No.1926
Its dead only to vocal synth companies. Theres a reason for the sudden UTAU renaissance recently, people love their robotic voices. You probably won't see normal voicebanks anytime soon but you will definitely see more people using UTAU.
No.1927
>>1926Also I actually have no idea why Crypton hasn't jumped on the AI craze. Miku V6 is in development yeah but it's surprising to me since they're the biggest voicebank company.
No.1952
>>1927Concerns of her sounding like Fujita or "un-Miku"-like was the main reason, I've heard through the grapevine that due to this they originally claimed they'd "never" make a Miku AI bank due to this but I guess it's possible now. At the same time, Vocaloid was probably chosen due to the fact it's the
shittiest most unrealistic sounding AI in the vsynth sphere right now, ahead of only Piapro NT itself. CFM probably originally assumed Vocaloid would've progressed past it's current state by now. I AM curious about how Miku V6 will sound in vx_beta though, it might sound nice.
No.1953
>>1952>most unrealistic sounding AI in the vsynth sphere right nowCeVIO is must worse in that regard but I think the reason people hate V6 more is because the vocals sound extremely flat and there's no actual way to tune them (Unless that has changed now. I did see they updated V6 but im not sure if that was one of the updates. When i pirated it there were no tuning options.)
No.1997
outside of utau yes they're dead
No.1998
>>1953There's direct pitch editing now
No.6465
>>1925No, Piapro Studio is a thing
No.6467
Dat necrobump
Honestly i don't think sample based voicebanks will ever die as long as utau exists
No.7986
Kagamine NT soon bayyybeeee
No.8786
Rin and Len NT are sample based. I guess Crypton really wants to keep them alive. That's good.
No.8818
>>8786Only a matter of time before they switch to AI. I think what they're afraid of is that the new voicebanks lose their appeal which is their robotic tone. But they did well by going with the most robotic AI engine right now (V6) for the new Miku voicebank.
No.8823
>>8818I'm not against AI voicebanks but it's sad that sample based ones are likely going to stop being made at some point. But maybe this is a good thing and producers will finally use those UTAUs which have been begging to be used for decades now.
No.8824
>>8823Hyper-robotic UTAU are in vogue right now. UTAU that are half robotic, half realistic should come back into fashion eventually.
No.8825
>>8824I've definitely been seeing a lot more Rei, Defoko and UTAU Teto (who isn't a robot but since she was the first UTAU I guess she does sound a bit robotic).
I wish people cared about the lesser know UTAUs though.
No.8826
>>8823A lot of them are. Haraguchi for example uses UTAU Teto. Rei blew up after the Hear Abnormal song and is now possibly the 2nd most popular UTAU. Defoko doesn't get much use outside those Koronba copycat songs but it's still more than other UTAU such as Ruko and Ritsu get.
No.8846
AI synths are the hot thing right now for most people but honestly I prefer a well-made concatenative synth (good UTAU, most get hidden anyways in the sea of them and no one digs beyond surface level popularity unfortunately). But like other anon said, hyper robotic UTAU are super in right now for some reason, I guess AI fatigue + charm? Personally I would like to see more solid quality unknown UTAU get into the spotlight.
No.8848
>>8846I especially want the non-Teto Vipperloids to have the spotlight for once. I know they're relatively popular compared to other UTAUs but they barely have any original songs. I recently saw a new song with a decent viewcount that used Teto and Ruko, felt like a breath of fresh air.
No.8857
>>8846>AI fatigue + charmOh definitely. I would trace Rei’s popularity back to the avogado6 songs about her character. She has a core novel appeal in the physical robot thing.
To your point I think a good concatenative synth has unique strengths above AI and CV UTAU. Good blend of strength/staccato/timbre.
No.8872
I know it's been suggested in the Leekaloid thread already but we really should make a "robot" synth, sampling something unusual like a printer or a vehicle.
No.9023
>>8786Kinda? NT’s synthesis method - resynthesis - is spectral synthesis in a similar fashion to V1. While there’s samples, it’s also not entirely sample based surprisingly. It’s fucky.
No.9025
>>9023Technically all voicebanks use samples but AI voicebanks are trained on samples of the voice provider singing while older voicebanks + most new crypton voicebanks just use short, one second clips of the voice provider singing a syllable.
But yeah, without counting UTAU, I think Crypton are the only ones still making robotic voicebanks. It's a little sad because part of the appeal of vocaloid to me is how unrealistic they sound. But I do appreciate some AI voicebanks like Teto, Gumi and Eleanor.
No.12198
Rei saved sampled voicebanks
No.12202
>>8846Seconding this, it's quite sad how the library of UTAU voicebanks is so massive yet only like 4 are ever used. For original songs at least, because most whotaus have at least one or two covers.
No.12214
>>12202some of the "popular" utaus were whotaus at some point
No.12884
A lot of producers who dont like the AI sound use UTAU
No.12919
They're not dead, it's just that no one's making them
No.12920
>>12919Worded this poorly but I'm saying that AI voicebanks are the cool new thing and they sell so they have no reason to make sample based voicebanks but there's always a possibility like with Rei V5
No.12932
>>12920i'd love it if they made more shit for v5. that thing has no voicebanks.
No.13502
>>12932V5 is the PS5 of vocalsynths except instead of being at the end of its life and having no games, its at the end of its life and has no voicebanks
No.23514
>>23512Winning what? The only thing people use UTAU for these days is for pumping out Rei/Teto songs because if they used any other voicebank they wouldn't get views
No.23523
>>23522man i love avogado
No.24598
>>23514And Defoko thanks to Koronba copycats
No.24711
>>23514There's also the occasional Momo song