models.