The company set out in 2017 to create artificial intelligence that would produce the most realistic artificial voices in the world. The software the company has created requires a few minutes of audio recording, which the artificial intelligence and machine learning then disassemble and mimic. Deep learning algorithms sort the voice and its characteristics to allow it learn on its own.
Lyrebird's main usage is within the commercialization industry. The technology is useful for companies that want to represent their brand with a specific voice. The algorithms that Lyrebird uses are able to produce different voices, and are also able to differentiate between two separate voices. The algorithms in use are PyTorch and TensorFlow which are deep learning frameworks on Amazon Web Services (AWS) to train the artificial intelligence models.
Lyrebird can be found on chatbot assistants, audiobooks, hotlines, video games, text readers, and several other similar voice applications.
Alexandre de Brébisson