More pitches or more appends?

[heynotloid]

This is a debate I've seen a lot... what do you think?

 

[SaKe]

More pitches. While appends can be useful, pitches are definitely more needed, since it nullifies the distortion caused by the resamplers.
Also, if you record pitches in a similar pattern to Mine Laru’s voicebanks, you can use your voicebank with TN-FNDS.

 

[Kiyoteru]

It depends on what you want to do with your voicebank and the kinds of songs you want to work on. Additional pitches for a neutral singing tone can smooth out the pitch transition and make it sound more natural, but then you still have only that one singing tone. Additional singing styles can make the vocals more expressive, and it's what I chose to do for Dominic's initial release.

If you want to save time on voicebank development, I would recommend recording a powerscale voicebank (also commonly called KIRE, named after Ritsu's KIRE bank). With soft low pitches and powerful high pitches, you get the benefits of spanning the vocal range more naturally, and also have the option to manually override the pitch when you want to make certain sections of the song be sung more softly or more powerfully.

If your voicebank sounds good with Moresampler, you can also make use of the tension flag to strengthen or weaken the tone. So if that works well for you, you would only need to record additional singing styles for particularly unique sounds that are hard to recreate with flags alone. For example, pure unvoiced whisper, real vocal fry, or screaming vocals.

 

[GreenPear03]

This is a debate I've seen a lot... what do you think?

I'm on the side of recording various appends at different pitches so that they can be combined into a multipitch and multiexpression VB.