New UTAU, with new engine

You guys, this does NOT work with Yuugi's new Act 2. I'm switching back again.

It kinda works pretty well Hayao at some places.
But at others it just breaks him OTL Must be the VCVs, I'm probably going to switch back before I export and mix this.

i have learned that nanashi going above E4 i sound like a keyboard |D which i sound epic in toeto XD

with the fresample anyway with normal the resampler i sounds scratchy~

i'm getting better at these oto.ini's using this fresampler :O

http://www.4shared.com/audio/LvGFGoKI/temp.html
like i said earlier i need to rerecord a few things like those dreaded r's =_= i can say japanese r's, spanish r and irish r's (i can do those better than i can english r's when talking normally XDD) but only if the r is in an actual word the sound alone like ra ri ru re ro it's troublesome so i'm gonna have to figure something out.

OOOH °A° I just tested it on my UTAU, she wasn't great on lower notes with the old resampler, but she sounds much better with this new one. And the long 'i's. They always sounded so scratchy, I had to lower their volume to have them sound decent, now I need to raise it again because they sound SMOOTH. Uwa~ *A* Overall, she sounds much more like me, between c4 and c5. *.*

But, I must agree with others here, the higher notes are so-so. She doesn't really sound nasal as it happens with other UTAUs, but from G5 up I pretty much cannot tell 'o' and 'a' apart <.< I could with the old resampler. Some consonants are a bit weird too, but maybe they could work with a different oto settings *needs to try*

OH GOD THAT THING KILLS QUALITY.

Tuwai's act 3 has the highest quality that my mic can reach and what came out? That: OTL

With normal resampler: Well.

But Mia's act one which is pretty bad quality sounds a bit better, but not more.


/switches instantly back

I had the same thing. Good to know I'm not the only one.

Well,thou know what's ze weird part?
I'm making an english UTAU voicebank right now (eewww oto.ini),and It doesn't sound that awesome,but the other voicebanks I downloaded work perfectly.
So either I'm doing something wrong,or...^_^"
Not to mention that not every voicebank sounds good.

so far the fresamp works lovely for me, of course my utau usually sings around c4 to lower sooo...

i did test a higher song and the nasally-ness made me laugh so hardddd

but in Natsukage, it wouldn't even render one of my "ta"s properly, it just sounded like a whistle <___> (and note that it was within the "normal" sounding range, just above c4)

at least it didn't do that screechy thing the older resampler did

Results for me thusfar:


Soft Aiko = improved smoothness (ohmygod no more scratches in her voice), less clarity

Normal Aiko = improved smoothness, MUCH less clarity, her A's sound like "apple" A's now

Ariella Act1 = SHE BROKE IT

Mathieu = only successful attempt, he lost his buzz and became more human ;u;


Well except for Matt, I think I'll wait until it's improved on a bit more...


EDIT: Also, Aiko sounds better when I adjust her otos... since she doesn't buzz anymore, I can lengthen the oto and it sounds a little crisper. AND I just read the other pages. Regenerating FRQ files? I have not tried this yet.
I guess I'll test editing her FRQs and OTOs before I mess with the fresamp anymore...

I love this~
It removed the screams and the mumbleness :'D I'm so happy right now

Except that the high notes sounds well, strained, but that was there from the begining xD

So I'm otoing Hana's English and when I test out some syllables they sound like unripe doo-doo covered banana. Wait, what? What's -

Guys, don't generate frq files with the new resampler (I'd forgotten to turn it off for frq generation) but play it with the old. It sounds horrible.

llllOTL tomi's first act doesnt even work with it.
at least i made some new voice banks... ;A;
see... its funny... cuz my normal voice that i recorded my new voicebank with made my utau a guy.... cuz my voice in manly Q.Q;

the freesampler softens out my newer voicebank, but i still cant tell much of a difference other than it sounding a hell of a lot better on low notes, and sounding.... beyond disgusting on high notes.

Kurisu act1 sounds more manry XP

for me i have no real qualms with this fresample it's amazing for nanashi right now anything above e4 though he sounds like a synth keyboard, but anything under e4 he sounds alright XD (c4 has never really been his range though so it makes sense for me), he does however start to sound a little compressed and hollow with some sounds in the c4 range.

but other than those i see no problem with it XD and remember this is still beta everyone XD

;.;..... OMG! Rikai sounds almost human now... ;u;
Thanks Aline-sama for putting up the link~ Even if its in beta... its works fine for me <3


Lol but she sounds funny in the higher notes XD;

Hmm...I think I'm going to test it...Thanks Aline

Listened to all the successful attempts.

Don't here a difference.

:I

try defoko
http://www.4shared.com/audio/ErSNWYLM/chaining_intention_defoko_new_.html

zinfandel wrote:try defoko
http://www.4shared.com/audio/ErSNWYLM/chaining_intention_defoko_new_.html

...Defoko with a BRE0 flag?

I've heard her sound way better and more human.

Yeah, still seeing no reason to so much as touch this resampler. It looks like it does more harm than good.

Though I might get the complete version just to smooth out scratches in 911's Whisper. >:

na i dont use flags much XP

zinfandel wrote:na i dont use flags much XP

I think Crossfrown was refering to what she sounds like, Zin. |D
Honestly, the fresampler likes applying an automatic BRE 0. You just can't see this flag.
And I actually do hear that strange nasal-ness in her higher notes.

oh lol,

here how kurisu sounds with the new one
http://www.4shared.com/audio/jaFQ-ygk/Fukkireta.html

Now Utau works, I tried this new sampler...


...I'm in the same case as smeen T__T

Just tested it XD
http://twaud.io/Dfm

Wow, Katarina sounds less robotic now :D I didn't have trouble with the new resampler, it works well