Comment by sockfish

I've always wondered if machine learning would help here - we already have language agnostic embeddings (LaBSE, LASER) - now all we need is a way to pronounce them. Perhaps you could learn one by training an autoencoder that is constrained somehow just like the human vocal chords are