Google TTS does not have a speed parameter (EDIT: it has a `speakingRate` parameter), but maybe the default should just be ffmpeg's `atempo` option?