Versions history
Version 1 - June, 2020
- Linguistic resources for Text Normalization, Phonetization, Alignment and Syllabification of SPPAS for the following languages: cat, cmn, deu, eng, fra, fraquebec, hun, ita, jpn, kor, nan, pcm, pol, por, spa, vie, yue.
- Data resources for face detection, face landmark and LfPC automatic annotations of SPPAS.
Version 2 - July 2020
- Updated data for face detection and face landmark.
Version 3 - Sept 2020
- Updated linguistic resources of Polish language.
- Add of this documentation.
Version 4 - Feb 2021
- A DNN model is added into Face Detection package
- The file fra.txt of LfPC package is modified – corrected by experts
- Lightness of LPC package hand pictures is adjusted
Version 5 - Sept 2021
- New acoustic model of Italian: it’s no longer a context-dependent
model. It’s a monophone model like for the other languages. The French
HMMs of
a~
andO~
, and the Naija HMM ofe~
were added to the hmmdefs. - The LPC file fra.txt is renamed cueConfig-fra.txt.
- The keys of the LPC vowels are coded with characters
b
,s
,m
,c
,t
instead of numbers. It’s not compatible with versions 3.x of SPPAS. - The keys of the LfPC consonant are coded differently, we now use the same as the previously defined ones for English.
Version 6 - Nov 2021
- New resources for Bengali language: vocabulary, pronunciation dictionary and acoustic model.
Version 7 - 2022
Updated resources for Bengali language.
Version 8 - 2023
Updated CuedSpeech resources (renamed from LfPC).
Version 9 - 2023
Added resources for Persian language.
Version 10 - February 2025 (under construction)
- Updated CuedSpeech resources.
- Added resources for Dutch.
- Added samples of audio files and transcripts in each language resource package.