added support for [[ipa-phonemes]] in text

[contentnation]

Added support for inline phonemes in IPA format.
Especially useful for mixing languages or words that create wrong phonemes.
In this case "feature" an English word in a German sentence.
echo -n "Dies ist ein [[fˈiːtʃə]] des pull requests." | piper -m ...
Translation: "This is a feature of the pull request."

[chriseaton]

I just wanted to put in my 2cents here that I tested this merge request and, at least regarding the feature change subject, it worked as expected and compiled without issue (tried with a few different voices).

[synesthesiam]

I'd prefer to use SSML for this instead. There is a dedicated tag specifically for phonetic pronunciation.

[sassanh]

@synesthesiam Thanks for all your efforts on this open source project!
I noticed the development on this repository has slowed down a little bit, there were a single commit in last 2 months, which was a line added to the README.md to mention a vim plugin.
Having that in mind, maybe it's not a bad idea to merge this PR until you can invest on a better alternative.
At least one user has reported it is working for them and it fills a gap in the feature set.
I understand if you have other considerations regarding this issue.

I may be wrong about development pace, I just checked the insights page.

[synesthesiam]

This has been added to the new repo 🙂
https://github.com/OHF-Voice/piper1-gpl/