First off, using Eloquence, shorten pauses is definitely on. Like, it's really freaking nice. When entering your passcode, it tells you, for example, one of six filled. Two of six filled. I mean I don't need that, but okay. You can now control how much VoiceOver speaks when exploring text with your finger. Paragraphs, lines, sentences. You can now have the VO cursor visible or not. Now, VO recognition! If VoiceOver is focused on the dynamic island, you can: explore screen, and ask about screen. For actions focused on an image, you can explore image, ask about image, or get intelligent image description. There's an -on-device” mode, where no images are ever sent to a server. Descriptions may be less detailed, of course.

Live recognition is like, totally new. Default question switch button. Default is “What is this?" "you can set a default question that will be asked automatically when you start an ask session in live recognition.” You can choose to prefer speaking or typing your questions. You can use volume buttons to recapture image. "press either volume button to capture a new image while in a voiceover ask session.”. For "scenes" you can have text, Braille, or both. Same for text recognition. For Point and Speak, you also get sounds and haptics.

For Braille, there's a new Translate after a delay switch. Good if you want to pause after you type that capital sign and when you write your sentence, you find that capital is gone, you can turn that off. Braille Access can now open text files, it's not just BRF's. I have the feeling this is gonna turn into a Braille-first book reader.

That's all for now.

#wwdc