Real quick: let's suppose I have an image that I'm writing alt text for. Within that image is some text that is stylized with unnecessary and superfluous punctuation, meaning it's largely readable to the average audience, but would sound very weird when, say, parsed by a text-to-speech client.
Is the best practice with alt-text to record the text displayed in the image faithfully, even if the text would not necessarily read properly? Or should I amend it so that the text reads in the way that most audiences would otherwise parse it?
Example in the image above: a title card from the most recent episode of NieR: Automata Ver 1.1a, in which the episode name has some strange punctuation around one of its letters. Should I record the text in the alt-text as "[Q]uestionable actions" or "Questionable actions"?
