Unpopular Opinions

[uncreepy]

My humble opinion is that "Galaco" is really hard to pronounce. (In both English and Japanese, which is "Gyarako".) It seems like when I try to say it, each syllable comes out slowly and painstakingly, like trying to talk after having just been stung by 100 bees on my tongue. Each time I say it, I try it with different intonations but nothing ever sounds right. Gal-eə-co, Gal-LA-co, Ga-leə-co, Gala-co...

Sorry, Galaco fans.

 

[DefiantKitsune]

it's like "galaxy", right?

that said, the Japanese version you posted definitely took me some tries lol

 

[crtstatic]

I have an opinion that all Vocaloids are good. Sure, they’ll have their bugs, but computer software IN GENERAL has plenty of bugs. Its just the nature of technology

Saying that a Vocaloid “isn’t good” because of some small bugs isn’t a very good statement. Think of them as little quirks in the software

SONiKA with her muffled voice? Think of it as a voice that a REAL human couldn’t have, and that its perfect for a robotic type of voice for a sci-fi or autotuned song

There’s probably others here I could mention, but my point still stands

 

[EPROM]

–––

 

[Stigbn]

I think vocal synths should strive for both realistic voices - and especially in English we need more 'understandable' voices - and voices that sound synthetic.
Miku sounds often synthetic and a little weird to most non-vocaloid-fans, but without her we wouldn't have such a big fandom and probably not nearly as many voices. Her voice is unique - like it or not. But at the same time, I hope that more realistic and understandable voices, especially English (and other larger languages) will continue to be developed.
It's a bit like when the first Synthesizers came out. In the beginning they were always trying to make them sound like a 'real' instrument (violin, horn or whatever), but at some point in the 80'ties many realised that a synthetic synth-sound in itself could be just as interesting as a 'real' instrument, and today we see many trying to recreate the old synths like minimoogs or similar. I wonder if Miku will be recreated in some fancy synth in 40 years or so?

 

[EPROM]

–––

 

[Stigbn]

EPROM:

What I meant was, that in the future, they might recreate her voice with a more advanced technology - not recreate the exact vocaloid technology, because I agree that the vocaloid technology is already very outdated. I just think that her voice might live on in other kinds of technology.

And yes, you can have realistic without understandable, but every time I play an Engloid for someone not into vocaloids, the reaction is that "she's hard to understand" or something like that - so in the future I hope for a little better pronounciation (also without having to edit for hours).

 

[TheStarPalace]

Don't know if this is really an opinion but the posts above reminded me of Neutrino AI...an attempt to make a super realistic sounding voice synth with little to no user input (it's not even a program, you just input notes and lyrics and a wave file is automatically generated).
Neutrino can create very realistic results with little user input, but it tunes everything the same way, basically giving the same singing style to everything....not ideal for a lot of music...also you'd have to use autotune or something similar to actually change how the vocals sound.

It's been trending quite a bit in the Japanese vocal synth community, but I can't help but feel that the lack of an interface and no editing kind of removes the magic of using a vocal synth in the first place, at least for me.

And, the final product doesn't sound like Kiritans voiceroid or utau at all. To me it just sounds like some random lady, even if it is relatively realistic. I wouldn't recognize it as kiritan if her name wasn't attached to the product. Funny enough, neutrino uses the same samples as her utau does, I think?

The point of this post is that realistic doesn't equal better. All voice synths will sound unrealistic to a degree (especially those made by chopping up audio of a singers voice and stitching it back together) and either way it won't appeal to people who don't like the idea of programs like that in general. Also people were bringing up Miku as an example, which just confuses me because she was never meant to sound realistic. Judging her by how much of a real singer she sounds like is like judging a fish by its ability to climb a tree! Vocals such as Cyber Diva, Cyber songman, Sachiko, Fukase, even the V5 starters to a degreexl have more natural tones (once read a review of the cybers on a music site that said they were almost perfect so at least some people agree with me)

Thanks for writing things that made me think.

 

[Rylitah]

I don't really have anything to add to the conversation other than NEUTRINO does have an editor tool now that allows actual tuning. (Here's an English guide on how to use it with DL links.)

(Though I guess my unpopular opinion since I've seen multiple people say this is that NEUTRINO Kiritan does sound like her...? I actually think it's her UTAU that doesn't do her voice any justice (though I haven't heard it in a long time, I may be wrong but I remember it not sounding much like her), but NEUTRINO and Voiceroid are very similar -- the only difference being that Kiritan speaks in an emotionless tone, which is on purpose. So you could say that her NEUTRINO is out of character, but it very much sounds like her to me.)

 

[morrysillusion]

i never post here but i feel like i have no where else to just quickly throw out this opinion lol so

- i dont like eleanor forte's design, i dont think its fitting to her voice. its a nice design by itself but, her voice just feels so... basic/neutral to such a complex and 'over the top' design in comparison? i dont wanna say i hate the anime look for her english american tone, bc im sure i could get on that style like i do with dex/daina, but man. i just do not like her design at all. when i say her voice is basic/neutral i dont mean boring or bad, i just mean it sounds like, to me, the most average and genre neutral voice- which is good for her use as one of the 'main' vocals for synth v. but her design.... hm. would love to see it in a general sense but just simplified.

- i am a v flower fan so here two: 1) i think her V3 design gets her genre/tone across THE BEST visually and while i like her V4 and Talk designs to some level bc she is cute, i think theyre a bit nonsensicle with some details and dont fit her as well as V3. 2) too many producers put flower in the wrong range (too high) and then dont even tune her and it sounds soooo bad and i wish ppl realized this... some actually potentially good songs are ruined by her being strained and flat.

- also really dont think V5 is bad. it is missing a few things from V4 but it honestly has functioned all the same for me and i often go finish my files in V5 because i think the expression tools are great! its really useful for someone like me who often found myself exhausted by all the hand drawn pitch bends and things i couldnt figure out how to do without drawing in 4 different parameter windows. i think they knew what direction they were going in, in terms of the actual music production audience and making it easier for those who have never touched vocaloid before. i think they really only failed in...being too late/not pushing that properly to begin with.

 

[FluoroLime]

Vocaloid is actually the worst of the singing synths rn. Like Cevio and Neutrino have this cute engine noise plus realistic vocals (with even more realistic vocals coming with Cevio AI). UTAU offers so much flexibility in tuning (pitch bends and append banks come to mind) for free. Synth V honestly changed the game for me with its ability to draw pitchbends right on the notes.
Idk I'll always love the Vocaloid characters, but the engine.... yikes.

 

[andantina]

- i dont like eleanor forte's design, i dont think its fitting to her voice. its a nice design by itself but, her voice just feels so... basic/neutral to such a complex and 'over the top' design in comparison? i dont wanna say i hate the anime look for her english american tone, bc im sure i could get on that style like i do with dex/daina, but man. i just do not like her design at all. when i say her voice is basic/neutral i dont mean boring or bad, i just mean it sounds like, to me, the most average and genre neutral voice- which is good for her use as one of the 'main' vocals for synth v. but her design.... hm. would love to see it in a general sense but just simplified.

Big fat agree. That overly complicated hairstyle, and weird dress, and everything...shes too standard. I agree with you on not minding an anime-ish look for English synths, but even the "anime style" has variety within its bubble. I mean, think of all the options character designers have! They could go with a Western-inspired style like this or this, a more mascot-ish figure thats not entirely human. (or even something non-anime, like the ele.os fanmade design). but instead they went, "nope, lets create yet another Floating Moe Goddess (C)!". And that's not even considering her supposed personality. Eleanor's voice and character description are complete opposites. The voice is very neutral, while also coming across as really lively and spunky (at least to me). It sounds like a voice that a tomboy or really energetic person would have. But noooo. "Eleanor is old fashioned and loves antiques". It's so contrary its almost funny.

 

[peaches2217]

Big fat agree. That overly complicated hairstyle, and weird dress, and everything...shes too standard. I agree with you on not minding an anime-ish look for English synths, but even the "anime style" has variety within its bubble. I mean, think of all the options character designers have! They could go with a Western-inspired style like this or this, a more mascot-ish figure thats not entirely human. (or even something non-anime, like the ele.os fanmade design). but instead they went, "nope, lets create yet another Floating Moe Goddess (C)!". And that's not even considering her supposed personality. Eleanor's voice and character description are complete opposites. The voice is very neutral, while also coming across as really lively and spunky (at least to me). It sounds like a voice that a tomboy or really energetic person would have. But noooo. "Eleanor is old fashioned and loves antiques". It's so contrary its almost funny.

To tell you the truth, I like because it's not what you expect from a voice like hers. Maybe it's personal; when I was younger, I was really soft-spoken and had long hair and a cute face, and my deep, loud singing voice was hardly what people expected from someone who looked and acted like that (or so I was told, quite a bit). But I honestly love the direction they took with Eleanor! I agree that her OG design is quite cluttered, and while I'm not as big a fan of her new design (it feels almost stark and cold to me), it's definitely an improvement, removing all the clutter.

 

[andantina]

To tell you the truth, I like because it's not what you expect from a voice like hers. Maybe it's personal; when I was younger, I was really soft-spoken and had long hair and a cute face, and my deep, loud singing voice was hardly what people expected from someone who looked and acted like that (or so I was told, quite a bit). But I honestly love the direction they took with Eleanor! I agree that her OG design is quite cluttered, and while I'm not as big a fan of her new design (it feels almost stark and cold to me), it's definitely an improvement, removing all the clutter.

that's a really interesting point of view! its cool how you can connect with her character on a personal level because of that; it probably really helps with her charm. and yeah, removing the clutter was a good step in the right direction. it just makes more sense to have simple designs as opposed to complicated ones, considering that vocal synth characters are going to be drawn by many, many people. plus each person would ideally be giving their own interpretation/fitting it to their artstyle. and a fun design to draw means a variety of people who want that synth in the first place.

speaking of a variety of people: since the vocaloid scene is slowly regaining popularity, i'd love to see a popularity boom with Eleanor Forte or any other English SynthV character. like a modern retelling almost. instead of Hatsune Miku on Niconico Douga in 2007, it could be Eleanor Forte on the creative corner of YouTube in 202?. that's mostly wishful thinking though... but who knows? anything could happen!

 

[Scarlet Illusion]

Vocaloid is actually the worst of the singing synths rn. Like Cevio and Neutrino have this cute engine noise plus realistic vocals (with even more realistic vocals coming with Cevio AI). UTAU offers so much flexibility in tuning (pitch bends and append banks come to mind) for free. Synth V honestly changed the game for me with its ability to draw pitchbends right on the notes.
Idk I'll always love the Vocaloid characters, but the engine.... yikes.

I truly agree with this, I haven’t gotten to use CeVio or Neutrino yet, but when I come back to Vocaloid after working with UTAU and SynthV it just really lacks some very important things—especially due to its irritating methods of pitchbending. It’s hard to get the pitchbends to do what you want and I really struggle with getting the vibrato to do anything at all. The fact that SynthV lets me draw exactly what I want is nice (even if I do have trouble getting it to cooperate sometimes).

On a similar note, I find Vocaloids incredibly difficult to mix. No amount of compression or EQ is ever enough to make them audible over the backing track without making their consonants god-awful loud. After going back to Vocaloid for the first time in ages, I find them such a struggle to work with. (Praying that my beloved Cryptonloids will no longer be hell to mix once all the bugs in Piapro NT finally get worked out.)

 

[Prism]

I still think vocaloid 4 is great I do wish that synth v allowed for midi export with pitch bends because I would totally tube in synth v and export it out. @Kazumimi have you tried a vocal rider you could also try something like melodyn

 

[DefiantKitsune]

I'm gonna be honest and admit that I think CeVIO in it's current state is heavily overrated; it has an awful pink noise and a terrible voicebank selection, and I don't think the vocals sound any more "realistic" than Vocaloid (they are more emotive, but less clear and real)

In addition, it seems like it tends to heavily struggle with mature female and/or male vocals (Minato is good but there were a few good male V2s as well), a shocking percentage of incredibly poor quality vocals overall (at least 4 banks are virtually unusable, and around 4 more are somewhat debatable or mediocre- which is pretty horrid when you only have 15 banks on the whole engine) , and honestly it's Talk function varies from "slightly worse than AITalk" to "turn this choppy crap off" depending on the voicebank.

(Note that CeVIO can sound incredible with good tuning, and I really want to try Minato and Kasumi, but Vocaloid delivers, by default, much better results with much less work imo- especially in V5.)

AI might be better overall, but the lacking voicebank selection- including missing the "Miku" of the software and the second most popular voice being Talk only- doesn't bode well for it's success, especially with SynthV2 having a much better JP launch around the same time.

(Edit: making it clearer that there is not in fact a Miku voicebank for CeVIO)

 

[___]

Aaaaaas someone who got frustrated with tuning vocaloid just few hours back...I gotta disagree. Worst editor to get the grasp of? Worst editor in terms of the learning curve? Yeah probably but it undeniably creates one of the best singing synth results on the market.

And maybe it's just me but when I hear Cevio's engine noise "cute" isn't necessarily the 1st thing that comes to my mind.

 

[Stardust]

I have to agree about CeVIO. I was rooting for it until we got news about AI. My biggest gripe with it was it's engine noise, something that I only dealt with listening to because of ONE. With her missing and no news of if it's engine noise is going away, or at least lessening, I've taken it off of my list of synths I want. I hope they're just being quiet and not actually giving up on many of their banks, but I'm not going to bet on it.
V5 can be hard to learn, like patuk said, but right now it's the best synth right now, and has the biggest library of options in terms of vocals*. I feel once Neutrino has developed more and it's editor becomes better, and SynthV gets more voicebanks, Vocaloid will have a lot more competition. I heard it had competition with CeVIO, but I don't know enough about that to know if it was big.

*Not counting Utau here since the topic is more about commercial synths

As for some synth opinions unrelated to the current topic;

Long

I know it's not finished yet, nor is it's first voicebank finished, but Piapro Studio NT and Miku NT have been really disappointing so far for me. I know they're trying to recreate her original tone, but to me it sounds nothing like her V2. She sounds like a V4 append. Add on the editor itself being hard to work with currently from what I've heard from users, I'm not feeling very hopeful about it. I thought it would be so wildly different, but it really just sounds like V4. Nothing wrong with V4 banks, but for a new synth and all the hype Crypton has been making about it, I suppose I expected something big.

I feel like it's inevitable that Neutrino will go commercial at some point. Which is a shame because it's nice to have such a good synth be free, but with the quality it puts out and the work they're putting into it, I wouldn't be surprised to see them start charging for voicebanks.

I wish Plogue had put more effort into Alter/Ego, or at least didn't abandon it. It had so much potential, but they treated it too much like a VST. That works with something like Chipspeech, where realism or different tuning styles aren't expected since most people who purchase it want extremely robotic voices that other synths couldn't produce. But with Alter/Ego, it feels like they re-skinned Chipspeech's UI completely. It doesn't help that it lost one of it's few voicebanks for reasons unknown, and anyone who still has it isn't allowed to use it. That part alone doesn't make me feel okay with using it in case Marie or Bones gets pulled and I didn't know they did and accidentally use them.
(If Crusher actually did reveal the reasons for Daisy Bell being pulled then please inform me. I'd hate to speak about something without knowing the whole story )

 

[___]

I know it's not finished yet, nor is it's first voicebank finished, but Piapro Studio NT and Miku NT have been really disappointing so far for me. I know they're trying to recreate her original tone, but to me it sounds nothing like her V2. She sounds like a V4 append. Add on the editor itself being hard to work with currently from what I've heard from users, I'm not feeling very hopeful about it. I thought it would be so wildly different, but it really just sounds like V4. Nothing wrong with V4 banks, but for a new synth and all the hype Crypton has been making about it, I suppose I expected something big.

Oh big fat agree on her not sounding like her V2 at all. I remember in the first days of it's beta release ( or whatever that was ) people were saying how much she sounds like her V2 and I was wondering if we were listening to the same vb whether you find her pleasing to listen to or not is another topic but in terms of her being a recreation of her V2...well I think they failed pretty hard so far, imo this is the most un-Miku sounding Miku out of all her vbs. I genuinely think that if they wanted her to sound ""as Miku-ish as possible"" putting her on different engine was number one mistake. To me it feels like Crypton was trying to fix something that wasn't broken in the 1st place.

And just a heads up, this doesn't reflect on like...my overall feelings on Miku NT. I'm persoanlly excited to see CFM experimenting with new stuff even if I happen to not like the end results so far and I have fingers crossed for them but I'm just stating that for the goal they have in mind, I don't think their actions make much sense.