What are the difference of multi pitch voicebanks vs mono voicebanks?

[Akira Satoru]

I've noticed that some voicebanks have 3-4 pitch folders in them. Are they necessary? And if I were to use them, how am I going to utilize them? Does UTAU automatically know where to sort depending on how high I set a note for an UTAU vb to sing? or do I have to specify the note that it would be using a different pitch folder amongst the 3-4 pitch folders I mentioned earlier?

 

[Lovely peaches chan]

Yeah do have to add the pitch at the end of the sample alias for utau to recognize the pitch and you have to create a pitch map i think

 

[vlbonnie]

Having more than one recorded pitch for your voicebank isn't necessary, but it helps UTAUs to sound better at wider ranges. If you pitch up a voicebank recorded at B2 to B4, it might sound a bit strained, but if you also record the voicebank at higher pitches, you can have UTAU use a sample recorded closer to the pitch you're shifting to. That was kind of a trash explanation, but I hope it made sense...

Like Teto-chan said, most people put a suffix at the end of the oto.ini entries for a pitch (but it can also be a prefix.) A multipitch UTAU should have a prefix.map file that tells UTAU when to use what pitch. You can edit this file by going to Tools>Edit prefix.map.

So an UTAU recorded at A2, C3, E3 and A3 might have a prefix.map that looks like this:

And the oto.ini files for each pitch would be aliased as [- あA2] [- あC3] [- あ3E] [- あA3].
They don't necessarily have to have each pitch in different folders--for instance, Sonata's VCCV has all of her pitches together, just with different file names for each pitch. However, most people separate pitches into their own subfolders for better organization.
I hope this was helpful!

 

[cloudyyskie]

Recording UTAUs in lower pitches makes it sound more strained, robotic, and falsetto-ey in my experience. My main UTAUloid, Akagane Karou, was recorded in C3 (somewhat intentionally) to give his voice a more falsetto yet somehow somewhat masculine tone if that makes any sense.
I haven't noticed much of a difference in multi pitch vs. mono
pitch voicebanks, it makes it somewhat nicer sounding from what I've heard though.

 

[vlbonnie]

Imo powerscale/kire voicebanks give a really good example on the strengths of multipitch voicebanks, where you start with a normal low pitch and a belting/shouting high one. Or voicebanks that have falsetto/soft pitches for very high/low notes to give you a little more versatility in how the singer sounds

 

[Akira Satoru]

Yeah do have to add the pitch at the end of the sample alias for utau to recognize the pitch and you have to create a pitch map i think

Oh I see. That makes a lot more sense. I'll do my manual research on how to do that exactly for all types of UTAU vb.

Post automatically merged: Dec 20, 2021

Having more than one recorded pitch for your voicebank isn't necessary, but it helps UTAUs to sound better at wider ranges. If you pitch up a voicebank recorded at B2 to B4, it might sound a bit strained, but if you also record the voicebank at higher pitches, you can have UTAU use a sample recorded closer to the pitch you're shifting to. That was kind of a trash explanation, but I hope it made sense...

Like Teto-chan said, most people put a suffix at the end of the oto.ini entries for a pitch (but it can also be a prefix.) A multipitch UTAU should have a prefix.map file that tells UTAU when to use what pitch. You can edit this file by going to Tools>Edit prefix.map.

So an UTAU recorded at A2, C3, E3 and A3 might have a prefix.map that looks like this:
View attachment 11021
And the oto.ini files for each pitch would be aliased as [- あA2] [- あC3] [- あ3E] [- あA3].
They don't necessarily have to have each pitch in different folders--for instance, Sonata's VCCV has all of her pitches together, just with different file names for each pitch. However, most people separate pitches into their own subfolders for better organization.
I hope this was helpful!

Oh! Now I get it. It is really convenient that UTAU does that automatically. I'm fairly new to the software, so this really is an enlightening fact for me. Thank you for the concise explanation,

Post automatically merged: Dec 20, 2021

Imo powerscale/kire voicebanks give a really good example on the strengths of multipitch voicebanks, where you start with a normal low pitch and a belting/shouting high one. Or voicebanks that have falsetto/soft pitches for very high/low notes to give you a little more versatility in how the singer sounds

Recording UTAUs in lower pitches makes it sound more strained, robotic, and falsetto-ey in my experience. My main UTAUloid, Akagane Karou, was recorded in C3 (somewhat intentionally) to give his voice a more falsetto yet somehow somewhat masculine tone if that makes any sense.
I haven't noticed much of a difference in multi pitch vs. mono
pitch voicebanks, it makes it somewhat nicer sounding from what I've heard though.

Oh I see. So if I were to record a multipitch VB, how many variants do you recommend me to do?

 

[cloudyyskie]

Oh I see. So if I were to record a multipitch VB, how many variants do you recommend me to do?

Maybe two or three. Not too much work but it'll do the job nicely. Just make sure to space the pitches out properly, as in maybe do one in a lower range for you, then do one about your normal talking/singing range, and then a higher pitch. You can maybe even do a falsetto if you'd like!

 

[Akira Satoru]

Maybe two or three. Not too much work but it'll do the job nicely. Just make sure to space the pitches out properly, as in maybe do one in a lower range for you, then do one about your normal talking/singing range, and then a higher pitch. You can maybe even do a falsetto if you'd like!

Thanks! I'll follow your advice. I plan to create one without the falsetto, so that would be 3 variants in total.

 

[cloudyyskie]

Thanks! I'll follow your advice. I plan to create one without the falsetto, so that would be 3 variants in total.

Cool! I wish you luck!

 

[feelingiconiic]

Hi! If you don't mind, I'd like to offer some advice too! Basically, multi-pitch voicebanks not only add a wider range to voicebanks, but they provide different tones for the pitches! Basically what I mean by this is that the natural human voice shifts and moves muscles depending on the pitch, and change the overall placement of your mouth and throat. Adding multiple pitches, either a natural, soft, or powerscale voicebank will add much more realism in voicebanks. Also, while you might want to just do three pitches, you may need to add "transitional" pitches as well. For JADE☆SICK's bipitch voicebank, it's the pitches E3 and C4, however there is a noticeable jump in the voice's (however I recorded that voicebank while I was sick, so I can't really updated it) tone and pitch that makes it sound strange. Power scales definitely require these transitional pitches as well. Ahhh, I hope this helped somewhat, and if you need further support don't hesitate to contact me! It's 12:27 am and I have more finals tomorrow so I'm going to head to bed lol