UTAU Help otoing

kukuism21

Momo's Minion
Hello, I've been familiarizing with UTAU the past few months and am currently trying to make a VCV UTAU bank. Despite the tutorials I've been seeing, I can't really understand the exact way how to oto. I've been using OREMO and SetParam to make a cleaner and accurate voice bank and I have some understanding on where each part goes and its function, but I feel I'm really missing something. I heard from one source that the overlap had to be exactly between the preutterance and left side but I took a look at some voice banks and it's all different. So here are my actual questions:
  • To be entirely sure, preutterance goes between the consonant sound and the vowel, right?
  • Does the overlap really stay exactly between the preutterance and left side (forgot what it's called)?
  • Is there a diagram I can see for what vowels, hard consonants, and soft consonants look like when oto'd?
Sorry if this is seen anywhere else on the web that I may have overlooked. Thanks for the help!
 

Kiyoteru

UtaForum power user
Supporter
Defender of Defoko
With VCV, a lot of the process is simplified, since you don't need to take consonant type into account at all. For all normal [V CV] samples, you can refer to these images that I posted a long time ago in St. Defoko's School of UTAU.

Diagram
unknown.png


One real-world example
Screen_Shot_2018-04-02_at_6.06.18_PM.png


Generally, you can set the overlap to a fixed value like 80msec. Too short, and there won't be enough material to crossfade smoothly. Too long, and the crossfade will cause a noticeable dip in volume in the middle of the note. This is the value I've been using lately, but feel free to experiment with it.
Once the overlap is fixed, you'll be adjusting every other parameter to fit the particular sample.
Offset/left blank: Area from offset to overlap should be consistent/stable vowel, right before it fades out. Refer to the spectrogram to find horizontal bands (ie. no diagonal lines)
Preutterance: End of consonant, beginning of vowel
Consonant: Where the vowel begins to stabilize again
Cutoff/right blank: Before the vowel fades out

For a [- CV] sample, put the offset where the consonant starts and the overlap at 0. Placement of preutterance/consonant/cutoff are the same as before.

To learn more about the function of the oto parameters, you might want to read this guide: https://utaforum.net/resources/anatomy-of-the-oto.321/
 

Similar threads