eugeneloza
commented
6 years ago Also provide for manual initialization of gibberish sound, e.g. by providing SIRAT____KREENA_OROOK sequences to the gibberish interpreter for the phrase (and autogenerate those, of course).
Maybe, a good idea would be to finally implement gibberish speech engine. We need 5 languages to gibberish. Each word should be transformed into gibberish syllables (e.g. by a kind of Base64/Hash encoding it - so that every word (excluding dots/commas) would sound the same every time, but the amount of gibberish syllables may be different than in original word). E.g. if the "language" has 15 syllables, it would project all possible letters of the words in English onto these 15 syllables. Second, it shouldn't be too long or too short (variable, e.g. Kerf language should be very short). With automatic pitch-shift and other effects to support different "voice acting".