#clue talks vsynths | Explore Tumblr posts and blogs

mayordeas-clone · 2 months ago

Text

saw a take that peeved me in the vocaloid feed on bluesky but i don’t wanna be confrontational there + my thoughts are too powerful for a 300 character per post thread so im making it tumblr’s problem

due to my intense vsynth brainworms this is going to sound incredibly pedantic and annoying which is why i am posting it here okay here we go

the post was sort of a ‘callout’ of a creator who uses synthv for covers of fnf songs or whatever. the general body was like “they write their own lyrics but they make the vocals using ai with a program called synthesizer v” (the implication being, “this person is an ai-using fraud, don’t support them”)

if ur in the community or like have a little bit of an idea of how ai is associated with synthv specifically this is a nonsense uninformed take (the ‘ai’ training is used for autotuning/cross language synthesis/some other shit idk but the point is it’s not the type of ai that like, writes the melody/lyrics for you and is voiced from random samples from unknowing providers but i’ll get back to this later)

ai is mentioned on synthv’s website (also vocaloid’s much to the detriment of news articles that report on hatsune miku) so it’s a common misunderstanding, whatever. it would be nice if people actually looked into the ways ai works in vsynth at least a little bit but in an age where artists are on the defense against the ai boogeyman i guess i sorta understand the knee jerk reaction to assume the ai in vsynth software like synthv is the Bad kind. annoying, but nothing new.

but i think what made this post rotate in my mind to the point of crystallizing into unnecessary rage is an add-on that was like “i just prefer my music to be sung by real people!” followed by “by the way, vocaloid isn’t ai, so no you’re not bad if you like hatsune miku music”

ohhh man this actually did piss me off a little cuz wow way to really shout to the heavens that you have no damn clue what you’re talking about

firstly, i COMPLETELYYYY understand that the premise and general sound of vocal synth music is not for everyone. the computerized words, the unrealism, or even the fact that it’s not an actual person singing, like whatever that’s fine i get it. but like you can’t go around and be like “by the way vocaloid [aka hatsune miku] is chill don’t worryyyy” bc that line of thinking from my understanding doesn’t really make sense? miku isn’t a real person either 😭😭 her latest release even uses ai! (at least i think it does cuz vocaloid 6 has ai support but idk i guess that doesn’t matter) does she just get a free pass cuz she’s an internet darling?

miku is a piece of software where you write the midi for her vocals and input the lyrics so a slurry of samples recorded by saki fujita can be played back in an in-character singing voice. synthv is a piece of software with similar character voice databases where you do the same thing, write a midi and input lyrics so a database’s voice samples put it all together to sing back. the inclusion of ai in this process is to streamline the workflow for the user to automatically work out the pitch transitions between notes to save on time manually tuning, creating a more realistic playback. this ai, perhaps merely the inclusion of these two letters in the proximity of synthv’s marketing, is what makes synthv both Significantly Different and Inferior to miku.

the lack of ai with vocaloid (as far as op knew, cuz need i remind you vocaloid as of its 6th installment uses similar ai to synthv lol) making it Better than synthv sent the implication, TO ME, that op thinks that the ai used in synthv is the type that like, takes samples from other people without their permission to create the voices (then leading to the “i just want my songs sung by REAL people” aka not sung by dubiously gathered samples). i recognize this means of creating ai voices to impersonate other people and make them say things they should not without their knowledge or permission is CERTAINLY bad but like that’s not how any of this works if we’re talking synthv (you could argue this applies to those utau banks that make voicebanks for like sonic the hedgehog and shit but those usually sound like ass and are about as convincing as a sentence mixed youtube poop). maybe that’s not what they meant (i don’t wanna dig up the post cuz i don’t wanna be bothered and it’s probably gonna piss me off to the point of starting something which i don’t want) but based on my understanding of their point this is like. wholly untrue. in fact synthv voicebanks (particularly those from eclipsed sounds whose terms of service are free to read on their website) have special terms specifically designed to protect the voice providers of their databases BECAUSE of the extra realism in their voices (terms such as, “you can’t credit any voice work done with Solaria to Emma Rowley [her voice provider] or anyone else besides the software itself” in the case Solaria is used to say something objectionable and cast it on Rowley, that would be a breach of the terms and might even get your license revoked). shit like that is likely what people who are afraid of ai voices expect to happen but there were contracts signed to protect and inform the voice providers of what their vocal likeness is being used for + compensation paid to them for their work providing vocal samples SPECIFICALLY for this product. and ofc miku is the same. ITS THE SAME. BUT MIKU IS A FUNNY INTERNET MEME SO SHES FINE. GRRRRR!!!!!!

#mayor talk #IM SORRRYYY i’m so insufferable with this shit #what a 4 year consistent attachment to a niche ass music subculture does to a mf #don’t seek out op if you find the post etc etc i have fought in the internet argument trenches and im NOT doing it again #i might find the post to block them tho. nothing personal kid but you have awakened my wrath #anyway i mentioned in the post but i GET the defensiveness against the mention of ai in association with art and music #vsynth on the surface sounds like it’s perpetuating this ai rise. ‘oh i didn’t have a real person sing it i used software to do the singing #for me so i could save on hiring a vocalist’ [vocaloid being an alternative to hiring a vocalist per hour is a legitimate selling point]#i can see people who sing or do music or appreciate either bristling at that. ‘a computer sung this? hm…..’#i try to say it out loud and i feel like i’d piss people off! but as we know the difference in workload + sound makes it less black and whit #it’s just. frustrating to see people immediately assume the worst whether they’re talking about synthv or miku. :[#on the flip side it’s also annoying seeing people argue for the difference between generated ai vocals and vocaloid by saying #‘vocaloid is not only different but BETTER and MORE VALUABLE bc it’s hard to use/you have to do it yourself’#yes gen ai and vocaloid demand different workloads and learning curves but uh nooo please don’t say that #please don’t imply that art only has value or goodness depending on how much an artist suffered to make it noooooo

3 notes · View notes