Let's start by talking about screen readers! A screen reader is a form of assistive technology that renders text and image content as speech.
Screen readers use APIs/frameworks offered by the platform (such as UIA on Windows) to know what content is there and read it out loud!
Screen readers exist on Playstation, Xbox, PC/Windows and mobile devices!
... but not the Switch. So we've already hit our first roadblock. 🤷♀️
Secondly, how games are developed and packaged aren't done in a way that plays nice with these frameworks so they can't actually be used in a lot of cases.
Meaning again, we've hit another roadblock.
Some platforms, like Xbox, do offer systems that can do speech synthesis to "fake" having a screen reader, which is great! But it's also not without it's draw backs.
First, they're platform specific, so if you're multi-plat, you're going to have to find multiple solutions.
Secondly, a lot of speech synthesis options have to be manually hooked up to each string. It's not like a screen reader where it will just work as long as you've formatted it correctly.
This of course, leaves a lot of room for human error.
There are also some cloud based options (like Playfab/Azure) that at least can be used on any platform, even the Switch!
But, there is still a lot of manual work and they aren't free - they're a service and it comes with a fee.
For some games, cloud services aren't going to work and the devs decide to go a different route, such as generating audio files and shipping them with the game.
This is actually what Naughty Dog did for Last of Us 2. They used Read Speaker to generate the files & shipped them!
So, where does that leave us as game devs wanting to make great screen reader experiences? Well, we need support.
Maybe you could start by emailing @UnrealEngine and asking what it would take to make this a supported feature instead of experimental 👀
So separating audio, I can meant from the video itself. We already support separate audio channels (which is how we do multiple languages for our big shows) but ...
@LesterCovax@brycej@LittleRabbtPaws@Twitch@awscloud@WatchMixer@Azure@ianhamilton_ 2/ For streamers using Xsplit, OBS or Native Broadcasting on the Xbox, although the sources are individual on the streamer side, when they make it to us, it's all collapsed together, making it all jumbled. They aren't sending their voice along like we do for big shows.