Depending on the speech provider you are using (Edge, Google, Azure, Amazon, ElevenLabs or even just basic Windows TTS) you may find occasions where the speech engine is not pronouncing a word correctly.
To adjust this, you can edit the file : EDCoPilot \ User Custom Files \ Pronunciations.txt and add entries for words you would like to override / change, to force it to say a word using the phonetic version you provide.
Note : as of Mar 20, 2024 the location has now changed of the file. It was previously created at EDCoPilot \ working \ Pronounciations.txt if a user created one. I have modified EDCoPilot now to use the User Custom Files version, and to always generate a default file, making it easier for users to edit the pre-existing file. Any entries you had in a working folder version will have been added to the file in User Custom Files.
Just add a line for each string / word you want to override as follows:
<originalstring>=<newstring>
EDCoPilot will match the original string, case sensitive and including any spaces, with the phrase it will use for generating the speech, and will replace with the newstring.
Example lines:
(space)4(space)=(space)four(space)
This will speak 4 as "four" - note the (space) should be an actual space
(space)a(space)=(space)A(space)
This should make the speech engine pronounce the "a" as "ay" and not "ah"
Advanced : multiple overrides
You can also provide alternate overrides of strings as well such as :
Thargoid=Thar-goid|murder flower
Note the | to denote an alternate override. With this line EDCoPilot will sometimes pronounce Thargoid as "Thar-goid", or sometimes as "murder flower" (EDCoPilot will randomly select from the options you give it)
You can have as many alternates on a line as you like, delimited by the | symbol. You can use this to add more variation to speech strings, such as how it announces the name of your wing-man, for example!
Comentarios