probably I don't understand anything?

[riot3121 (trying to reg)]

Hello, I'm riot3121, and I have a problem with using Otouchi Tsurara's voicebank
Otouchi Tsurara has 4 voicebanks (Hatsuyuki, Shirayuki, Shirokuma, Miyuki)
As I understand, this UTAU is CVC (judging by "a, あ" construction, but I'm not sure)
When I'm trying to use some of Otouchi Tsurara's voicebanks (I didn't try Shirokuma, Miyuki, but probably, their situation are the same), I can't hear any sound
The only one that works with changes for me is "Shirayuki F4", but her sounds are delaying (or some sounds are not cutting and playing like "sa, se, si, so ,su")

The demonstration:


I'm sure I'm doing something wrong, but what exactly?

 

[Dangosan]

I'd say that your issues are caused by the "little squeaks" glitch if I had to guess. The glitch is caused by incorrect locale settings that prevent the oto.ini from being parsed correctly, hence the silence and offset values appearing to be ignored. To defuse the glitch you have to modify your decimal separator to a point/period under Time and Region in your Settings app (Regional and Language Options in Control Panel). Alternatively, you can invoke the Region Settings dialog by running intl.cpl (demonstrated here).
ameya should consider getting rid of the rendering batch file imho it's outrageous that number formats cause software to malfunction in 2024

 

[riot3121 (trying to reg)]

OMG SHE SPEAKS!!!!!
I'm not sure it was your advice that helped me, or I was blind, because I just wrote "あ_A4" and it worked, but thank you anyway!

 

[riot3121 (trying to reg)]

Next problem I have, Otouchi Tsurara doesn't say some syllables (as example "て"), but the line is exists at the oto file:
< _た_てぃ_とぅ_て_と.wav=て_A4,4830,240,-490,90,0 >

I had a similar problem with Amanaki Tsuneki, but that voicebank doesn't say "あ" (and "て" too, if I'm not mistaken). Amanaki Tsuneki is a CV voicebank

 

[Dangosan]

Do the voicebanks have a prefix.map file? The prefix.map file contains the pitch prefixes and suffixes to be set for each note. If the voicebanks do not come with a prefix.map then you will have to set pitch suffixes using SuffixBroker.
I would like to take a look at the て_A4 sample in the visual oto editor to see if it is otoed correctly.

 

[riot3121 (trying to reg)]

Yes, it includes a prefix.map file, but it's on the general folder (where the "A3", "A4", "C4" are)
I'll try to paste this file to "A4"

Post automatically merged: Apr 10, 2024

And yes, it works, thank you again
But Amanaki Tsuneki doesn't have "prefix.map"
As I said, Amanaki Tsuneki doesn't say "あ" (I just noticed, "い" too), so there is a line from oto file:

あ.wav=,15.481,84.518,100.068,4.518,0.0 (doesn't work)
い.wav=,37.67,62.329,100.068,2.329,0.0 (doesn't work)
いぇ.wav=,0.0,100.0,99.274,40.0,0.0 (doesn't work)
う.wav=,0.0,80.0,139.274,40.0,0.0 (works)

 

[Dangosan]

Could you show me the oto editor waveforms for "あ", "い" and "いぇ"? I have a hunch that the offset values are greater than the duration of the samples, causing said samples to be skipped over. Either that, or the cutoff values are set incorrectly.

 

[riot3121 (trying to reg)]

Yes, there is