Beta version https://soundcloud.com/utaufrance/lyse-beta-sweet?in=utaufrance/sets/lyse-ufr-v01-millefeuille-diffsinger-demo

Song: Synthimentale www.youtube.com/watch?v=i72UaUSygvs
Music: @conslation
Lyrics: @utaufrance
Tuning: @Hibya-tweek-twix
CV: @valshamr

SWEET is a sugary mode that has similar power to the CORE vocal mode while offering a kinder tone. Its warm and lively highs will pop making it perfect for energetic love songs as well as softer ballads.

www.lyse.utaufrance.com/
utaufrance.com/

LYSE 0.5 SWEET

Beta version https://soundcloud.com/utaufrance/lyse-beta-solid?in=utaufrance/sets/lyse-ufr-v01-millefeuille-diffsinger-demo

Song: Synthimentale www.youtube.com/watch?v=i72UaUSygvs
Music: @conslation
Lyrics: @utaufrance
Tuning: @Hibya-tweek-twix
CV: @valshamr

SOLID is LYSE's powerful mode. It offers a solid and energetic tone that feels like it can't falter. The switch to head voice happens higher than in other modes, making its high range fierce and strong. This tone is perfect for emotional choruses and for songs with heavy instrumentals.

www.lyse.utaufrance.com/
utaufrance.com/

LYSE 0.5 SOLID

Beta version https://soundcloud.com/utaufrance/lyse-beta-soft?in=utaufrance/sets/lyse-ufr-v01-millefeuille-diffsinger-demo

Song: Synthimentale www.youtube.com/watch?v=i72UaUSygvs
Music: @conslation
Lyrics: @utaufrance
Tuning: @Hibya-tweek-twix
CV: @valshamr

SOFT is a smoother counterpart to the CORE vocal mode. Elegance and mellowness are kept and refined with a softer touch in the high range. Able to sound both melancholic or warm depending on usage, it's most suited for ballads and acoustic songs.

www.lyse.utaufrance.com/
utaufrance.com/

LYSE 0.5 SOFT

Beta version https://soundcloud.com/utaufrance/lyse-beta-cute

Song: Synthimentale www.youtube.com/watch?v=i72UaUSygvs
Music: @conslation
Lyrics: @utaufrance
Tuning: @hibya-tweek-twix
CV: @valshamr

CUTE is a youthful vocal mode with a playful tone, whose consistency and energy will capture your heart. This mode is best for J-Pop with a cutesy aesthetic, as well as EDM or hyperpop.

www.lyse.utaufrance.com/
utaufrance.com/

LYSE 0.5 CUTE

Beta version https://soundcloud.com/utaufrance/lyse-beta-core?in=utaufrance/sets/lyse-ufr-v01-millefeuille-diffsinger-demo

Song: Synthimentale www.youtube.com/watch?v=i72UaUSygvs
Music: @conslation
Lyrics: @utaufrance
Tuning: @hibya-tweek-twix
CV: @valshamr

CORE is LYSE'S default vocal mode, meant to reproduce Valshamr's natural singing voice. Elegant and mellow, with an airy high range - a versatile vocal mode that will suit a broad range of genres!

www.lyse.utaufrance.com/
utaufrance.com/

LYSE 0.5 CORE

Beta version https://soundcloud.com/utaufrance/lyse-beta-japanese?in=utaufrance/sets/lyse-ufr-v01-millefeuille-diffsinger-demo-language-extensions

song: はるかかなた
ust & tuning: @makkusandesu

song: Synthimentale
Music: @conslation
Lyrics: @utaufrance
Tuning: @hibya-tweek-twix

CV: @valshamr

Millefeuille's extra phonemes allows LYSE to sing in additional languages, with a slight French accent!

www.lyse.utaufrance.com/
utaufrance.com/

LYSE 0.5 JAPANESE

Beta version https://soundcloud.com/utaufrance/lyse-beta-english?in=utaufrance/sets/lyse-ufr-v01-millefeuille-diffsinger-demo-language-extensions

song: はるかかなた
ust & tuning: @makkusandesu

song: Synthimentale
Music: @conslation
Lyrics: @utaufrance
Tuning: @hibya-tweek-twix

CV: @valshamr

Millefeuille's extra phonemes allows LYSE to sing in additional languages, with a slight French accent!

www.lyse.utaufrance.com/
utaufrance.com/

LYSE 0.5 ENGLISH

Original by EZFG https://www.youtube.com/watch?v=c3s7wLrSJq0
UST & tuning by AshitaTsukiloid https://www.youtube.com/watch?v=nKvJbh3B1fE
VBs used: Flora Renji DS, Mimosa DS, Yura Sashoune DS
Mix, art by @mimsynth

Joyeux Noël Flora !!! 
Made for the 2024 UFR SECRET SANTA

[SECRET SANTA UFR 2024] Morose Make-Believe (EZFG) - Diffsinger cover ft. Flora Renji

Tuning based on a singing sample from @valshamr
French lyrics: @valshamr
Song: Sleepwalking Orchestra - BUMP OF CHICKEN (Dungeon Meshi OP1)

Happy Birthday LYSE!

For her 1st anniversary, we wanted to showcase the evolution of the project and how we built her. We started from scratch and changed software halfway through but we are so happy with where LYSE is standing today and looking forward to see where she will be in the future! 

LYSE 0.0 - 2023/02/22 - NNSVS
First model ever, she was still named "valshamroid" and used only two and a half minutes of singing data.

LYSE 0.1 - 2023/02/24 - NNSVS
Second model, she used some more data and was pretrained on a stronger model (Yura Sashoune NNSVS)

LYSE V0 - 2023/05/08 - NNSVS
First time she was named LYSE.

LYSE V1 - 2023/05/09 - NNSVS
She received a first round of audio normalization & had a lot more singing data by then.

LYSE Alpha 4.5 - 2023/12/15 - DiffSinger
One of the first LYSE models on Diffsingers, it used the same corpus as the NNSVS ones.

LYSE Beta 1.0 - 2023/12/31 - DiffSinger
Another round of audio normalization was done... and it wasn't a success! Beta 1.0 was supposed to be the release candidate.

LYSE β (Release) - 2024/01/13 - DiffSinger
The model that was used for LYSE's release on January 19th! It's actually Beta 2.0.

LYSE 0.6 - 2024/10/19 - DiffSinger
Her first public reflow model with a complete relabeling done on her corpus, she can now sing in 12 languages!

LYSE (UFR V01) - Model History

French vocal synth fan-project striving to unite the French-speaking community and promote French vocal synths!
We made LYSE!

LYSE 0.5 JAPANESE

Utau French Resources

Related tracks