Blendin' Vowels~

dyne 智優

Shoutbox Lurker™
Defender of Defoko
I don't have Vocaloid, neither have I touched it before. But somewhere off the Vocaloid wiki states that it's possible to blend vowels to create a new vowel which would be non-existent in its phoneme inventory by tweaking some stuff/the properties of the syllable e.g. Japanese a+e almost sounds like English cat etc. I saw this done once.
So, my questions are:
Is it really possible?(Optional to answer) and
How do you do that in UTAU?
 

na4a4a

Outwardly Opinionated and Harshly Critical
Supporter
Defender of Defoko
With utau, you can't layer sounds. however you can create a very strong crossfade between two vowels through the use of consonant velocity and/or note property editing.
 

oteto

Tetoholic
Defender of Defoko
Yeah, there is a crossfade option where you can specify the milliseconds of crossfade, and where it happens, by crossfading two vowels you can come up with vowel sounds the voicebankk doesnt naturally have!
 

dyne 智優

Shoutbox Lurker™
Defender of Defoko
Thread starter
With utau, you can't layer sounds. however you can create a very strong crossfade between two vowels through the use of consonant velocity and/or note property editing.
Cool! I didn't know you could do that with the existing settings! Thanks for the help:sing:
 

JVウタウ

Ruko's Ruffians
Defender of Defoko
UTAU's engine is based on a "resampler" and "wavtool" that get WAV files and join them together in sequence.
Vocaloid's engline is based on a synthesis engine that uses samples to layer sounds and smoothly transition between phonemes.
As a result, not everything Vocaloid can do with blending vocals can be done in UTAU.

However, as JeremyB796 said, a good crossfade and the editing of note properties (like overlap and stuff) might be able to produce decent results!
Blending vowels are easier with VCV and CV+VV voicebanks though due to the nature of those recording systems (having vowel-to-vowel transitions).

But that is just based on what I know, which isn't much ^^;
 
  • Like
Reactions: dyne 智優

dyne 智優

Shoutbox Lurker™
Defender of Defoko
Thread starter
Yeah, there is a crossfade option where you can specify the milliseconds of crossfade, and where it happens, by crossfading two vowels you can come up with vowel sounds the voicebankk doesnt naturally have!
Thanks again!:uhuhu:
 

dyne 智優

Shoutbox Lurker™
Defender of Defoko
Thread starter
God, why didn't I think of that?
UTAU's engine is based on a "resampler" and "wavtool" that get WAV files and join them together in sequence.
Vocaloid's engline is based on a synthesis engine that uses samples to layer sounds and smoothly transition between phonemes.
As a result, not everything Vocaloid can do with blending vocals can be done in UTAU.

However, as JeremyB796 said, a good crossfade and the editing of note properties (like overlap and stuff) might be able to produce decent results!
Blending vowels are easier with VCV and CV+VV voicebanks though due to the nature of those recording systems (having vowel-to-vowel transitions).

But that is just based on what I know, which isn't much ^^;

Thanks so much for the preferences of voicebank types!
 
  • Like
Reactions: JVウタウ

JVウタウ

Ruko's Ruffians
Defender of Defoko
No problem! ^_^ (also, not all CVVC voicebanks have vowel-to-vowel transitions, but those that do will also be useful for blending vowels, and I think all CVVC voicebanks should have vowel-to-vowel transitions because what's the point of realistic vowel-to-consonant transitions without vowel-to-vowel ones? X'D)
 
  • Like
Reactions: dyne 智優

dyne 智優

Shoutbox Lurker™
Defender of Defoko
Thread starter
No problem! ^_^ (also, not all CVVC voicebanks have vowel-to-vowel transitions, but those that do will also be useful for blending vowels, and I think all CVVC voicebanks should have vowel-to-vowel transitions because what's the point of realistic vowel-to-consonant transitions without vowel-to-vowel ones? X'D)
Ikr