Recent studies have shown that deep learning architectures, particularly recurrent neural networks, can dramatically enhance the identification of language in audio samples.