search

Marak / say.js

TTS (text to speech) for node.js. send text from node.js to your speakers.
MIT License
1.49k stars 214 forks source link

Updated windows to support the `voice` parameter. #68

Closed jimbuck closed 6 years ago

jimbuck commented 7 years ago

This adds support for specifying a voice when using Windows, working with #67 to round out the Windows features in #46.

I do have some questions as well, mostly about how missing or unknown voices should be handled. I can only speak to Windows right now, but getting a list of available voices (#16) is very simple. I know the other platforms have their own challenges, but should a "get list of voices" feature be implemented? Even if just used internally to first check if the voice specified exists. If the specified voice does not exist then ignore the parameter. The ability to write truly cross platform statements that gracefully revert to using the system defaults when needed is pretty nice.

tlhunter commented 6 years ago

Thanks for the PR!

tlhunter commented 6 years ago

Regarding voices, my goal is to eventually make them sort of "disappear" for most users. Check out #63 for a proposed interface.

Basically, a normal user would select either a masculine or feminine voice, and if they want they can then provide something more specific (e.g. a platform specific voice). This would be akin to selecting fonts in a browser, e.g. there's a generic "monospace" or "cursive" font, or more specific "Consolas" or "Courier New" font.

tlhunter commented 6 years ago

Published in v0.14.0.