When the recommended TTS fields changed from
style, we completely removed the
voice_name from the API description but it's still supported in the API itself. A more accurate way to describe the API would be to include the field but clearly mark it as deprecated.
(see also: an open issue to have our docs mark this more clearly. Other platforms should do it though https://github.com/Nexmo/nexmo-oas-renderer/issues/87)