Text this: End-to-end audio-visual learning for cochlear implant sound coding simulations in noisy environments.