search

benetech / MathShare

MIT License
9 stars 3 forks source link

Choose voice #1005

Open benetechMatt opened 4 years ago

benetechMatt commented 4 years ago

Students can pick different TTS voices from a drop down in personalization settings.

benetechMatt commented 4 years ago

@jscholes @alexrcabral are there specific voices that work best for different LDs? @johnhbenetech is there a list we can choose from? I assume the voice options will be based on the system used.

benetechMatt commented 4 years ago

mockup of the settings. "Voice Selection" with a drop down menu, the default is 'female'. A toggle in the right column for adding it quick settings

benetechMatt commented 4 years ago

changing the quick settings toggle to a checkbox

johnhbenetech commented 4 years ago

Here are the AWS polly voices: https://docs.aws.amazon.com/polly/latest/dg/voicelist.html

If there are better suited voices in other services we can explore those.

alexrcabral commented 4 years ago

@benetechMatt It's not so much specific voices that work best for different LDs, so much as control over the voices -- namely voice, speed, and pitch, as well as adding pauses between words.

If you're interested in reading further: https://www.researchgate.net/profile/Ea_Draffan2/publication/39995271_What_do_we_need_now/links/53e348320cf2b9d0d8332706.pdf

jscholes commented 4 years ago

@benetechMatt Screen reader users will definitely have a prefered voice. For example, @sinabahram and I would find it to be an outright productivity hit to write code or read documentation with anything other than ETI Eloquence.

However, as the scope of this relates specifically to people not using a screen reader, the above probably doesn't apply. Some users will undoubtedly find certain voices more pleasant than others, and there may be some better suited to mathematics. But you'll only find that out by asking what people think, either of samples or of the chosen voices once they've been implemented.

Worth noting that some users may find it enjoyable or even comforting if the voice matches that of their Amazon Echo at home, particular if they have additional cognitive needs and find the Echo to be a supportive piece of tech.

alexrcabral commented 4 years ago

Updates from call today: 1) Separate settings for math vs text (rate, pitch, prosody/volume, pauses) 2) Verify standards on speaking out math (eg ClearSpeak) for the default