Lule Sami Text-to-Speech

Finite state and Constraint Grammar based Text-to-Speech processing

View the project on GitHub giellalt/speech-smj

Page Content

TTS-møte 5.5.2022

Til stades: Katri, Sjur, Flammie, Inga

Flammie

Framover:

Test for underlesingsfonematiseringa:

echo "åhpadusradiojda" |  modes/trace-smj-txt2ipa.mode

Katri

Plan til korpusprosessering:

  1. Rename textfiles after soundfiles, find correct texts, rf. the P(x)
  2. Cut audio from unnecessary parts
  3. Edit texts to match audios
    NORMALIZE numbers/abbreviations etc. automatically to orthography first!
  4. Final normalizing of volume, denoising (or noise gate), removing echo etc.
  5. Force-aligning automatically (WebMAUS or Montreal) for finding sentence/word boundaries
  6. Check automatic alignments and fix boundaries if needed (Praat TextGrids)
  7. Split audios and texts to sentences/utterances
  8. Tacotron pre-process!

Framover:

Inga

Framover:

Sjur

Framover:

Opptak med svensk røyst

Neste møte

12.5.2022 09.30 norsk tid