Resamplers and wavtools for screaming voice banks? (Hakaine Maiko/CO2 Like)

[heynotloid]

Yesterday I finished a bank of screams, and I'm wondering which resampler would suit it best, or which resamplers I should use to test the voice bank.

It's a single-pitch CV.

 

[SunnyWolves]

Moresampler with the U flag or tn-fnds with W-1 typically produce the best results for voicebanks of that type.

 

[FelineWasteland]

IMO tn_fnds with W-1 is best since in addition to removing pitch shifting, it also loops back and forth rather than back to front, so it works better with the irregular frequencies of those type of vocals. Moresampler with U is fine for short notes but IIRC it won't stretch the sample for longer notes since it's just playing it as-is.

 

[SaKe]

IMO tn_fnds with W-1 is best since in addition to removing pitch shifting, it also loops back and forth rather than back to front, so it works better with the irregular frequencies of those type of vocals. Moresampler with U is fine for short notes but IIRC it won't stretch the sample for longer notes since it's just playing it as-is.

Seconding this, also, use original UTAU, not OpenUTAU, to render the samples, as the two programs have slightly different engine noise.

 

[Kiyoteru]

Seconding this, also, use original UTAU, not OpenUTAU, to render the samples, as the two programs have slightly different engine noise.

This is factually incorrect. When using the same resampler and wavtool, you will get the same result, possibly with some minor differences between automatic and manuual crossfade.