Speaking issues Coqui TTS Tacotron2 DDC model

From Voice Technology Wiki
Revision as of 00:44, 13 November 2021 by Thorsten (talk | contribs) (Just an edit for demo)
Jump to navigation Jump to search


General

Most models are trained with a dot, exclamation or question mark at the end. So always end a sentence to avoid model synthesizing weird output.

  • TTS model: tts_models/en/ljspeech/tacotron2-DDC
  • Vocoder model: vocoder_models/en/ljspeech/hifigan_v2

Input string formatting

Phrases ending in "ah"

"ah" at end of sentence generally produces strange results. Short names produce a 12 second clip.

Examples:

  • Nelson Mandela
  • pergola
  • Demo

Mitigation

If at the end of the input, adding punctuation to the end synthesizes correctly:

Example "Nelson Mandela" > "Nelson Mandela."

Acronyms

To speak acronyms as letters it needs to be formatted as:

"A. B. C. news"

Not:

"ABC news" "A.B.C. news"

Mispronounced Words

  • video