Command Line TTS

I will keep on accidentally referring to it as CLI TTS - so I just named the repo that to get it out of the way

This is a basic Command Line based Text-To-Speech application. This application interfaces with Amazon Polly, and requires the use of standard voices. It has the ability to use SSML (Speech Synthesis Markup Language), and the configuration has the ability to set up text to replace.

Replace Text

The replace text section of the configuration allows you to set specific strings that will be replaced by other strings. This can be very useful for when you want to use SSML and set a shorthand for the beginning tag and ending tag. This can also be useful for having 'macros'/'commands' with different responses

Note:
It is advised to only replace specific combinations of characters, and not common combinations/single characters
For example don't replace "e" with "𝔼" otherwise the tag "__" will break because it will now be <𝔼mphasis l𝔼v𝔼l=\"strong\"> instead of <emphasis level=\"strong\">
You can mitigate this issue by surrounding the letter with a space like this: " e "

Voice Prefixes

The voice prefixes section allows you to set a prefix that will switch the voice to a specific Polly voice. Then, if you use that prefix at the beginning of a message, the application will switch to that voice until another prefix is used.

AWS

Using this requires an AWS account - https://aws.amazon.com/ You can use a free account - Please do read and check the limits and such though. You will need to go to Identity and Access Management -> Access Keys You will get an Access ID and Access Secret from there.

References:

• Amazon Polly Description of Limits/Price
• Setting up a user and how to use IAM

Twitch

In order to use this to interface with twitch, you will need to create an application on Twitch's Dev Console

Once you have created an application, you will be given a client ID and a Client Secret, you will place these in your tts-config.conf

Default configuration

aws-region= "US_EAST_1"
aws-access-id= ""
aws-secret-key= ""
connect-to-twitch= false
twitch-channel= ""
twitch-app-client-id= ""
twitch-app-client-secret= ""
replace-text {
  "**"= "<prosody volume=\"x-loud\" pitch=\"low\" rate=\"slow\">"
  "/*"= "</prosody>"
  "*/"= "</prosody>"
  "~~"= "<amazon:effect name=\"whispered\">"
  "/~"= "</amazon:effect>"
  "~/"= "</amazon:effect>"
  "__"= "<emphasis level=\"strong\">"
  "/_"= "</emphasis>"
  "_/"= "</emphasis>"
  "++"= "<prosody volume=\"x-loud\" rate=\"x-fast\" pitch=\"x-high\">"
  "/+"= "</prosody>"
  "+/"= "</prosody>"
  "!!"= "<say-as interpret-as=\"expletive\">"
  "/!"= "</say-as>"
  "!/"= "</say-as>"
  " - "= "<break time=\"300ms\"/>"
  "<3"= "heart emoji"
}
default-voice= "Brian"
voice-prefixes {
  "Sal:"= "Salli"
  "Kim:"= "Kimberly"
  "Bri:"= "Brian"
}

Todo:

• Make closing tags less annoying
• Manage exceptions better
• Probably unify the config into one file
• Add configuration for audio output destination