The recent update from ElevenLabs features the introduction of a new Speech to Speech (STS) tool, designed to convert one voice recording to sound as if spoken by another, allowing users to manipulate emotions, tone, and pronunciation more effectively than with traditional text-to-speech (TTS) prompts. This innovation aims to enhance expressiveness in voice applications, such as professional narration and character portrayal, by replicating emotional nuances and providing reference for speech delivery. Additionally, changes to premade voices include the introduction of over 20 new voices alongside improvements to voice sharing and compensation features. Other updates include the Eleven Turbo v2 model for real-time interactions, the addition of a pronunciation dictionary to Projects for precise pronunciation control, and the capability to embed metadata, aligning with industry audiobook standards. These enhancements are part of ElevenLabs' broader strategy to improve AI voice synthesis and customization capabilities.