Special signs in any other language than English

Do you know how to implement SPECIAL signs?
In Polish there are several of them. I got lexicon containing WORD+ SPELLING.
For example, there is word:
białagłowa
with spelling:
b j a w a g w o v a (as in English)
Should I change every word in my data set (metadata.csv) to a such spelling?

I have found only this thread: #93, but nothing specific here.

Go to utils/text/symbols.py and make sure your character set is inclusive of your characters. You don’t have to transliterate everything manually.

1 Like

or you can use the phonemizer setting it true in config.json. It will handle the rest.

Please allow me to chime in: in symbols.py there is a list of phonemes. I am not a linguistic expert - is this a “universal” list of phonemes ?

In other words: can I use it safely with "phoneme_language": "de" or do I have to edit it and insert “german phonemes” (is there such thing?)

It should work for DE and any language covered by phonemizer.

1 Like