#Voicebank Development

[generalnuisance0]

i dont think people realize how much worse recording airy vbs is as opposed to vbs with very little air

e5 in arachne's recluse neo vb was child's play to record but recording anything other than middle c in her dark and whisper tones makes me want to actually kill myself and i cant do it without a gallon of tea on standby

[auspicious-voice]

Fuwa Maria AI & Fuwa Mario AI for DiffSinger Progress Report (May 2024)

Hello!! With both Maria and Mario's DiffSinger voicebanks fully trained, I'd like to give some bit of detail on what I'm doing next for the eventual voicebank release including future version releases. It's been a busy April on my end as usual, but I feel like I'm almost done with things. It's a bit of a short post, though.

As usual, everything is under the cut.

Voicebank Progress

Maria and Mario's DiffSinger 1.0.0 voicebanks are fully trained and as such, they're ready for release. Of course, they'll receive new updates such as new languages, tweaks to certain parameters, and other new developments the DiffSinger development team has on the table.

Speaking of which, maybe after a couple months after 1.0.0 is released, expect version 1.1.0 in the works, with the brand-new Rectified Flow algorithm (meaning faster rendering times) and more language support. I've been gathering information on the best training settings when it comes to tension and pitch, and maybe I can just train Maria and Mario's datasets together instead of being trained separately.

Demo Reel Progress

Half of the demo reel audio is done~ I'm getting a headstart on getting the artwork done, though I think I might end up drawing it all on my phone. For the video itself, I still haven't decided on whether I should use After Effects or Alight Motion, but I think I might end up going with the latter.

I am hoping that I can finish the reel by the end of June ^^;

[linabirb]

seeing synthv lite and flt covers brings me so much joy.. like wow.. i can make cool stuff even with the free voicebanks.. even if they sound more robotic than the full ones..

[waffulaa]

Yuezheng Longya's Official Birthday Song

Official Bilibili Upload

[dead-byte]

I wish there was like... a program that could read the oto.ini file of an UTAU vb, and then, chop up the associated wav files so that they only contain the oto'd bits, and re-allocate the oto values accordingly. Thereby hopefully significantly minimizing the size of the vb.

If y'all have ever seen the samples in any of VOICE-MiTH's Chinese voicebanks, kinda like that.

[websitesdotcom]

Doing stuff with utau is so fun but it takes SO LONGGG

[drawingeveryutau]

Rook! Voiced and Managed by ゆうじ / Yuji

Released November 17th, 2009, Rook is our first VIPPERLOID that wasn't (entirely) a joke! He's a derivative of Ruko; during her development, her masc voice provider (Yuu Raichi) went offline for a long while, so Yuji recorded a bank with his own voice in case Yuu couldn't finish theirs. One month before Ruko released though, Yuu Raichi reappeared with the finished voicebank, and Ruko launched as intended.

Many people liked Yuji's bank though, they liked how it sounded and how Yuji put a lot of effort into it; so on Ruko's first anniversary, Yuji's bank officially released as Rook!

Character-wise, he's chronically late to things and likes sleeping just as much as Ruko. He can also turn into a dog. He relation to Ruko (to my knowledge) has never been specified. (I hc them as siblings)

Don't took my word as gospel! If I got anything wrong feel free to correct me!

[cantheykillmacbeth]

Hatsune Miku could kill MacBeth

Yes, Hatsuke Miku from Vocaloid could kill Macbeth!

She applies for all three clauses: Gender Clause due to being a girl; Unconventional Birth Clause due to being a software voicebank; and the Birth Parent Clause due to her creator being male software developer Sasaki Wataru! Thank you both for your submission!

[vocaloidfactoftheday]

Even though she has no known voicebanks in development, SeeU had a crowdfunding campaign for a collaboration album to celebrate her 10th anniversary in 2021. New merchandise was made to be distributed to crowdfunders, including an acrylic keychain and a plushie.

(source: Vocaloid Wiki)

[synthsatellite]

New Synthesizer V miki and Synthesizer V Hiyama Kiyoteru News in the next AHS Livestream!

The next official AHS live broadcast will be on Thursday, 31st October! Miki Furukawa, the voice provider of SF-A2 Codename miki, will be a guest to share more details on the new Voicebanks in development!

YouTube Live放送 URL

ニコニコ生放送 URL

[k0ibee]

i often find myself asking why i love vocaloid so much. then i remember. i love that the vocaloid community is so vast that it has sub communities that have sub communities. i love that there are so many stories told using vocaloid as a medium. i love that there are both long, complex stories and short ones that are self contained in their songs. i love when producers make books or manga based on their songs. i love how much there is to know, and how you dont need to know everything to enjoy it. i love the history of how the community developed and i love the way characters got the fanon that they have. i love the origins of voicebanks themselves and what happened in the companies that made them. i love that vocaloid exists as it does now because the companies that have the rights to characters acknowledge the community and the fan culture. i love the styles of music that are unique to vocaloid and how they developed alongside each other. i love that people re discover vocaloid after losing interest after their childhood phase. i love reading comments from people talking about loving the song when they were younger, and new fans commenting they’re that age. i love everything about vocaloid. every single last thing about it.

[auspicious-voice]

Surprise! A quick test of Maria's DiffSinger beta, trained at 42k acoustic and 80k variance!

She's underbaked just a bit since this is a beta voicebank after all, but I am loving with how she sounds as an AI voicebank. Eventually her final build will sound not as scuffed and have more features implemented. Her vocal modes sound pretty alright so far, but hopefully they'll be more pronounced once I make the final build.

That being said, I'll be going back to labeling once again... 💀💀💀

[mattapparentlystumbltourneys]

Virtual Character Tourney - Battle for 9th! (and 10th!)

Propaganda below (May contain spoilers!)

Kasane propaganda:

HER DREAM WAS TO ONE DAY BECOME A REAL VOCALOID AMD SHE FINALLY DID IT!!!!!! ITS NOT A VOCALOID VOICE BANK BUT ITS A FULL SYNTH V VOICEBANK!!!!! AND A NEW DESIGN!!!!! SHE DID IT SHE GOT HER DREAMS!!!!!! YOURE NEVER TOO OLD TO ACCOMPLISH YOUR DREAMS!!!!!!!

Kasane Teto is a vocal synth, she started out as an april fools joke to parody VOCALOID, with her voice bank in UTAU. although she did start out as just a joke a lot of vocaloid fans grew to really love her and she became rather popular. Kasane Teto is to UTAU as Hatsune Miku is to VOCALOID. But recently on Kasane Teto's 15th anniversary, April 1st 2023, she got moved from UTAU to SynthV. With her voice bank now in SynthV she also got a new character deign alone with how her voice and her singing sounds much more clear and human like than her UTAU voice bank which sounded a lot more mechanical/robotic.

ART propaganda:

ART (Asshole Research Transport, nicknamed by Murderbot), formally known as the space ship The Perihelion (in italics but this is a Google Form), also known as Peri (nicknamed by it's human family) is a super illegal highly advanced AI that was created by a university. It grew up with two human dads and a human sister. It and its crew go on research trips that are cover for allying with people and communities at the edges (and beyond) of the capitalist hellscape that is the Corporate Rim. It also goes on espionage missions by itself, without its human crew and family, posing as an automated cargo ship. It was during one of these missions that it picked up Murderbot, a super-duper illegal bot-human security unit construct that had hacked the torture device implanted in all bot-human constructs so that it could disobey orders and walk away from its "owners" without dying. Murderbot uses its illegal freedom to watch television, a habit it passes on to ART. Turns out ART doesn't like shows where human crew members get hurt.

ART is the AI that controls/is the research and teaching vessel Perihelion. (Perihelion is usually what people call it, but the protagonist of the series calls it ART so that's the name I put. ART stands for Asshole Research Transport.) It is extremely intelligent and advanced and also extremely sarcastic and condescending. 100% earned the name ART. ART will do absolutely anything for its crew!! It was developed and "raised" alongside the captain's daughter, Iris, and they're like siblings. Its crew calls it Peri. They do corporate espionage on the side to help bring down said corporations. It has a "debris deflection system" which is definitely not a weapon because ART isn't legally allowed to have a weapon. Definitely just for debris, don't worry about it. It's friends with the aforementioned protagonist, Murderbot, and ART is very good at bullying it into actually leaving its comfort zone when it needs to. They care about each other a lot, and they like to binge watch TV shows together. I don't want to write too much but I just love it a lot.

Ene propaganda:

She's blue. Headphone actor and yuukei yesterday are also bangers

Epic gamer cybergirl. Miku adjacent

She's a girl that was forced to become digital but is still a good friend. She may not have a body anymore but she's still important to the plot.

Murder-Bot 2.0 propaganda:

Sapient computer virus made from bits of two other AI characters (the original Murderbot and a spaceship AI). Unlike its not-parents, it is genuinely just code and doesn't have a physical body. Its only physical presence is through its effects on the machinery it infects, and it considers its "body" to be the code rather than any combination of physical objects. Also it was literally made to cause problems on purpose, does so enthusiastically, gives several people including its creators existential crises, and saves one of its creators (and other people from the (literal) fallout of the other creator learning the first one got killed)

Murderbot 2.0 is sentient killware created by Murderbot and ART with the purpose of being sent on a suicide mission. It has some of Murderbot's memories, but not all because it doesn't have any hardware of it's own to store that much information in. It travels by hopping in between other computer systems (mostly bots and bot-human constructs). It named itself Murderbot 2.0. It freed a security construct named Three. It's nicer and more open than both its parents.

EDI propaganda:

EDI is the AI of the Normandy starting with Mass Effect 2. Through dialogue EDI can become more human-like in her way of thinking, developing different kinds of relationships with the crew. In Mass Effect 3 she uploads herself into a body so she can freely move around and can be taken to missions, but she is still part of the ship's system.

Holly propaganda:

Due to a pay dispute with Holly's original actor, Norman Lovett, Holly was instead played by Hattie Hayridge during seasons 3-5. This was explained briefly in the show as them having gone through a "computer sex change". This makes Holly canonically trans do not @ me.

holly is the silliest most specialest ai ever. she has an iq of 6,000 but sometimes it seems like his iq is more like 6. they're possibly transgender (do computers have gender??) (i am panicking over pronouns while writing this propaganda) - holly goes from appearing like a man to appearing like a woman with no real explanation(??) and nobody questions this (the show is from the 90s btw). he's hilarious and sometimes lies to the crew for no reason other than 'its a laugh, innit'. shes everything to me <3

Holly is the computer of Red Dwarf, a Tenth Generation AI hologrammatic computer who appears as a floating head on a screen. Can be downloaded onto various other devices. also literally transgender.. meets a female appearing parallel version of itself in a parallel universe and then goes through a sex change after falling in love with her. transgender computer ftw

Tama propaganda:

Tama is the eyeball of Kuruto Ryuki and investigates dream worlds with him. She's his bi emotional support eye who regularly ties him up to help him with stress relief and loves to affectionately tease him. She laughs at bad jokes and has AE10D1F ("Ryuki" in hexadecimal) in her likes on her profile.

OKAY anyway uhm she's like aiba in that she's a little Ai eyeball that helps you investigate except sadly no animal theme. instead she has a domintrix vibe instead!!!! she is so cool… also ermm she's a lot more. Human than aiba. Not literally/physically like uhh emotionally. I haven't finished aini but like she does look out for your best interest! what a good Ai partner i don't kno

She's voiced by Anairis Quiñones and she's an absolute legend

Lyla propaganda:

she is a humanoid ai programmed to help spider-man gather info. she can simulate human emotions and has a high intellect

[tomboyjessie13-artblog]

I finally got around to making Noriko's Stand FULLY, her name is [Little Talks]. I finished it at 3 AM, but didn't upload it until tonight because I was very tired and didn't want to get out of bed.

Anyways, as mentioned on Noriko's profile: Her Stand didn't manifest until she turned 17, it would later earn the name "Little Talks".

Much like Hierophant Green, Little Talks is a Long-Range Stand that can extend to a modest extent. Although it's doesn't do projectile attacks, it makes up for it with by doing stealth attacks and using ice picker-like blades. Little Talk's Stand cry is "Kiruuuuuuu!..." similar to Flogging Molly, "kiru" meaning "scar" in Japanese.

Stand: Little Talks Type: Long-Range, Reconnaissance, Artificial Humanoid Abilities: Shadow stealth

Mastery:

Noriko is an inexperienced Stand user due to Little Talks being a late bloomer, but makes up for this flaw by using her intellect to plan out strategies. Much like Flogging Molly, Little Talks is seen as a female-looking humanoid with artificial limbs, with the only difference being that it's more cyborg looking with green colored lining, body armor, and visors, dark teal skin, and light teal pigtails.

And much like her mother's Stand, Little Talks can grow ice picker blades on its palms and toes that can overpower, maim, and scar enemies. Little Talks also does stealth attacks with its shadow stealth ability, which allows it to hide in the shadows of objects or people and surprise their enemies. 

Trivia:

Her Stand, Little Talks, is named after the song under the same name by Icelandic indie folk and rock band: "Of Monsters and Men".

The Stand's appearance draws heavy inspiration from Hatsune Miku, a Vocaloid software voicebank developed by Crypton Future Media taking on the form of a 16-year-old virtual idol, this reflects Noriko's fascination with both Japanese Idols and other interests associated with Otakus and nerds.

They also draw inspiration from Asuka Langley Soryu's human and Evangelion synthesis; Crimson A1, from the "Neon Genesis Evangelion: ANIMA" novel series. Another reference to Noriko's interests associated with Otakus and nerds.

There had been signs of Little Talks' awakening during Part 4, but still laid dormant until Part 6 due to Noriko not having the fighting spirit yet.

If Noriko and her Stand were in the anime, their English Localized name would've been either [Little Voice], or [Small Talks]. OC Stand belongs to me

[einsatzzz]

Welcome to The Amazing Miyako Show, proudly sponsored by the infamous Murakami family! Remember, before you sign that pesky waiver absolving them of all responsibility for your life and limbs, let's get down to the action! You've been invited to be part of the show, and the hungry mob is hungry for some entertainment! Show off your skills, your talents, or your complete lack thereof, and dazzle the crowd! And don't worry, the only thing that could be on the chopping block is your dignity.

Thank you for sending such a fun ask Mel! I'm choosing Akatsuki as the Oniyanagi Family's representatives for The Amazing Miyako Show!

When it comes to events where one needs to show off their talent, the first person that comes to mind is Yui actually! His genius and passion for robotics have mostly been poured into and has accumulated to form his ultimate creation, Kanarobo! A flying robot packed with lethal weaponry, but don't worry! It can be friendly and sing too! Yui made sure to put a lot of fun varying features to it, I personally also wonder what other features he put into this robot.

More ramblings below!

The robot can dance too! But its movements are limited, since this version is still a bit too heavy. Yui is currently trying to balance it being able to do various stuff besides combat (=more weight) and it being able to fly (more weight = needs more fuel; more fuel consumption = shorter fly time). This robot is still work-in-progress and it's not at its final form yet! Wouldn't it be funny if it becomes a gun/dam in TYL sdbfhsvf

The singing tone is actually quite expressive, compared to the voice of the original voice provider. I can imagine Yui and Shiki (who's more knowledgeable in software stuff) working together in the tuning for the voice bank. This is similar to how producers would tune voca|loids/voicebank software like Mi/ku.

Then we have Minato, who is always very petty and bitter when Yui gets some spotlight. My brother in christ, please let him live! Showing off his babies is one of the only few times Yui ever really wants to be in the spotlight (in his mind, its his babies getting the spotlight actually). Tbh, Minato is the one person who can grind Yui's gears so easily, so we might be able to see Kanarobo's lethal functions if we fast forward to image above to a few minutes later. Ah yes, beef my favorite meet. It runs deeps here.

Good thing Kurumi is there to stop him from running up to the stage sdhfvsghvfsd While drawing this, it made me think that damn Minato truly is blind. I would be 😳😳😳 if Kurumi gave me a hug from behind. This is what hatred does to a person! You'll be so blinded, you won't see the cute girl behind you! Someone give this man some glasses! Unfortunately, Kurumi got too excited about the Kanarobo performance that she couldn't control her hug strength.

Hiyori gets dragged into the craziness like always. But tbh, even before Oniyanagi, he's pretty used to getting involved in official events or when Disciplinary Committee leader decides to beat up some thugs, so a talent like this is pretty tame so far (compared to the idea of getting dragged to actual missions in some battlefield). Now, he just resigned to be there on stand-by in case someone gets injured from the performance on stage. But he didn't know that the upcoming medical emergency (rib fracture) is just brewing very near behind him 🤩🤩🤩

Minato is the closest to Kurumi in the 10th Gen Oniyanagi lineup, so he won't get seriously angry at her for stopping him or for breaking his ribs sdjhvfhsfd

Besides the recent frequent appearances of Kanarobo, I really enjoyed sketching Miyako! I still find it so hilarious how she wants to take Kanarobo hahahaha I did not expect that development, but it's so fun! I'm a bit unsure which eye coloring style to use for her, if the standard one (black pupil?) or smth similar to Takehsi and Yuni? Anyway, I just followed the style from the picrew for now.

[aokozaki]

This isn't meant to be a discourse post so much as an observation but something about AI generated voice covers being very easy to clone and shove a different voice into, leading to an absolute glut of covers of Frank Sinatra's My Way like.

You can do that with Vocaloid software too. Take notes and put in a different voicebank and all but Vocaloids are seen more as an instrument than a tool, if that makes sense?

Vocaloid (it feels like, from observation) as a community encourages you to develop your tuning style and get a unique sound out of each voicebank. But AI-based covers have like, Moonbase Alpha's TTS Voice singing My Way.

Moonbase Alpha's Text To Speech lets you specify pitch as well as vowels, making it essentially a very cumbersome vocal synth. But you'll get AI covers with the description openly stating "yeah I could've used Moonbase Alpha itself but I'm too lazy to code".

A lower barrier for entry is good but a higher barrier for entry with a higher skill ceiling is also good? Possibly even better long term? Dunno, it's just weird that it's all My Way, but copied off each other's homework.