DSD is able to decode several digital voice formats from discriminator tap audio and synthesize the decoded speech. Speech synthesis requires mbelib, which is a separate package.