Pavlichenko, Stelmakh, Ustalov, 2021. CrowdSpeech and VoxDIY: Benchmark Datasets for Crowdsourced Audio Transcription.