Open tonyahowe opened 1 year ago
@tonyahowe I will do some research on the best way to get this working.
@tonyahowe There are a number of options here. I think we can start with something basic (and free/low cost) and see how we like it. Two options that seem pretty simple: Google WaveNet voices: https://cloud.google.com/text-to-speech/ natural speech enabled by deep learning, free plan will translate 1,000,000 characters each month.
Amazon Polly https://aws.amazon.com/polly/ For Amazon Polly’s Standard voices, the free tier includes 5 million characters per month for speech or Speech Marks requests.
Let me know if you have a preference for which you would like to test. I lean towards Google, as I find AWS pricing structure to be slightly more convoluted, but either should be fine.
Cool! Let's ask at our meeting today. It might be useful to keep it in-house (aws) since we're already using that, but let's see what everyone has to say.
We did not ask at our meeting. @jobrien62--what do you think? Google, or Amazon?
I think we agreed on Amazon to keep any future payments needed on the same platform?
@tonyahowe Lets revisit this at the next meeting.
We would like some way to enable text-to-voice of the main text (not annotations, for instance), or to stream a librivox audiobook. Ideally, this would be a tab at the top of each file or in some other way--like a floating box that follows the scroll and could be clicked on to text-to-voice the current page text. A librivox link could easily open from the tabs in a new page, but not sure about the best way to enable readers to listen WHILE reading. Maybe a floating iframe of the librivox page? Not sure... The end result is to have the ability, both for assistive purposes and for universal design, for readers to listen as they read.
Again, not sure what this would look like, but the winkfield-female-american.xml has a librivox link embedded in the text just after the title. Ideas for best doing this in the XML would be most welcome!