Cadencii Download

May as well give it a try. I heard it can do english quite easily, is that true?

The reason that Cadencii doesn't get much attention is that it is harder than UTAU to use, and has lots of bugs. Also has many locale issues, getting to not even work depending on your OS language, and version (I had lots of issues on Windows XP). Also, there are functions that will nto work properly, and it requires too much RAM memory to work, being slow to execute many tasks. It also doesn't have portamento AT ALL (at least, talking about UTAU and v-Connect), and you need to make it manually using the pitchbends.

Also, for some reason, it won't import some USTs at all, and some that it gets to import will create very strange bugs that can't be fixed. In other words, you would need to make a voice sequence from scratch. There are also some bugs with vsqs, but not as frequent.

About English, no. It isn't any easy to do English on Candecii, and may even be impossible. The way it works is optimized for VCV only, and it tends to bug with CV, and there are even some instances when trying to use CVVC, the whole vocie sequence will get mute due to some samples.

I have done lots of research, mainly because I want Cadencii to work. Nevertheless, here’s what I know. Cadencii is a program that allows you to use VOCALOIDs, UTAUloids, and other voicebank types with other resamplers and synthesizers. For example, you can use an UTAUloid with a VOCALOID synthesizer (as long as you actually have VOCALOID, because Cadencii doesn’t come with a VOCALOID synthesizer). However, the best part about Cadencii is the synthesizer it comes with, V-Connect.Stand. 

The V-Connect.Stand synthesizer produces better quality sounds than VOCALOID synthesizers, because vocals are smoother than VOCALOID1 and 2 (the only versions Cadencii supports at the moment). Its optimum voicebank type for Cadencii is a Japanese VCV UTAUloid, because it blends the notes (like a VCV in UTAU would, but with better sound and a smoother voice automatically). This is why it gets “buggy” with other voicebank types. For instance, with a CV voicebank, you are probably going to lose some of the consonant sounds, because there is no sound before it. This is also because of the way that CV voicebanks are oto-ed.

I myself have not tried a CV VC voicebank with Cadencii, and it’s funny, because I got to this thread by looking this question up on Google. I imagine that a CV VC voicebank created using Chezzie-Chan’s recording list would work good, because these voicebanks are recorded in “strings”, like VCV voicebanks are. That is just a personal guess, not based on actual fact, so please don’t quote me on it. However, for CV voicebanks, it does work somewhat, but it kind of just sounds like a CV voicebank that has been oto-ed badly. You can import your oto.ini file into Cadencii, but it still doesn’t sound great.

Some more great features include the ability to use VOCALOID-styled pitch bends. Cadencii’s layout is very similar to that of VOCALOID2, and this goes for the piano roll, which is exactly the same. Like VOCALOID, Cadencii automatically inserts vibrato into notes that are long enough. This is seen below each note, which has a line to indicate the pitch bends. It also allows you to use UTAUloids on .VSQs, and keep the pitch bends intact. It can also convert to different writing systems like hiragana to romaji automatically. Lastly, you can change things like portamento, breathiness, brightness and more the same way VOCALOID2 and VOCALOID3 do.

However, while Cadencii does all these nice things, it does them in a rather sloppy way. First off, getting the program to actually work took two tries, or a long period of time. The second time I got it to work by copying the exact settings some other Cadencii user had through a video. Second, it is very picky as to which voicebanks it will, and won’t have work. Most of my voicebanks do not work well in Cadencii at all. When I try to use them, it gives me a very low-pitch voice. However, with more “professional” voicebanks like Kasane Teto or Namine Ritsu, or even Nizimine Kakoi, it’ll work. Again, Cadencii can only work so good with CV voicebanks, so that still isn’t great.

Another problem is Cadencii is picky as to whether it wants to work, or doesn’t want to. If it doesn’t feel like it, then Cadencii will not work. The user interface is also somewhat confusing, and for VOCALOID newcomers (like myself), it’s going to take a while to get used to. Along with this, Cadencii does play .USTs, but only some, and it barely works with .VSQs. It also doesn’t support VOCALOID3 in any way yet, so you can’t use a .VSQx or a V3 voicebank in Cadencii. Lastly, V-Connect.Stand takes forever to render even the littlest bit of a sequence.

So overall, Cadencii, when just looking at the plus side, is an amazing program that any UTAU user will want to own. But when you get to the negative side of things, it sounds awful, and the negatives cancel out the positives because they are massive negatives. However, if you do get it to work, it’ll sound great, and you’ll be able to use this program the way it’s meant to be used. But for those who haven’t gotten it to work, it is very frustrating.

Also, sorry about the crazy long post, I just have a lot to share I guess.  ._.U

How can I convert the UST to romaji to hiragana (and vice versa) with Cadencii?

You might wanna write down a tutorial how to make it sing cause it won't sing for me.

I'm actually interested in this! I've heard a lot about it in the past (like I think 1 year ago?) but never got a grasp around it. If I'm correct, to convert your UTAU for Cadencii, you goto Tools>Create vConnect-STAND DB. You choose the oto.ini of the UTAU you want to convert, then the place where you want to store the converted UTAU, and press Convert. Not too sure, I'm testing it right now.

EDIT: It also warns you that you shouldn't convert if it's against the Voicebank's/Creator's rules which I think is a nice touch. Also, the conversion process is a little long. I will come back with the result of a very short song. (I like how it's like the VOCALOID2 interface and that you don't need to manually put rests.)

---

Did some more looking, you could import the UTAU as just an UTAU through Setting>Preference>UTAU Singers. Also on the Preference Window, you can go to the Platform tab to add resamplers and the wavtool. Some peoples' Cadenciis might not be working because they didn't use/put these. (WINE also located here) vConnect-STAND is a completely differently formatted synthesizer/resampler than UTAU ones, so if you have trouble with it, convert it. You can also change the font and language of the program under Appearance. (Default=English so originally a project from those who knew English?)

EDIT: Languages are:
-ja (Japanese)
-kr (probably Korean)
-zh-CN (Chinese?; not too sure)
-zh-TW (Taiwanese?; not too sure)

I'm rendering the whole song of Torinoko City (sung by my UTAU) and it takes a long time. (Using UTAU rendering system with bkh01 resampler) A short 2-second takes 1 minute. This rendering of Torinoko City is actually pretty long... o.o

It's been at least two hours and it's not even 15% done rendering... Rendering in Cadencii takes a long time.

I like cadencii as a converter, the newer versions are able to convert midis and vsq/ vsqx's into usts. Which is really helpful to me as utau doesn't like taken it in.

My UTAU won't work in Cadencii, while Ritsu and Sonone Linako can... I hope they can fix this

I just downloaded the latest version, & it won't open VSQXs or USTs. I read elsewhere that it can't read VCV USTs, but I don't know why it won't open VSQXs. I really need a VSQX to VSQ converter & was hoping this could do it, but no.

Cadencii is beta morelikeunfinishedearlyalpha, you can't expect it to work properly.

To open usts, first open them in utau and export to midi.
Import them as "standard midi" through cadencii.

I honestly cannot understand why people like it. First, it lacks crossfade tools so all the notes/envelopes must be manually edited.
Second, after the update (both utau and cadencii update) it can't open ANYTHING. VSQx for V4 is a totally new format, so it doesn't support that. VSQx for V3 is practically unfinished and doesn't always work. Opening UST files is hit and miss, HOWEVER vcv lyrics are not the cause because I have opened those just fine.

I don't see why people like vConnect-STAND so much. It's a World Toolkit based engine, much like EFB-GT for utau. It's mostly a proof of concept and is barely to be considered a finished/usable product. Missing many features and glitching out during analysis.
Furthermore, It's better at downpitching notes, where pitching them up often results in a harsh hissing noise. The quality of the samples are some of the cause, it's less prominent in cases where you have absolutely perfect samples.

All the statements before about it only working with VCV are incorrect, loss of consonant sound is caused by nit using the apply utau parameters option.

Pupperoni Pizza said:

VSQx for V4 is a totally new format, so it doesn't support that. VSQx for V3 is practically unfinished and doesn't always work. Opening UST files is hit and miss, HOWEVER vcv lyrics are not the cause because I have opened those just fine.

Click to expand...

The 2 VSQXs I wanted to work with are a V3 & V4. Both appear to be XML, just that the tags are different. Still, I can't import the V3 one anyways.

Just about all my USTs are VCV & can't open any of them, so don't know what else is different. But that's not a big deal for me - just wanted to be able to convert those VSQXs. Probably just a matter of time before I get V4 myself & then I can do it, but would like to do some covers with them in the meantime.

There is a VSQx to UST conversion tool, it will work for VSQx files with ASCII characters only (so no hiragana). but it's an option.
But it only works with V3 files.
You could try that out, perhaps?