Show Thoughts

Speech recognition models are usually trained with noise and reverberation free speech samples.