I couldn't fit the lyrics perfectly into the syllable structure, and there were some words that just straight up didn't have a direct translation. The English lyrics displayed below the Latin ones in the video have been translated back from the Latin, to sort of show what I had to do to make it work.
@@user-jt7ic1je3rit's like old english. Despite it not being used as commonly as it was back then,it can still be learned by history and the people who have used it before
I use SynthV. It's essentially the same thing as Vocaloid; machine learning technology _is_ involved in SynthV, but importantly, unlike with most AI deepfakes that I'm aware of, the voice providers both willingly let their voice be sampled _and_ are being compensated for its use. Solaria's voice provider is named Emma Rowley, if you want to learn more about her. Edit to add: I would sing them myself, but they're out of my range, and I can't actually roll my R's 😔
@@efhiIt's hand tuned, yes. When you're working in English lyrics, the default settings don't sound _awful,_ but some tweaking improves them a whole lot. (And the default settings are the same for every note - there's no auto-adjust option, and the closest thing there is to one gives me results that actively sound awful IMO.) Working in Latin, though, is a lot more difficult. None of the voices I use are made for singing in Latin, so there's a lot of fine tuning needed - especially when it comes to the rolled R, which doesn't have a dedicated phoneme in the ARPABET chart that SynthV uses, so I have to approximate it by repeatedly alternating two other phonemes in quick succession. Pitch is its own matter; this cover is basically the most complicated pitch-wise of all the ones I've posted so far, because I really wanted to replicate the pitch bends in the original.
honey wake up crpyologicalmystic posted another latin cover of a mili song wake u
now vergilius can understand lyrics!
I remember coming up with touhou characters while listening to these songs, glad a Latin cover is available now!
Every time there's a new Mili song in Limbus I look forward to seeing your Latin covers of the songs. Very nice translation/vocaloid use!
this cover is really really cool, it kinda makes me wanna learn latin myself!!
I always await these with eager ears.
nice!
absolutely nice
Great cover!
Cool
Pense que estaria en español hasta que vi bien y era Latin, igual el cover esta buenisimo
Latin voice really fit this music. It sounds so much emotional. But why the lyrics is different from the original?
I couldn't fit the lyrics perfectly into the syllable structure, and there were some words that just straight up didn't have a direct translation.
The English lyrics displayed below the Latin ones in the video have been translated back from the Latin, to sort of show what I had to do to make it work.
@@cryptologicalMystic oh ok
@@cryptologicalMystic how do you even translate to latin? isnt it a dead language?
@@user-jt7ic1je3rit's like old english. Despite it not being used as commonly as it was back then,it can still be learned by history and the people who have used it before
do you use AI for the voice or do you sing it yourself because it sounds so good
I use SynthV. It's essentially the same thing as Vocaloid; machine learning technology _is_ involved in SynthV, but importantly, unlike with most AI deepfakes that I'm aware of, the voice providers both willingly let their voice be sampled _and_ are being compensated for its use.
Solaria's voice provider is named Emma Rowley, if you want to learn more about her.
Edit to add: I would sing them myself, but they're out of my range, and I can't actually roll my R's 😔
@@cryptologicalMystic is it hand tuned or does the program do all the work?
@@efhi It's hand tuned
@@cananrhys very nice, people need to understand the difference between voice tuning and AI
@@efhiIt's hand tuned, yes.
When you're working in English lyrics, the default settings don't sound _awful,_ but some tweaking improves them a whole lot. (And the default settings are the same for every note - there's no auto-adjust option, and the closest thing there is to one gives me results that actively sound awful IMO.)
Working in Latin, though, is a lot more difficult. None of the voices I use are made for singing in Latin, so there's a lot of fine tuning needed - especially when it comes to the rolled R, which doesn't have a dedicated phoneme in the ARPABET chart that SynthV uses, so I have to approximate it by repeatedly alternating two other phonemes in quick succession.
Pitch is its own matter; this cover is basically the most complicated pitch-wise of all the ones I've posted so far, because I really wanted to replicate the pitch bends in the original.
sextans meus est curiositas
😍😍😍😍