Enhancing text-to-speech synthesis by controlling voice characteristics through prompt-based methods.