JAY

Tea Enthusiast
Defender of Defoko
Damn Jay, back at it again with the shitty forum threads...

Anyway hey guys! I wanted to ask you guys your opinions on different English UTAU styles and your thoughts on UTAU English overall!

I mean no disrespect to reclist creators such as CZ, Delta, etc. I understand that many reclist creators in general have worked really hard to create effective, clear, and realistic reclists, and I have really high respect for these people. I just want to discuss the differences of styles and see what people personally think, and what the hype of UTAU English is really all about!

That being said, this is my thought on UTAU English.

I personally am not a huge fan of voice synth English as a whole. It's really quite difficult to make sound good due to the languages phoneme complexity. Japanese is simple phonetically and can be made to sound realistic with much more ease. That being said, good synth English isn't impossible. English banks keep on evolving to improve on what we already have.

In terms of reclists, CVVC makes most sense to me technically. I've worked with Aiko a bit, and I've found that it's really easy to use a CVVC reclist with English because the reclist is most phonetically similar to English. I personally don't think VCV is as effective for English. I've heard it done really well, like here for example, but I think VCV should primarily be left for the simplistic phonetic nature of Japanese.

Then there's VCCV, which I have a whole other opinion on. First off, I have a LOT of respect for CZ for making a really accessible, clear way to do English UTAU. That being said, I don't see the hype. I think that it's basically the same thing as Delta's CVVC but with a prepared oto. It's quite convenient, but people seem to rely on the oto too much.

UTAUloids like CZloid, Al!ce, etc. sound awesome with VCCV. They rock that. But I've heard other UTAU banks that are VCCV banks that sound choppy, unotoed, and not understandable because they rely so heavily on the base oto even though their samples aren't the same as CZ recorded hers. I of course understand that otoing English is a real challenge, but there's always room to edit otos to make them better.

I honestly think the type of reclist (of these three kinds) determines the quality of an English UTAU bank. I've heard some really incredible VCV and CVVC banks in my day that still sound better than a good amount of VCCV banks. To my understanding, CZ's intention was to create an easier way to achieve clearer UTAU English for all, and she has done that, assuming that VCCV users make their banks following her tutorials, and editing otos.
Please reply to this and let me know what you think! I want to hear your opinions! ouob
- Jay
 
  • Like
Reactions: na4a4a and keopi

SOHBlue

Master of Doom
Defender of Defoko
I'm quite certain that if enough effort is put into an English bank it can sound quite realistic. Yes, English has over 200 sounds while Japanese has, I believe, 50 or 150 (Sorry my memory sucks), but if someone is really set on making one they can. Yami (Yes, I know not many people like Yami, but he-he's not a girl-sings well) has an amazing English bank that sounds quite real to me. Listen:

 

CarrotBunz

Trash
Defender of Defoko
I'm really interested in English Voicebanks it might be noticeable But i don't really have an opinion on different reclists other than it might be harder for me to use since I'm already used to VCCV ^^"
 
  • Like
Reactions: JAY

JAY

Tea Enthusiast
Defender of Defoko
Thread starter
I'm really interested in English Voicebanks it might be noticeable But i don't really have an opinion on different reclists other than it might be harder for me to use since I'm already used to VCCV ^^"
Nah it's more about the oto. I think that if the oto is crappy then it doesn't really matter what kind of reclist you use. I don't think any reclist is relatively better than the others, it's about the bank itself.
 
  • Like
Reactions: Nohkara and Zoku

CarrotBunz

Trash
Defender of Defoko
Nah it's more about the oto. I think that if the oto is crappy then it doesn't really matter what kind of reclist you use. I don't think any reclist is relatively better than the others, it's about the bank itself.
Yeah the oto is very important and it's also about how the user feels about the reclist since many people like a different reclist and say that some is harder than the other
 
  • Like
Reactions: Nohkara

JAY

Tea Enthusiast
Defender of Defoko
Thread starter
Yeah the oto is very important and it's also about how the user feels about the reclist since many people like a different reclist and say that some is harder than the other
Agree 110%. I personally would either make a CVVC since it makes sense to me more than VCCV, but since VCCV is so accessible and has tutorials I could do that too. Who knows. Not that I'm actually gonna get off my lazy ass and do english ahahahha
[doublepost=1460589569][/doublepost]
I'm quite certain that if enough effort is put into an English bank it can sound quite realistic. Yes, English has over 200 sounds while Japanese has, I believe, 50 or 150 (Sorry my memory sucks), but if someone is really set on making one they can. Yami (Yes, I know not many people like Yami, but he-he's not a girl-sings well) has an amazing English bank that sounds quite real to me. Listen:


I actually really like Yami too! They sound great in this. But you're right, it is about the effort rather than the reclist. A big fancy reclist doesn't make English good, the quality of the bank makes it good.
 
  • Like
Reactions: Nohkara and SOHBlue

keopi

Ritsu's Renegades
Defender of Defoko
tbh i really enjoy english utaus using whatever reclist as long as theyre understandable
however, when using the voice banks.. i usually go for cvvc/vccv mainly because its what im used to. plus cz is a goddess and made some nice reclists.
i cant say vccv is the easiest to use, but with the right configuration, it sounds very nice!!
im still getting used to vccv, but its really close to being my preferred method.
 
  • Like
Reactions: Nohkara

JAY

Tea Enthusiast
Defender of Defoko
Thread starter
tbh i really enjoy english UTAU using whatever reclist as long as theyre understandable
however, when using the voice banks.. i usually go for cvvc/vccv mainly because its what im used to. plus cz is a goddess and made some nice reclists.
i cant say vccv is the easiest to use, but with the right configuration, it sounds very nice!!
im still getting used to vccv, but its really close to being my preferred method.
Understood! I really like how a lot of VCCV banks sound. I just don't think quality is limited to VCCV like so many users think. I feel like there's this mentality in the fandom that if "oh I made a VCCV that means it's good right" when it might not be that good depending on how much effort and skill is used while creating the bank, if ya follow.
 
  • Like
Reactions: Nohkara and keopi

SOHBlue

Master of Doom
Defender of Defoko
Agree 110%. I personally would either make a CVVC since it makes sense to me more than VCCV, but since VCCV is so accessible and has tutorials I could do that too. Who knows. Not that I'm actually gonna get off my lazy ass and do english ahahahha
[doublepost=1460589569][/doublepost]
I actually really like Yami too! They sound great in this. But you're right, it is about the effort rather than the reclist. A big fancy reclist doesn't make English good, the quality of the bank makes it good.

You're making an English one?! I can't wait to hear it! :-D
 

Zoku

making doper vocaloid music than the rest
Defender of Defoko
Let me start off by saying this, the type of reclist doesn't affect the quality of an English voicebank. It's the oto. In fact, one could argue that VCCV™ CZeng, CVVC, and Delta are essentially the same, just under different names and different aliases.

Between VCCV™ and CVVC, there almost the same. But one has transitional VC, ending, VC, a GuideBGM, and a base oto; the other doesnt exactly have all of those. However, any reclist can theoretically be adapted to VCCV otos.

Now for English voicebanks in general--English is a very complex language, as you said. Several dialects and accents, and there isn't even a universal pronunciation for several words in the language. There are dipthongs, variants; there are several consonant clusters, differences in the pronunciations of this and that before and after a vowel.

Needless to say, perfecting English vocal synth from scratch is a challenge. Even when working from samples, it's basically hell to start developing.

Now, CZ's lists focus more on accessibility and ease of development. Even she said it herself--CZ wanted to make a universal reclist that everyone could use. That's the primary reason why she released VCCV. Not so that more people can use it, but so that more people can create with it.

This isn't necessarily a bad thing, however. But it should be said that so many people refrain from making VCCV banks or even just using VCCV because it's hard. This is the list's fatal flaw--it's pretty hard to use.

'Prefect' is two syllables, separated as follows (written phonetically in X-SAMPA): pri ・Bekt . In VCCV, this is separated into five notes:
[-pr][_rE][Ef][fe][e k][kt-]
How come it can't be just:
[-prE][E f][fe][ekt-] ?

VCCV adds too many notes when it's unnecessary. Not to mention, the transitional VC oto is wrong in CZloid and the base oto.

I feel that VCV English (with transitional and ending VC) can be achieved and be even easier. Imagine the five notes in VCCV being reduced to only 3.

[- prE][E fe][ekt-]

It would not only accelerate usage but make it easier to use in general. VCV English would be so much easier and would promote usage.

If I may, allow me to take this opportunity to shamelessly self-promote.

My list, based on the Arpabet phonetic system, has VCV capabilities. VCV capabilities lessen the need for more notes, larger ust files, and simply more work on the user end. In addition, it can fix many 'smoothness' issues many people have with some CVVC/VCCV banks. My list is designed with the user in mind, rather than development. I focused more on ease of use so that English UTAU is easier. It's true that the list is incredibly big (22 pages according to Google Docs atm and counting), my goal is to make English much more accessible than VCCV is at the moment. pm me for the link

Honestly, it's the userbase that matters.

tl;dr: English vocal synth (specifically UTAU) has so much potential, but people are kinda scared and English usage could be a lot easier.

I probably trailed off and I should've added more pertaining to the topic more, but these are my three cents.
 
Last edited:

JAY

Tea Enthusiast
Defender of Defoko
Thread starter
Let me start off by saying this, the type of reclist doesn't affect the quality of an English voicebank. It's the oto. In fact, one could argue that VCCV CZeng, CVVC, and Delta are essentially the same, just under different names and different aliases.

Between VCCV and CVVC, there almost the same. But one has transitional VC, ending, VC, a GuideBGM, and a base oto; the other doesnt exactly have all of those. However, any reclist can theoretically be adapted to VCCV otos.

Now for English voicebanks in general--English is a very complex language, as you said. Several dialects and accents, and there isn't even a universal pronunciation for several words in the language. There are dipthongs, variants; there are several consonant clusters, differences in the pronunciations of this and that before and after a vowel.

Needless to say, perfecting English vocal synth from scratch is a challenge. Even when working from samples, it's basically hell to start developing.

Now, CZ's lists focus more on accessibility and ease of development. Even she said it herself--CZ wanted to make a universal reclist that everyone could use. That's the primary reason why she released VCCV. Not so that more people can use it, but so that more people can create with it.

This isn't necessarily a bad thing, however. But it should be said that so many people refrain from making VCCV banks or even just using VCCV because it's hard. This is the list's fatal flaw--it's pretty hard to use.

'Prefect' is two syllables, separated as follows (written phonetically in X-SAMPA): pri ・Bekt . In VCCV, this is separated into five notes:
[-pr][_rE][Ef][fe][e k][kt-]
How come it can't be just:
[-prE][E f][fe][ekt-] ?

VCCV adds too many notes when it's unnecessary. Not to mention, the transitional VC oto is wrong in CZloid and the base oto.

I feel that VCV English (with transitional and ending VC) can be achieved and be even easier. Imagine the five notes in VCCV being reduced to only 3.

[- prE][E fe][ekt-]

It would not only accelerate usage but make it easier to use in general. VCV English would be so much easier and would promote usage.

If I may, allow me to take this opportunity to shamelessly self-promote.

My list, based on the Arpabet phonetic system, has VCV capabilities. VCV capabilities lessen the need for more notes, larger ust files, and simply more work on the user end. In addition, it can fix many 'smoothness' issues many people have with some CVVC/VCCV banks. My list is designed with the user in mind, rather than development. I focused more on ease of use so that English UTAU is easier. It's true that the list is incredibly big (22 pages according to Google Docs atm and counting), my goal is to make English much more accessible than VCCV is at the moment. pm me for the link

Honestly, it's the userbase that matters.

tl;dr: English vocal synth (specifically UTAU) has so much potential, but people are kinda scared and English usage could be a lot easier.

I probably trailed off and I should've added more pertaining to the topic more, but these are my three cents.
YES THANK YOU I AGREE 100% I COULDN'T HAVE SAID IT BETTER BLESS YOUR SOUL
tumblr_mtzmeqC16Z1qcga5ro1_500.gif
 
  • Like
Reactions: keopi

keopi

Ritsu's Renegades
Defender of Defoko
Let me start off by saying this, the type of reclist doesn't affect the quality of an English voicebank. It's the oto. In fact, one could argue that VCCV CZeng, CVVC, and Delta are essentially the same, just under different names and different aliases.

Between VCCV and CVVC, there almost the same. But one has transitional VC, ending, VC, a GuideBGM, and a base oto; the other doesnt exactly have all of those. However, any reclist can theoretically be adapted to VCCV otos.

Now for English voicebanks in general--English is a very complex language, as you said. Several dialects and accents, and there isn't even a universal pronunciation for several words in the language. There are dipthongs, variants; there are several consonant clusters, differences in the pronunciations of this and that before and after a vowel.

Needless to say, perfecting English vocal synth from scratch is a challenge. Even when working from samples, it's basically hell to start developing.

Now, CZ's lists focus more on accessibility and ease of development. Even she said it herself--CZ wanted to make a universal reclist that everyone could use. That's the primary reason why she released VCCV. Not so that more people can use it, but so that more people can create with it.

This isn't necessarily a bad thing, however. But it should be said that so many people refrain from making VCCV banks or even just using VCCV because it's hard. This is the list's fatal flaw--it's pretty hard to use.

'Prefect' is two syllables, separated as follows (written phonetically in X-SAMPA): pri ・Bekt . In VCCV, this is separated into five notes:
[-pr][_rE][Ef][fe][e k][kt-]
How come it can't be just:
[-prE][E f][fe][ekt-] ?

VCCV adds too many notes when it's unnecessary. Not to mention, the transitional VC oto is wrong in CZloid and the base oto.

I feel that VCV English (with transitional and ending VC) can be achieved and be even easier. Imagine the five notes in VCCV being reduced to only 3.

[- prE][E fe][ekt-]

It would not only accelerate usage but make it easier to use in general. VCV English would be so much easier and would promote usage.

If I may, allow me to take this opportunity to shamelessly self-promote.

My list, based on the Arpabet phonetic system, has VCV capabilities. VCV capabilities lessen the need for more notes, larger ust files, and simply more work on the user end. In addition, it can fix many 'smoothness' issues many people have with some CVVC/VCCV banks. My list is designed with the user in mind, rather than development. I focused more on ease of use so that English UTAU is easier. It's true that the list is incredibly big (22 pages according to Google Docs atm and counting), my goal is to make English much more accessible than VCCV is at the moment. pm me for the link

Honestly, it's the userbase that matters.

tl;dr: English vocal synth (specifically UTAU) has so much potential, but people are kinda scared and English usage could be a lot easier.

I probably trailed off and I should've added more pertaining to the topic more, but these are my three cents.
tl;dr :wink:

nah but fr, this is understandable.. you made good points. and wow.. thats a lot of info lol
 

Kiyoteru

UtaForum power user
Supporter
Defender of Defoko
Reclists can be written for a lot of different purposes. I spend way too much time writing reclists, and there are some meant to reduce recordings, some meant to have complex usages, some meant to be ridiculous but fun to work with. I actually really enjoy using VCV english voicebanks such as Adrian, but there's so much work that has to go into the development just to reach that point. It's just not practical, and unfortunately we have to subject users to ridiculous and complex things to cut down on recordings. A balance must be struck.

Also, if Zoku's going to advertise his unfinished reclists, I might as well advertise mine :U
https://drive.google.com/folderview?id=0B3Gri677vajxZ1pJbmVBQWZHMjg&usp=sharing
 

kamonohashi

a small person
Defender of Defoko
Sometimes English voicebanks sound really understandable, but other times it sounds like an entirely different language. That's just my opinion, but I also have no idea how to use English UTAU.
 
  • Like
Reactions: JAY

Zoku

making doper vocaloid music than the rest
Defender of Defoko
Most likely to the fault of poor otoing and/or accents.
 
  • Like
Reactions: JAY

WyndReed

Dareka tasukete!
Global Mod
Defender of Defoko
The only way to sum up my opinion is that none of the current methods are any good and you are better off creating something new for yourself. Just my two cents.

And also that Utau isn't really set up to handle the complexity of English.
 
  • Like
Reactions: kamonohashi and JAY