UTAU appends

[Astral-Arts]

Here's a big question that I think has been discoussed on VO before, but I don't think I've seen it here:

What is your opinion on appends for an UTAU voicebank?

Tbh, I think appends are a great way to increase an UTAU's genre range, it lets you cover a lot of songs you normally wouldn't be able to cover well with just a main voice. No 1 voice is fit for everything, so I personally think appends help. But when I think about it more, I also start thinking about how some UTAU have upwards of 5-6 appends. Do you think that can be annoying after a certain point, or is it just another form of voice enhancement for UTAUs?

Just wanting some discoussion, don't mind me. ^_^ I'd love to hear your opinions on it.

 

[Aline]

I think two contradictory thoughts: primarily, all UTAUloids should at least have a weak/strong mode. But then I contradict myself and think UTAUs should only have appends if the voicers can make them.

So, it depends on voicer, I think. If said voicer can't strengthen or weaken their voice, they shouldn't force it. But if they can, they should go for it like bosses.

 

[SakisCookie]

Honestly, and sorry if this applies to anyone here, but I like appends. I just wish there was a wayto turn them into flags. Does that make sense? Like, it's all one voice, but somehow you add a flag and the voice becomes softer or stronger.

....actually, that'd be awesome. DO WANT.

 

[Astral-Arts]

Honestly, and sorry if this applies to anyone here, but I like appends. I just wish there was a wayto turn them into flags. Does that make sense? Like, it's all one voice, but somehow you add a flag and the voice becomes softer or stronger.

....actually, that'd be awesome. DO WANT.

Yeah, actually. I've thought about that a lot too. xD but flags just don't cut it atm. Maybe with some improvements to m4 and the other 'unreleased' new technology for UTAU flags will improve over time.


@Aline I do tend to agree, some voicers can't pulla it off. And I'm going to get flamed for this, but example A:

Spoiler

Utaune Nami. All of her appends sound exactly the same IMO. The only one I can even begin to tell is different is her soft append. And that just sounds like her normal voicebank with a BRE flag added. T_T

 

[Aline]

Honestly, and sorry if this applies to anyone here, but I like appends. I just wish there was a wayto turn them into flags. Does that make sense? Like, it's all one voice, but somehow you add a flag and the voice becomes softer or stronger.

....actually, that'd be awesome. DO WANT.

You usually can do it by naming them with a suffix, but this would make solo songs with them become pretty annoying, as you'd have to suffix the whole UST all the time.

Another way is to estabilish prefixmaps for them, to, let's say, make a weaker voice sing lower notes and a stronger voice sing high ones. But if it's too discrepant, the results might be disappointing.

And yes, I wish there was a plugin to make multiple singers sing in only one UST. Not only for appends, but for choruses, medleys and all that stuff ;_;

 

[Astral-Arts]

Another way is to estabilish prefixmaps for them, to, let's say, make a weaker voice sing lower notes and a stronger voice sing high ones. But if it's too discrepant, the results might be disappointing.

Actually, what I find is a great thing to do is to download a plugin for UTAU (i forget what it's called otl) that lets you use multiple resamplers on one ust. I plug in the usual resampler for the UTAU I'm using, and then add the 'flag' "/2" to the low notes so it becomes tips.exe, making the low notes more compatible.

And yes, I wish there was a plugin to make multiple singers sing in only one UST. Not only for appends, but for choruses, medleys and all that stuff ;_;

Thisssssss. My god. Do want.

 

[shinami]

[quote author=SakisCookie link=topic=454.msg2255#msg2255 date=1329073147]
Honestly, and sorry if this applies to anyone here, but I like appends. I just wish there was a wayto turn them into flags. Does that make sense? Like, it's all one voice, but somehow you add a flag and the voice becomes softer or stronger.

....actually, that'd be awesome. DO WANT.

Yeah, actually. I've thought about that a lot too. xD but flags just don't cut it atm. Maybe with some improvements to m4 and the other 'unreleased' new technology for UTAU flags will improve over time.


@Aline I do tend to agree, some voicers can't pulla it off. And I'm going to get flamed for this, but example A:

Spoiler

Utaune Nami. All of her appends sound exactly the same IMO. The only one I can even begin to tell is different is her soft append. And that just sounds like her normal voicebank with a BRE flag added. T_T

[/quote]

Oh gosh Rave you said exactly what I was thinking.

I really, really like well done appends - for example, there's this Japanese male UTAU (I think his name is Makamiya or something, I'm not on my UTAU computer so I'm unsure) who has a bunch of appends, but all are different sounding VCV banks of amazingness. He actually managed to do well on a whole bunch of appends that sound fairly different.

But I think with most people, two to three appends for a bank is starting to push it. With Kaisho, I'm possibly only planning two since I know beyond that would either be a different voice type or just inflections of another append. Some people just don't do appends well. Like, while your example is one of that they just aren't voice acting well enough to make them sound different, Sorane Yui recently released six appends for a cover of Division Destruction, and I'd say three sound great while the other three sound too similar to the other ones.

On Aline's note, I've considered suffixes a lot for integrating banks together. But that creates a reeeeally hefty bank in VCV.

On the note of affecting softness and intensity of a voice, V-Connect Stand does that really well. Ritsu songs done with it are phenominal and, just by using his Strong append, make him go from soft to intense. Though barely anyone has any clue how to use it. Someone needs to make good tutorials for it, ugh. DX

 

[SakisCookie]

Actually, what I find is a great thing to do is to download a plugin for UTAU (i forget what it's called otl) that lets you use multiple resamplers on one ust. I plug in the usual resampler for the UTAU I'm using, and then add the 'flag' \"/2\" to the low notes so it becomes tips.exe, making the low notes more compatible.

I have that, too. It's that thing that lets you use two resamplers at the same time.
Except I can't get mine to work with anything other than resampler/fresamp or resampler/TIPS. And the person that distributed it to me said she had never heard of a problem like that before, so she couldn't help me. Why does technology hate me? orz

 

[DianaTan]

Personally I really love appends. But I wouldn't do appends, that doesn't fit the UTAU at all, like if you have an UTAU who is totally shy, quiet and I don't know and gets an append who represents something like sexy or violent. No, this is just no for me. And bad made appends. As everybody already said if you can't recognize them, they're bad. And the amount of appends is another point to discuss about. I think more as six is too much. You make appends and appends but you them only for one song and in every song it changes? I would be really confused.

 

[LupinAKAFlashTH]

Let me say something right now that CAN be considered a jump to a different sub-topic under "what do I think about appends"

Don't make a new design for every append you make.

Like seriously, if you have to make a new design for your UTAUloid's appends, only make one append design.  Like Gumi EXTEND, she had one design for it, four different poses.
And even then the design isn't all that important cause creative freedom.  How often do you see, I don't know, Miku, in her original outfit, seriously

Upon the note of actually HAVING appends for an UTAUloid, I think the idea is just fine!  However, you have to understand somethings.
1.  An append is an append.  Its not meant to sound totally different from the original voice.  Its simply meant to change how the timbre of the voice sounds to fit other genres.
1b.  This argument can be argued against, but in all seriousness sounding similar to "other appends" is a good thing.  Its not a good thing if its EXACTLY THE SAME, however.  Miku's softer appends go like this, DARK: 10% breathy, low tone; Light: 30% breathy, middle tone; Soft 60% breathy, high tone:  this is something few people understand, is that the appends also are recorded on different timbres, which is something that is hard to control without practice.  If you're recording an append that you think sounds too similar to a previous one, you really need to do some sample-comparison before releasing a new one.
2.  Don't think just because your append is complete that your core bank can now sing those genres.  If it can, you integrated the suffixes, congrats.  Usually, people don't do that.
2b.  Don't ask for examples.  I'm not going to drop names.
3.  Ever get the feeling the append doesn't sound like the original voice at all?  Too much variance can actually blend in with the original voice, proving you wrong. 
3b.  There was a video where someone combined all of Miku's appends + the core some point after Miku Append came out, and it sounded like she was changing her timbre instead of singing separately as each append, like a chorus.  I'm pretty sure it used voice-morphing instead mixing together, so that probably effected the result, I'm pretty sure the result would've been totally different if even one sounded too unlike Miku.

...now someone is going to swoop in, quote me, and find some way to make my argument invalid.  Its not an argument, its an opinion, and I know some people will jump on twitter going on about how ridiculous my opinion is (because yes, this has happened before, I -stalk- twitter, don't run away and criticize me behind my back), but the problem is -its an opinion- so there is no right or wrong.  Thank you kindly in advance.

 

[PandaLolii]

In case you guys didn't know ( which it seems like so! ), Masao was the one who made that different sampler in one song plugin.

But oh my gosh, yesss. I agree with what LupinAKAFlashTH said about different timbres (???). To be honest, that's more of what I've always considered about banks on top of strongly different because real singers ...you know, kinda actually do that??? XDD
Though I don't quite understand you meant by 2, I might've read it too fast though. :'D

Anyway, I think I might just be repeating everyone else but I think about 2 or so for completely different appends is good enough. Depending on what voice type you already have? Utauloids having any one strong and then any one soft bank is good enough to me? You don't have to go making shoujo or shonen voices on top like how I've noticed more users than necessary have been "inclined" to doing. And yessss, to those who said you need good voice acting skills if you really want to make more vbs. Q A Q

I honestly think everyone said the basics of what I have on mind at the moment but in short??? Don't make appends all silly nilly just because you "can", and then I think I could settle with that from someone? u v u

 

[Astral-Arts]

[quote author=Rave link=topic=454.msg2264#msg2264 date=1329074024]
Actually, what I find is a great thing to do is to download a plugin for UTAU (i forget what it's called otl) that lets you use multiple resamplers on one ust. I plug in the usual resampler for the UTAU I'm using, and then add the 'flag' \"/2\" to the low notes so it becomes tips.exe, making the low notes more compatible.

I have that, too. It's that thing that lets you use two resamplers at the same time.
Except I can't get mine to work with anything other than resampler/fresamp or resampler/TIPS. And the person that distributed it to me said she had never heard of a problem like that before, so she couldn't help me. Why does technology hate me? orz
[/quote]

I think this is where the argument over personality also starts coming in, which is why I think it's best if your UTAU has a personality, to say that it's optional to use it /or that it is just for RP/character related puproses, not for song covers. That's what I do anyway.
Hmm...Really? I've never had that problem before either. 0_0 sorry, wish I could help.

@LupinAKAFlashTH I see what your saying there (and I totally failed at the very first point you made. Amagaku normal vs Moonlight append, woah man. Big difference) and i certainly wont say you're wrong or your opinion is crazy. Actually, I think your 2nd point is -extremely- valid. Why do you think that because your append can sing a genre that it means your core can? THEY'RE DIFFERENT. So I don't think you're off base at all.

 

[jellophish]

Personally, I like Appends :D It's a way for a person to make many voicebanks with a similar voice. I can't voice act very well for example, and I have all these UTAU ideas. Just no voicers. Plus, it's hard, as I've learned when it's not me. I don't like bothering people ;u;
So I make Appends of my UTAU Taiyo xD I gave her a soft/sweet voice, and an airy voice so far ouo
Appends also show different sides of an UTAU's voice o3o

 

[Pianodream]

I like Appends, except like, as you said, they all sound the same...
I think that a good UTAU voicer can make up to about 4 or 5 different-sounding appends, but the average number seems to be around 3.

(You remind me that I have to make Ciel's OPERA Append and NiKORU/NiKO's Sweet Lolita Append) o3o

 

[smeen]

I like the idea of appends. If you really add different voice types I think it's a good thing to try. However, going from CV to VCV isn't an append for me, even if you call it like that.

I also do not like how some people change their design into the Crypton-y futuristic mecha sl- plugsuit idea. (My apologies, bad habit.) And having a different design for every append is even worse. It is still one bank people.

However, what I truly dislike? Not being able to get only the appends/banks I want. No, I do not want every append in one folder, thank you. It's the biggest turn off you can give me. It's what made me not buy Miku's appends, since I only like two of them. Not all six.

 

[Dark3li]

I like the idea of appends (I'm making one for my utau xD) but, when you make a lot of appends they might sound all the same D: I think that 1 or 2 appends are ok

I agree with the outfit thing, I don't like when a utau have a different design for each append but well, that depends on what the owner of the utau wants

 

[Cdra]

I'm gonna pretty well rehash what's been said here

Append voicebanks are just that: voicebanks that APPEND the original, basic (or core) bank.  They aren't for just... giving your UTAU a new dimension.  They're for enhancing their ability to sing, giving them room to change timbre during a song like a real singer.  It gives them more freedom to sing and express their voice, and gives them more potential for emotion.  Appends let UTAU try out a little bit of a new sound--but not a WHOLLY new sound.

My example: Halt's core bank is a fairly soft, mature voice.  Her one append is still a very mature voice, but now a strong, loud version of it.  When I played them next to each other in the same song, they still sound like the same person singing--just in different modes of their voice, for effect.  I think that's really what appends are all about.  An append can sing a song on its own, sure.  But how to know if your append is /really/ an append?--have it sing back to back with your original voicebank.  If it sounds like the same singer switching vocal modes, you're doing it right.  If you can't tell a difference, you might want to rethink the append.  If they sound like two different people, well, maybe you made a new UTAU instead of an append! (that happened to me once. ;w; )

As long as you can still get voices that sound like the same singer, but not the same timbre of them singing, well, record all the appends you like, I say.  But don't redesign your UTAU for them all :3 just like... do what Myst does.  Have a tiny change for each voice.  You know, like how Aiko Soft keeps her hood up?  I kind of like that way of doing it...

...I plan to make several appends with different timbres so that Halt can sing rock and metal songs with all the correct timbres required there.  I think it'll take like, 2-3 new appends. P:

Also sometimes I use masao's patcher plugin thing (that you mentioned) to put some parts of a song on different resamplers for an append-like effect.  TIPS usually makes my voices softer, and fresamp or w4u will make them stronger.  So, there's another idea...

tl;dr - If your append sings back to back with your core voice and it sounds like one singer changing their timbre, then it's an append.  If it doesn't sound any different or they sound like two different people, it's not.  That's pretty much the only thing in here that's terribly pertinent

 

[MystSaphyr]

I'm all for appended banks, but not really for calling them "appends" XD;

Like, idk, "appends" and "acts" seem to be the last 2 Crypton-style trends that are still going strong in this fandom. I guess now you might get the occasional hipster who will use the term "Extends" because it's Internet Co. and not Crypton, but... there's just so much else you could do xD;

And the extra outfits never made sense to be. I adore alternate designs and I do like most of the append designs I've seen, but to change one's entire outfit simply because one is singing in a different tone...? Even alternate version/act designs makes more sense to me, since as a singer matures I'm sure their sense of style might change, I guess..?

I'm abandoning Aiko's SOFT Append because I've been having fun with thinking up new names for her other banks. So far I have RockLoud (which currently is her core bank, since she's always been a bit loud), BalladSoft and PopMellow. I might end up adding in AriaBold for opera if things go well, since opera isn't just a loud voice but a loud voice with a specific pronunciation/thickness to it (which is why Tonio and Prima most often sound like derps trying to sing in other genres). Each name corresponds to a general genre that the bank will suit best (and AriaBold is a small typography pun because why the hell not)

 

[willowispembers]

Personally I kind of like the idea for having different outfits for appends... However, if the user has multiple append banks, I can see where it can get kind of excessive. I think it's alright to have alternate outfits if you maybe only have one append bank, but anything more seems like a bit much.

In terms of the actual voicebank though, I really like them in that they sort of expand the range for an UTAU, it gives the user more options to work with, and it's fun to experiment with your voice a little while recording I suppose. xD

[sub]Though that's just my opinion... orz[/sub]

 

[Hentai]

I am with Myst on this one. The idea behind Appends is great, recording more samples to give your UTAU more diversity. That being said I don't like the name "append" "act" or "extend". No need to take our naming conventions from Crypton especially when the idea of "act" implies the first release was a dismal failure we would all rather ignore.

My issues with Crypton aside its your UTAU you should come up with your own name for your additional voices, something unique, representative, or maybe a signal to the kind of music that bank is specialized in.

RockLoud indeed Myst, rock on Aiko!