Does anyone record CVs like this?

CatastrophicMelons

Momo's Minion
Are there any CV voice banks recorded in a manner similar to VCVs or CVVCs? To keep the file size down, why don't we record a string of kana slowly and break them apart. For example, we could record kakikukeko and duplicate them to get each syllable. Need romaji aliases? Duplicate the hiragana duplicates and change it to romaji. Also, a basic hiragana bank with no extras has about 104 samples. With the string kana way, we can have only 21 samples if I did the math right. 20 strings of five and one string of four. Or we can reduce the amount to 20 if one is a string of six. What do you think?
 

violetR

Teto's Territory
This is called "Rentan" style. Oversea UTAU who have this method are limited to just NOIR. I know Shirasagi Merlot/Mellowly have this style on the JP Side. The only issue is double aliasing- one for Kana input and one for Romaji input, as you said.
 

violetR

Teto's Territory
Because Aliasing is pretty boring. XD You'll go through and make one set of aliases then have to duplicate them all and replace the romaji with the kana equivalent. Also not sure, probably because of the aliases..
 

Cdra

possibly dead
Global Mod
Supporter
Defender of Defoko
Rentan, huh?  I didn't know this had a name :'D

I do this, actually.  Halt Tanner CV2 is "pseudo-VCV", aka Rentan style with 34 strings (with consonants, yoon, and VCV-style vowels) oto'd for CV, a very small amount of VCV, and CV VC.  It does involve a lot of aliasing, but since I alias in Setparam it takes no time at all and I don't care xD  I think it keeps the samples more consistent and higher quality, and it also makes otoing easier in a way.

Kirrin wrote a nice little reclist for this style, but he'd have to give it to you owo  The one I used for that bank is a little outdated and derpish, so yeah.  It's 6-moora for the sake of the CV VC component--but if you didn't want that, you could make it 5-moora.

I know one of Kirrin's UTAU has this method as well but somehow I don't think it's Arco? *derps*

Here's my bank using this method--with the tiny VCV bits and the CV VC bits added in, too.  Because I like to make things more complicated than they should be :U Safe Word  ...and also I did this for Amon Ira: Destruction
 

Berin

Momo's Minion
I recorded my UTAU Beryll Helidora's CV Act 2 in strings so that the recordings have the same pitch and to shorten the recording time (recording it took me 6 minutes). I used a self-made cv-vc/cv reclist for it. It's also pretty useful because i can use her recordings for her act 2 cv-vc, too :-D The aliasing is a pain, you have to alias the whole set in romaji and kana and the otoing is also more annoying than normal cv =w= My reclist has about 34 strings, because it has a few extras (L, V ,F + vowelstrings). The vowel transitions in cv-vc are very useful for cv, they make the voice sound much smoother~ Basically, you can use any cv-vc reclist for that, you just have to skip the consonant+n parts.
 

kirrinx

Teto's Territory
ooh yes xD this thing i recorded Koa like this a few months ago xD and shortly after i recorded Arco's latest VB the same way and my list has around 40 samples in total , it has some extra sounds such as L's , V'  ,and F's , and includes all VCV vowels =3= and Vowel + n endings as well end breaths too ~

So as of now  ( from the knowledge gained from previous posts =3=b)

:uhuhu: Beryll, Halt, Arco, Koa, NOIR , Harmon E , and Reicheru Hoshikone are the few overseas recorded in rentan xD style ~ [sub]( i oto'd Harmon E latest Act3 , and Reicheru's Act2.1 so i would o3o know they are rentan)[/sub]

I used this type of VB in here =3=b (its my V.2 release song thingy )
its more fluid than CV , since my list has VCV vowels and a random assortment of VCV samples xD;; since i recorded Arco in strings like Berin did o3o//

[soundcloud]http://soundcloud.com/kirrin_kage/utau-v-2-0-release[/soundcloud]


for me recording in this style is easier -3- than recording like  i used to for a 100+ sample VB , since  i record in strings ~ the oto'ing :'3 isnt that hard just a little more time consuming ~
 
P

PurinPuff

Guest
I'm thinking of recording my voicebanks using Rentan encoding, although the novelty of it is intimidating to me. If I record a string, but oto it so it includes VCV beginning notes (ex. - あ) would that make a good method? Or is it redundant to do add those VCV notes?

Sorry, I always ambush people with questions. :sick:
 

Cdra

possibly dead
Global Mod
Supporter
Defender of Defoko
PurinPuff link said:
I'm thinking of recording my voicebanks using Rentan encoding, although the novelty of it is intimidating to me. If I record a string, but oto it so it includes VCV beginning notes (ex. - あ) would that make a good method? Or is it redundant to do add those VCV notes?

Sorry, I always ambush people with questions. :sick:

I think this is a good method because the [- a] will have the tiny beginning of a vowel on it, which makes it sound more natural, where the [a] won't. *nodnod*  I have all my vowels with the various VCV settings, so yeah.

But yeah, I'm with the other people here who say it makes the vocalization more natural, because it does in a few ways. *nod*
 

PandaLolii

Teto's Territory
Defender of Defoko
Yeepp, I use rentan too. Depends on my mood though lmao. And I switch from 5 to 3 moras, depending on what list style I'm up for.

As for alias problems, it's definitely more annoying? Lots of people from what I know usually use Aster's kit to add aliases to "standard CV" recording. You can't do that with rentan.

@ cdra -
→ Higher quality? No, not necessarily. It's like saying VCV's /automatically/ make your VB sound higher in quality. Not if you don't do it properly. u v u;;; sdkjklsj I guess that's out of context thought.... But it DOES help making things more consistent. Oto is your own opinion probably. IMO it's more annoying (slightly only though!), but that's just me?

AND OMG, LOVE TO THE "-" KTHANKS. < V >

Also
Am
I
the only one who thinks Pseudo banks shouldn't be still considered fully rentan? Because I'm not sure...Rentan isn't recording style only, is it? But configs too? ...?
 

Cdra

possibly dead
Global Mod
Supporter
Defender of Defoko
PandaLolii link said:
@ cdra -
→ Higher quality? No, not necessarily. It's like saying VCV's /automatically/ make your VB sound higher in quality. Not if you don't do it properly. u v u;;; sdkjklsj I guess that's out of context thought.... But it DOES help making things more consistent. Oto is your own opinion probably. IMO it's more annoying (slightly only though!), but that's just me?

AND OMG, LOVE TO THE \"-\" KTHANKS. < V >

mweemweemwee that's not what I mean :uhm:;;;; I'm not saying it makes it better automatically :uhm:;; please don't misunderstand!  It's just that the recording style seems to help...

See, I record infinitely better in strings.  It's easier to keep your consonants natural, and not so forced.  And it keeps my sis from making s's and sh's too long (sssssssssaaaaa is a pain :U).  And my r's are clearer and more Japanese-y.  That's why it makes them higher quality, I guess. :uhm:;;;  I just assumed that that also works for other people.  Plus the consistency factor, which makes them even a bit "better" (though this is contestable).  and I like the otoing because I can place my negative overlaps where they fall naturally ovo;

and yes the - is good :D

[sub]also I said I think :uhm: I am not stating facts I am stating opinions *hides*[/sub]
 

PandaLolii

Teto's Territory
Defender of Defoko
Oh whoop, I read the "I think" more of a "If I can recall from past information" type of think, opposed to opinion-wise lmao.

Um, I'm not disregarding the consistency idea of course, like..I'm not saying it's completely false to have better results and how rentan is "wow so stupid, this doesn't help one bit, lol what are you talking about?" But like for every recording style, lots of things can go wrong at the same time. XD;; So yeah, completely opinionated for sure. Personally, I'd avoid thinking it'd give higher quality results for everyone though. It'd depend on the recorder's capabilities.. :'D

waaah lol I'm sorry but I meant "Oto is your own opinion probably" as in it's your own call. The generic "you," not you as in Cdra. =v=b
 

Aline

Teto's Territory
Viocchan link said:
This is called \"Rentan\" style. Oversea UTAU who have this method are limited to just NOIR. I know Shirasagi Merlot/Mellowly have this style on the JP Side. The only issue is double aliasing- one for Kana input and one for Romaji input, as you said.

In the overseas side, one of Kaneyama Shou's CV voicebanks and Kanne Bikuta Strong append also have this method of recording a kana string and then aliasing.

I recorded Aline v2 that way, but ended up separating the phonemes because I had to re-record some messed up phonemes and because UTAU bugged the strings (I think the length was too long for it). So, I'd tell people to be careful with this method, as too lengthy wavs might end up making UTAU crash.
 

Similar threads