Open stevempotter opened 1 month ago
We are using a somewhat eclectic mix of two letter ISO 3166 country code and and ISO 639 two letter locale code, so that would be en-IR, for English, Ireland, but that sounds fine and would be great. Some competition for "Emily" :laughing: Someday we'll have to move everything over to an proper standard... but that is a problem for another day ;)
Great. Can you help with instructions?
Basically you will want another entry in sounds.json
like this one...
https://github.com/EdgeTX/edgetx-sdcard-sounds/blob/46ced91e122d99006cefbb7e743fb509ced63fd0/sounds.json#L32-L37
And audio files for all the entries in /voices/en-GB.csv @ https://github.com/EdgeTX/edgetx-sdcard-sounds/blob/main/voices/en-GB.csv - which would go in a new folder in /SOUNDS @ https://github.com/EdgeTX/edgetx-sdcard-sounds/tree/main/SOUNDS
perhaps un-creatively named en_ir
unless you want to ... name... the pack somewhat... in that case, it will just need the en_ prefix.
Bonus points if you can convince her to do /voices/en-GB_scripts.csv @ https://github.com/EdgeTX/edgetx-sdcard-sounds/blob/main/voices/en-GB_scripts.csv also.
If you want to do words/phrases not in /voices/en-GB.csv, that is fine... just need to add them (and /voices/en-US.csv also as that is basically a copy at this point).
Hopefully the release script does not mangle the audio when it tries to clip any trailing silence / normalise the audio, else I'll need to teach it some new tricks so it leaves that alone. Audio files need to be PCM Wave files, no higher than 32khz sample rate, mono.
Yell out if you need any more information, or if something isn't clear. ;)
Thanks.
I hope this audio processing (what is a release script?) is something you will do? Or am I responsible for that? If so, I will need some hand-holding.
Looking forward to en_ir-mairead on my TX!
Yeah, that will be my problem to deal with.
It's the release.sh
script that is in the repo root directory, that does all the audio trimming, normalisation and builds the zip files for each sound pack. Whereas generate.sh
builds all the wav files via Azure TTS that are in the SOUNDS directory.
Description
I'd like to be able to CREATE A voice pack in the following language(s) and locale(s): English in an Irish accent.
This is not a request to use AI to generate a new dialect, but to get help on how my wife, who is Irish, can record a voice pack of the correct format. She does some voiceover work and has good clear enunciation in a fairly neutral Dublin accent.
Thanks!
en_IRL
Help make EdgeTX better!