Valdis Vitolins
Valdis Vitolins
Fix your text [encoding](https://stackoverflow.com/questions/2477452/%C3%A2%E2%82%AC-showing-on-page-instead-of) and pass proper Latin or UTF-8 encoded text to eSpeak NG.
Is not reproducible in [development version](https://odo.lv/Espeak?eslang=-ven-gb&speed=160&output=-w&debug=-X&text=100000000000000000000000000000000&formmethod=get).
In this tests not letters but sequence of phonemes (sounds) is written, which are defined in `phonemes/ph_` or general `phonemes/phonemes` file. To Espeak these are passed as `[[abada]]` where surrounding...
If you will provide list of needed hashes, I may update test to accept more than one value.
Latvian or Latgalian? Probably need both.
My understanding is that [breath](https://github.com/espeak-ng/espeak-ng/blob/master/docs/voices.md#breath) adds (pseudo-random) fluctuations to formant to make it less like pipe (precisely harmonic) sound. Probably in different platforms this randomness starts differently.
What prevents you putting `[[h@l'oU]]` in the file?
Cannot reproduce buffer overflow. `1ᵉʳ` is pronunciated in that way because for `superscript e` there is no specific pronunciation set, and it is spelled out using [Unicode](https://www.compart.com/en/unicode/U+1d49). But for `superscript...
Proposed stress changes do not conform to pronunciation in: https://www.youtube.com/watch?v=pXC4fXqGVQA