• Hey, guest user. Hope you're enjoying NeoGAF! Have you considered registering for an account? Come join us and add your take to the daily discourse.

Hatsune Miku & Vocaloid All-Stars Discussion | Concerts, Songs, Music making, etc

SimonM7

Member
Luka V4 English sounds nice from what I heard, still accented but not that much compared to her V2. :)
Most of the English voices require changing some phonemes anyway, the problem with Miku English is that most of her sounds are actually wrongly recorded (tons of L/R issues, volume issues, slurry pronunciation and stuff like that).

As for Gumi's trial, I'd upload it since it's just a trial but my upload speed is pretty much 20kb/s and the file is 1.6GB. :/

Yeah, Luka sounds mostly nice, it's just the odd few words that sound weirder than my Frankensteined versions of them did. Her "saw", if left unattended, is like "sauu". When I did it in the japanese trial I just put "sa", which she pronounced "saah".

But the English voice bank is *definitely* worth it, of course. In retrospect I'd prolly go loopy having to construct every word like I did in the trial. :D

My first try was pretty rushed, and I hadn't properly tuned my guitar for it, but it managed to sound pretty okay anyway. I'm working on a proper song from the ground up now.

I'd love to try out Gumi's trial, but if it's a hassle to upload that's okay.
 
I'd love to try out Gumi's trial, but if it's a hassle to upload that's okay.

Tell you what, I'll put the files on my laptop and I'll try to upload them using the university's connection this Monday since it pretty much has 50/50Mbps download/upload. It might not work since it disconnects quite often, but it's worth a try. :)

As for "saw" you could try a lot of phonemes, I think stuff like "s O:", "s Q" or "s @U" might work. Even "s V" if you want a sound closer to "saah".
 

SimonM7

Member
Tell you what, I'll put the files on my laptop and I'll try to upload them using the university's connection this Monday since it pretty much has 50/50Mbps download/upload. It might not work since it disconnects quite often, but it's worth a try. :)

That would be awesome! :)

As for "saw" you could try a lot of phonemes, I think stuff like "s O:", "s Q" or "s @U" might work. Even "s V" if you want a sound closer to "saah".

I'll give this a whirl. I'm still learning when to use what in terms of sliders and effects, too. The manual sure is handy, and actually accurate (!) now that I have the English voice bank as well. :) Only half of the phonetic symbols worked before.
 
That would be awesome! :)

Here you go, 14 days trial of Gumi English. :)

Her voice has a few quirks you need to manually fix, most of her vowel-vowel transitions are choppy for example and you need to add another sound to make it smoother.
Say for example "say I" is choppy, you need to use the phonemes "s eI j" "aI" to make it smooth. Same goes for some words ending with r (with the [x]@ phoneme like Q@, which is the sound in "are"), where you might need to add an extra "r" phoneme to smooth out the transition.
 

SimonM7

Member
Here you go, 14 days trial of Gumi English. :)

Her voice has a few quirks you need to manually fix, most of her vowel-vowel transitions are choppy for example and you need to add another sound to make it smoother.
Say for example "say I" is choppy, you need to use the phonemes "s eI j" "aI" to make it smooth. Same goes for some words ending with r (with the [x]@ phoneme like Q@, which is the sound in "are"), where you might need to add an extra "r" phoneme to smooth out the transition.

Awesome! Thank you so much! Superb timing, too, because I was thinking about expanding the backup vocals on this song I'm making, and I currently have Yukari on that, meaning there are a lot of words to frankenstein if I give her more to do. I think Gumi will sound perfect in this style, too.

edit: She sounds way better for this song than Luka does. :|
 

SimonM7

Member
I'm having the most awful of problems. This song I've been working on was nearing completion, and now it seems that the vocaloid tuning file was corrupted when my computer crashed. It doesn't matter if I roll back to an earlier version of the song in Mixcraft, either - it just crashes when I use the same tuning. Even if I copy the tuning to a new file, it gives the same result.

I dunno if this is a thing that happens and that can be fixed?

edit: Er. I reinstalled the vocaloids and now it works. SO WEIRD. Why would... Yeah I dunno. Prolly shouldn't question. :>
 
edit: Er. I reinstalled the vocaloids and now it works. SO WEIRD. Why would... Yeah I dunno. Prolly shouldn't question. :>

Huh, that's weird, never had that happen to me... then again I don't have Piapro Studio other than the trial version, so it might be something related to that. It's nice that you got it fixed though. :)

I've found an interesting video while browsing Youtube, it showcases every single English Vocaloid (+ Maika since she has some English phonemes) released up to January 2015 (so it doesn't include Luka V4 English and Cyber Diva, other than two short demos) with two songs, one official demo and one user created cover/original. While I can't say I like all the user created songs they chose it's still a nice video which showcases all the different types of voices available. :)

Video here
 

SimonM7

Member
While I can't say I like all the user created songs they chose it's still a nice video which showcases all the different types of voices available. :)

Video here

Very interesting! I'm noticing in the charts that they often have something in between the words they sing. I can't make out what it says, but I'm assuming it's a way to control the intonation of the next word? Are there any good resources for learning those things? The instruction manual doesn't seem to go into that stuff at all.
 
Miku Expo 2014 from New York is getting a blu-ray release.

I didn't know it would. I'm a little bit tempted, actually. It hits some bingo numbers in terms of songs I don't have in the concerts I own, and it'd be fun to see a western performance. I've seen some cam footage on youtube, and it seemed like a good crowd, too.

Just as a note that Amazon.co.jp and AmiAmi seem to be offering much better pricing than CD Japan. Both stores also guarantee a A3 side clear poster as an early preorder bonus; Mikufan claims CD Japan is offering this bonus as well, but I don't see mention of it in their product page.

You can also order this domestically in the US via Tokyo Otaku Mode, though they charge you immediately rather than when the item is available.
 

SimonM7

Member
Just as a note that Amazon.co.jp and AmiAmi seem to be offering much better pricing than CD Japan. Both stores also guarantee a A3 side clear poster as an early preorder bonus; Mikufan claims CD Japan is offering this bonus as well, but I don't see mention of it in their product page.

You can also order this domestically in the US via Tokyo Otaku Mode, though they charge you immediately rather than when the item is available.

Ah, thanks for the heads up! I had trouble making an account and ordering from the japanese amazon page, if I recall right, but I'll give AmiAmi a whirl.
 
Very interesting! I'm noticing in the charts that they often have something in between the words they sing. I can't make out what it says, but I'm assuming it's a way to control the intonation of the next word? Are there any good resources for learning those things? The instruction manual doesn't seem to go into that stuff at all.

You mean those short notes you can see for example in the Miku demo? Yeah, they just add that to have a less flat note, you make a very short note with the beginning of the word and then you stretch it on another note on a different key, that way you have a quick pitch change as if the word started on a lower/higher note. You can achieve that by manually editing the phonemes (you can use for example the "-" phoneme which extends the previous one, something like this). It's basically the same as editing the PIT parameter, but you're doing it with notes instead.

I don't think there are some real resources if something is not mentioned in the manual unfortunately, since most of the stuff available is catered to the Japanese audience. :/

EDIT: here's an example, I tried recreating part of that Miku demo song. First up is the completely flat one, second one is made using those short notes to make them less flat. The difference is subtle but it helps a lot imo (unfortunately Avanna is not really suitable for high notes like that, but it's the only English Vocaloid I have :/).
 

SimonM7

Member
EDIT: here's an example, I tried recreating part of that Miku demo song. First up is the completely flat one, second one is made using those short notes to make them less flat. The difference is subtle but it helps a lot imo (unfortunately Avanna is not really suitable for high notes like that, but it's the only English Vocaloid I have :/).

Oh yeah, wow. :O I've mostly been fiddling with the portamento and stuff, but I'll definitely try this, too.

The concepts may already exist in the manual, but I just haven't put two and two together. I'll have to give it a look.

With the vocaloids back in action after the reinstall, I'm nearing something I'm happy with with this song. I've decided to try to go all the way with designing an outfit for Luka and making a simple video for it, too. Nothing fancy - just so I get a feel for the full process. I'm having a lot of fun with this. :>
 

antibolo

Banned
Cross-posting a photo I just shared in the FigureGAF community thread. Took it a few weeks ago while I was *ahem* inspired by Magnet...

VeUGtg2.jpg
 

antibolo

Banned
If anyone here was on the fence on getting some of the Crypton vocaloids, Big Fish Audio (the official US distributor) is currently running a 15% off sale today with a one-time coupon code.

There's also a promotional discount for Luka if you own one of more of the other V3s, but I asked customer support and they applied the discount ($50!) on my cart because I was also ordering the Miku bundle.

In total I paid only $264.42 (before shipping) for both the Miku bundle and Luka!
 

SimonM7

Member
If anyone here was on the fence on getting some of the Crypton vocaloids, Big Fish Audio (the official US distributor) is currently running a 15% off sale today with a one-time coupon code.

There's also a promotional discount for Luka if you own one of more of the other V3s, but I asked customer support and they applied the discount ($50!) on my cart because I was also ordering the Miku bundle.

In total I paid only $264.42 (before shipping) for both the Miku bundle and Luka!

Awesome :O

I may bite. I'm making a song (and a video) with Luka at the moment, but I'll want to do a duet after this one.

edit: Meaning I'd want another vocaloid! I'm not gonna sing a duet with Luka. That would be so awkward I can barely type thinking about it.

edit 2: Hmm. No Megpoid though. I'm prolly looking to nab Megpoid.
 

antibolo

Banned
Awesome :O

I may bite. I'm making a song (and a video) with Luka at the moment, but I'll want to do a duet after this one.

edit: Meaning I'd want another vocaloid! I'm not gonna sing a duet with Luka. That would be so awkward I can barely type thinking about it.

Act fast, that sale is ends today (it's 3:33PM right now where they're located).
 
All the PowerFX Vocaloids got a permanent price cut and thus are now more affordable. :)

They're all English Vocaloids and Yohioloid also includes a Japanese voicebank (but then again I'm not really a huge fan of his voice :/).
Big Al and Sweet Ann are V2 voices (they come with the full V2 editor, they SHOULD be importable in Piapro but I'm not 100% sure), Oliver and Yohioloid are V3 voices (they come with the Tiny V3 editor and they're importable in Piapro Studio).
If you listen to their demos on the website just disregard Sweet Ann's first demo as it either uses and old unreleased V1 Sweet Ann or V1 Lola :p

Here are a few songs which I think showcase quite nicely what you can achieve with these voicebanks:
Sweet Ann #1, #2
Big Al #1, #2
Yohioloid #1, #2
Oliver #1, #2

EDIT: I've also added some new Vocaloid trials to my previous recap post, the new ones are in italic text. :)
 

SimonM7

Member
All the PowerFX Vocaloids got a permanent price cut and thus are now more affordable. :)

They're all English Vocaloids and Yohioloid also includes a Japanese voicebank (but then again I'm not really a huge fan of his voice :/).
Big Al and Sweet Ann are V2 voices (they come with the full V2 editor, they SHOULD be importable in Piapro but I'm not 100% sure), Oliver and Yohioloid are V3 voices (they come with the Tiny V3 editor and they're importable in Piapro Studio).
If you listen to their demos on the website just disregard Sweet Ann's first demo as it either uses and old unreleased V1 Sweet Ann or V1 Lola :p

Here are a few songs which I think showcase quite nicely what you can achieve with these voicebanks:
Sweet Ann #1, #2
Big Al #1, #2
Yohioloid #1, #2
Oliver #1, #2

EDIT: I've also added some new Vocaloid trials to my previous recap post, the new ones are in italic text. :)

Cool! Thanks for the heads up! I've been busy lately and haven't posted here as much as I'd like.

I noticed today on Mikufan.com that the set list for Magical Mirai 2015 was partially revealed by way of the official album release.

http://www.mikufan.com/magical-mirai-2015-official-album-announced/

DISC 1 (CD)
01. Hand in Hand / livetune feat. Hatsune Miku (Magical Mirai 2015 Theme Song)
02. Remote Control / Wonderful★opportunity! feat. Kagamine Rin & Len
03. Raspberry*Monster / HoneyWorks feat. Hatsune Miku
04. Unhappy Refrain / wowaka feat. Hatsune Miku
05. Nostalogic (MEIKO-SAN mix) / yuukiss feat. MEIKO
06. Ai Dee / Mitchie M feat. Hatsune Miku & Megurine Luka
07. Romeo and Cinderella / doriko feat. Hatsune Miku
08. Snowman / halyosy feat. KAITO
09. Love Trial / 40mP feat. Hatsune Miku
10. <bonus track>

Looks like it won't be a carbon copy of 2013 and 2014 songs wise, and with the bigger venue I'm thinking it's time for me to shell out for another blu-ray. :D
 
They've been busy behind the scenes! That's an excellent lineup already.

The duets must take especially long to choreograph and animate, so it's good to see they are pushing those even more.
 
Gah, been avoiding reading Magical Mirai news because I had to cancel my planned Japan trip this year. :/ But yeah, I guess another BD is in my future.
 
Everyone should make a point of watching Miku Expo Shanghai - it was livestreamed, and captures are widely available now.

It turns out I was wrong about the song selection for Miku Expos being limited by the smaller stage; this show proves that Crypton and Sega can rework the wide-stage projections from Magical Mirai for a single screen.

It's the first Miku Expo that wasn't beholden to any of the setlists from Japan, but rather just combined mass-appeal songs with no restriction. They're targeting China in a big way.
 

SimonM7

Member
Everyone should make a point of watching Miku Expo Shanghai - it was livestreamed, and captures are widely available now.

It turns out I was wrong about the song selection for Miku Expos being limited by the smaller stage; this show proves that Crypton and Sega can rework the wide-stage projections from Magical Mirai for a single screen.

It's the first Miku Expo that wasn't beholden to any of the setlists from Japan, but rather just combined mass-appeal songs with no restriction. They're targeting China in a big way.

Wow, exciting! :O

I'm a little wary of "spoiling" myself before Mirai 2015, since a lot of songs are kinda inseparable from my experience of seeing them for the first time in Mirai 2013, so I'd like everything there to be as fresh as possible going into 2015. I'll check out the setlist and see if there's a lot of overlap.
 
Wow, exciting! :O

I'm a little wary of "spoiling" myself before Mirai 2015, since a lot of songs are kinda inseparable from my experience of seeing them for the first time in Mirai 2013, so I'd like everything there to be as fresh as possible going into 2015. I'll check out the setlist and see if there's a lot of overlap.

The only new songs played in Shanghai were Chinese-language ones. Nothing debuting in Mirai 2015 will be spoiled.
 

SimonM7

Member
To those of you familiar with vocaloid tuning, any tips on how to get a good "oooh" sound? I'm trying to have Luka sing along to a part of the music by just going "oooh oooh oooh", but she's all "awoo awoo awoo". Cutting it short and keeping the "o" sound sounds really monotone and robotic, too.

EDIT: Ah! I just put "a" instead, and that sounds pretty good. :>
 

SimonM7

Member
That sounds really nice! And I managed to understand what Luka was saying, so that's always a huge plus when it comes to English Vocaloids :)

Thanks! And that's great to hear! When you're the one tuning, you get so indoctrinated after a while that it's easy to lose perspective on whether it's comprehensible or not. I noticed that a lot of word transitions sorta muddy the next word, or just clutter the transition and make it jarring, and that's probably where I had to do the most cleanup. Luka also tends to veer on the "e" sounding side on her "a's". "Eees!" for "as" and stuff.

There was quite a bit of just.. flailing about and trying to make sense of all the settings and stuff, too, and I definitely have a greater understanding that'll let me make a smoother thing out the gate now. Man, it's a lot of fun. I make some music myself, but I'm not a very good singer so it's almost always humorous stuff. Vocaloids kinda make all other genres possible. It's very inspiring. :D
 

Rayis

Member
Is anyone familiar with UTAU? Is a program like Vocaloid expect that it is free and you can make your own voicebanks with it, Kasane Teto is the most popular "UTAUloid" as they sometimes call the voicebanks.

There are some pretty amazing voicebanks such as

Gahata Meiji

Namine Ritsu

Matsudappoiyo

I even have made my own voicebanks, I made a female one and a male one

here they are singing Cendrillon. excuse their shitty Japanese pronunciation


Anyway, hello fellow Vocaloid fans, I'm glad to see there is a community thread about it, I've been a huge fan since 2009
 
Is anyone familiar with UTAU? Is a program like Vocaloid expect that it is free and you can make your own voicebanks with it, Kasane Teto is the most popular "UTAUloid" as they sometimes call the voicebanks.

I've tried making my own UTAU a bit more than 3 years ago, but never released it because it's a mess, full of uneven pronunciation and errors. I've recently tried to record a softer voicebank as well, but I only did it as a quick test and thus only recorded a few sillables. I don't think I'm cut for this stuff :p

Old UTAU
Softer UTAU

I do like how people record UTAUs in a lot of different ways, and they often record lower registers with a softer voice. It makes them a bit more realistic tone-wise even though the engine noise is often more noticeable than in Vocaloid :)
 

Rayis

Member
Your UTAUs sound pretty good, I like the older one better though because I find UTAU pretty terrible at handling whisper voices, though using different resamplers such as TIPS can make them sound better, Making UTAU's is fun but time-consuming as hell, recording can take a whole afternoon and then otoing them another half a day but hearing the final result is so satisfying, if enough work is put into them, UTAU can sound just as good as Vocaloids.

Also opinion wise on recording methods VCV>>>>CV VC>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>CV
 
Your UTAUs sound pretty good, I like the older one better though because I find UTAU pretty terrible at handling whisper voices, though using different resamplers such as TIPS can make them sound better, Making UTAU's is fun but time-consuming as hell, recording can take a whole afternoon and then otoing them another half a day but hearing the final result is so satisfying, if enough work is put into them, UTAU can sound just as good as Vocaloids.

Also opinion wise on recording methods VCV>>>>CV VC>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>CV

Yeah, TIPS distorted the long notes way too much though so I ended up using another resampler (can't remember which one though as I haven't saved the project).
The older UTAU is full of missing notes unfortunately, I was dumb and basically didn't record any "n + vowel or consonant" transition, and neither I recorded any variation of the sh/j/ch/ky sounds other than the standard "shi/ji/chi/ki". That sample is using a flag of something like G+6 to actually resemble my current voice, as my voice when I recorded it was a lot higher pitched than it is now :p
I should probably record a new one sooner or later, but OTOing takes a bit too much time imo which I currently don't have due to exams.

I agree on VCV being really good (it's what I used for both UTAUs), CVVC can work too but it's a bit of a mess to work with. CV is fine if you just want to record something quickly, but yeah, it's clearly more limited.

I'll stop now as I feel like I'm derailing the topic. Actually, to stay a bit more relevant, here's a demo of a new male English Vocaloid called Dex.

Video

He was supposed to be released by the end of September alongside a female Vocaloid called Daina, but they're probably delayed to September.
 

SimonM7

Member
...here's a demo of a new male English Vocaloid called Dex.

Video

He was supposed to be released by the end of September alongside a female Vocaloid called Daina, but they're probably delayed to September.

The differences in strength between different phrases is impressive, but man, male vocaloids sure often sound like they have a cold and are recording their singing inside a shoe.
 
The differences in strength between different phrases is impressive, but man, male vocaloids sure often sound like they have a cold and are recording their singing inside a shoe.

Yeah, I'm not really pleased with any English only Vocaloid 4 so far, they're all really nasal. So far Cyber Diva has been released and she has an extremely thin and nasal voice (she can sound nice in good hands though), and for the upcoming Vocaloids both Dex and Daina have that nasal quality. Same goes for Ruby, who's a really powerful and clear voice but still nasal :/

I'd personally like a voicetype similar to Sweet Ann's since she's stuck to Vocaloid 2 (the voice sounds a bit better if imported into V3/V4, but it still has some missing transitions).
 

SimonM7

Member
Yeah, I'm not really pleased with any English only Vocaloid 4 so far, they're all really nasal. So far Cyber Diva has been released and she has an extremely thin and nasal voice (she can sound nice in good hands though), and for the upcoming Vocaloids both Dex and Daina have that nasal quality. Same goes for Ruby, who's a really powerful and clear voice but still nasal :/

I'd personally like a voicetype similar to Sweet Ann's since she's stuck to Vocaloid 2 (the voice sounds a bit better if imported into V3/V4, but it still has some missing transitions).

Is it flexibility/dynamism that makes the Vocaloid 4's more prone to the nasal quality? I haven't been keeping up with vocaloids at all, and I jumped straight into Luka V4X for my own stuff.
 
Is it flexibility/dynamism that makes the Vocaloid 4's more prone to the nasal quality? I haven't been keeping up with vocaloids at all, and I jumped straight into Luka V4X for my own stuff.

Hmm, I don't think that's the case since there are JP voices for V4 which aren't nasal at all, I think it's more likely what their respective companies wanted to achieve with those voices. Which is a shame because while they do sound good I'm not a huge fan of nasal voices :/
 

SimonM7

Member
Hmm, I don't think that's the case since there are JP voices for V4 which aren't nasal at all, I think it's more likely what their respective companies wanted to achieve with those voices. Which is a shame because while they do sound good I'm not a huge fan of nasal voices :/

Hmm, yeah.

Is there any word on how Miku V4 is coming along? I dunno anything about whether there are like.. closed betas or whatever. I think I saw somewhere that she's nearing release though.
 
Hmm, yeah.

Is there any word on how Miku V4 is coming along? I dunno anything about whether there are like.. closed betas or whatever. I think I saw somewhere that she's nearing release though.

I'm not a huge Miku fan so I don't really follow her news, but I think I read that a closed beta will be given to certain people next month and that indeed they're prioritizing her right now. Maybe some new info will be available at Magical Mirai, but I honestly wouldn't know.
 

SimonM7

Member
I'm not a huge Miku fan so I don't really follow her news, but I think I read that a closed beta will be given to certain people next month and that indeed they're prioritizing her right now. Maybe some new info will be available at Magical Mirai, but I honestly wouldn't know.

Ah, that's alright. I should get more read up on this stuff myself.

I always imagined that I would get Gumi next. I want someone to create duets with. If Miku v4 is good, though, it may be a more dramatic contrast with Luka than Gumi would be. Gumi and Luka both have sorta hazy, somber voices. When I was using the Gumi demo and did some quick tests together with Luka, they really fit the same tone, though, which was also nice.

I dunno. It'll be difficult to decide. I would like to take Miku for a spin and see if you can wrangle something less squeaky toy-y and more natural sounding out of her than the demos I've heard.
 
Ah, that's alright. I should get more read up on this stuff myself.

I always imagined that I would get Gumi next. I want someone to create duets with. If Miku v4 is good, though, it may be a more dramatic contrast with Luka than Gumi would be. Gumi and Luka both have sorta hazy, somber voices. When I was using the Gumi demo and did some quick tests together with Luka, they really fit the same tone, though, which was also nice.

I dunno. It'll be difficult to decide. I would like to take Miku for a spin and see if you can wrangle something less squeaky toy-y and more natural sounding out of her than the demos I've heard.

Maybe something like Miku Dark would work if you want a less squeaky voice, but then again it kinda falls in the husky area both Luka and Gumi occupy. Nowhere near as much, of course, as it's still Miku.

What I really like about Miku's appends is that they genuinely sound different from the base voice, especially Miku Dark (probably her only voicebank I can appreciate), which is something that got lost in other Vocaloids (cough, Luka.) because people complained the appends were too different :/
 
Top Bottom