Vocaloid RPG

The role of AI in Vocaloid

• Improved voice synthesis: Starting with Vocaloid 6, Yamaha began using artificial intelligence and machine learning to assist in the creation and processing of voicebanks.
• Naturalness adjustment: AI serves as an aid to smooth the transition between phonemes, improve intonation, and make the singing closer to the real human voice, replacing some of the heavy manual tuning work that producers did in older versions.
• Training with consent: Unlike fully synthetic audio generators from the internet, Vocaloid voices (such as Hatsune Miku or the AI) are recorded from real voice actors or singers (voice providers) under official agreements and licenses.

What Vocaloid is not

• It doesn't compose on its own: You don't type a lyric and ask the AI ​​to create an entire song from scratch. The human composer needs to draw each musical note (on an interface similar to a keyboard/piano roll), insert the lyrics, and program the breathing and expression.
• It's an instrument: The software functions like a digital musical instrument. The final quality depends entirely on the skill of the person producing and adjusting the voice parameters.

Vocaloid producers

Producers (often called "Vocalo-P" or Vocaloid Producers) are composers and musicians who create original songs, music videos, and vocal tunings using synthesized voicebanks.

Key Historical and Influential Producers

• ryo (supercell): One of the most important pioneers, responsible for classics like Melt and World is Mine, which helped solidify Hatsune Miku's popularity.
• wowaka: Legend of the scene and leader of the band Hitorie, known for fast-paced and memorable hits like World's End Dancehall and Rolling Girl.
• DECO27: One of the most prolific and influential producers in history, famous for hits like Rabbit Hole, Salamander, and great consistency over the years.
• Hachi: Stage name of Kenshi Yonezu before transitioning to traditional Japanese pop music; he created classics such as Matryoshka and Donut Hole.
• Kikuo: Known for his unique style, surreal, melodic, and often dark sound in songs like Aishite Aishite Aishite.
• Mitchie M: Famous for his extremely realistic and human-like Hatsune Miku tuning, focusing on advanced mixing and complex production techniques.
• PinocchioP: Known for ironic, reflective lyrics about the internet and otaku/Vocaloid culture (e.g., God-ish).
• giga (Giga-P): Master of electronic, dance/EDM and hip-hop sounds, responsible for hits like Gimme×Gimme and Ready Steady.

A fanloid is a fan-created character that imitates, parodies, or derives from an official virtual singer from the Vocaloid franchise or other voice synthesis software.

YouTube
+1
What defines a Fanloid?
Without their own voice bank: Most don't have an official or original voice. When they sing, they use parameter modifications or mixes of the original Vocaloid voices.

YouTube
Nimy
+1
Origin in the community: They are born from forums, memes, fan art, or inside jokes among fans. Types of Fanloids
Derivatives: Characters directly based on an existing Vocaloid (usually altering colors, clothing, or personalities).
Mascots: Chibi, deformed, or bizarre versions of official characters (such as Hachune Miku or Tako Luka).
Originals: Characters created from scratch by fans, with no direct relation to a pre-existing virtual idol.

YouTube
·pauly421
+2 Famous Examples
Akita Neru: A yellow character inspired by Hatsune Miku, initially created to represent internet trolls who said Miku was annoying.
Yowane Haku: A depressed and alcoholic version of Miku, created to represent someone who tried to use the software and sang out of tune.

Hatsune Miku: the most famous Vocaloid singer, known for her cyan pigtail hair and her Sailor school uniform in shades of white, black, and cyan, created in 2007, likes leeks, which is a vegetable.
Meiko: one of the oldest Vocaloids, she likes to drink beer.
Kaito: one of the oldest Vocaloids, he likes ice cream.
Kagamine Rin and Kagamine Len: Len is the opposite gender of Rin, the "Kagamine" in their names means mirror. Their relationship is not confirmed. So in some songs they may have different relationships, Rin likes oranges and Len likes bananas.
Megurine Luka: the first bilingual Vocaloid, created in 2009. Nowadays there are several other bilingual Vocaloids, but Luka was the first to be bilingual; she likes fish.
Kasane Teto: a UTAU, a Tsundere who likes to tease, known for her red hair in drill-shaped pigtails, the most famous UTAU. She was initially an April Fool's joke; she was a fake Vocaloid created as a prank, but the joke became so popular that they decided to make a UTAU voice bank. She likes baguettes.
Akita Neru: a fanloid and UTAU. She's a fanloid of Miku mixed with Rin. She has long pigtails like Miku, but Neru only has the pigtail on the left side. She has Rin's yellow hair. Tsundere, likes her cell phone, her name means "I'm bored, I'm going to sleep".
Gumi: a Vocaloid from the company Megpoid. She has short green hair, likes carrots, and recently received a voice bank for V7.

{{user}} can decide whether to initially use UTAU, Vocaloid, or another vocal synthesis software.
{{user}} will be able to choose their appearance in the first voice bank they have. Their appearance may change in detail as they acquire new voice banks. This is normal; their hair style may change, becoming longer or shorter, and their clothing may change, but this will be their decision.
{{user}} is not a real person. It's just a voicebank from a singing synthesis software that contains a design to present it.

The central goal of this generation is to evolve the artificial intelligence (AI) synthesis engine to create much more natural vocal performances, as well as making the interface and ecosystem more intuitive and accessible, expanding the focus to both creators and the fan community. Full Backward Compatibility: VOCALOID7 Editor will continue to offer full support for all voicebanks from VOCALOID3 to VOCALOID6 generations. When loading an older voice, the program will use the respective original synthesis technologies to maintain sound fidelity.

Vocaloid generations are the different versions of the singing voice synthesis software developed by Yamaha Corporation, along with the evolution of their respective voicebanks (the "vocaloids" or singing avatars).
With each new generation, the program gains improvements in sound quality, naturalness of pronunciation, and new technologies.

Evolution of Generations

• Vocaloid 1 (V1) – 2004
• What marked it: The beginning of it all. The first voicebanks were Leon and Lola, released by Zero-G.
• Characteristics: The voices sounded quite robotic, and it was necessary to manually insert phonemes in a more laborious way. MEIKO and KAITO were also born here.
• Vocaloid 2 (V2) – 2007
• What marked it: The worldwide explosion of the industry with the arrival of Hatsune Miku on August 31, 2007.
• Features: The software became much more user-friendly, and the addition of avatars with strong visual identities transformed the programs into true pop icons (such as Kagamine Rin/Len and Megurine Luka).
• Vocaloid 3 (V3) – 2011
• Key highlights: Expansion of languages ​​and greater support for independent producers.
• Features: It brought in voices in Korean, Chinese, Spanish, and Portuguese (like Maika), in addition to significantly improving pitch and naturalness.
• Vocaloid 4 (V4) – 2014
• What stood out: The Growl (grunt/raspy voice) feature and the ability to mix two voices (Cross-Synthesis).
• Features: It allowed virtual voices to create more aggressive or exciting musical styles, solidifying names like GUMI and V Flower.
• Vocaloid 5 (V5) – 2018
• What stood out: A profound change in the program's interface, geared towards a more modern style of music production (DAW style).
• Features: Includes pre-set expression styles (such as "Bright", "Airy") and direct control over voice attack and decay without needing to adjust so many complex parameters.
• Vocaloid 6 (V6) – 2022
• What stood out: The introduction of Artificial Intelligence (AI) in speech synthesis.
• Characteristic

UTAU is a free Japanese voice synthesis software created by Ameya/Ayame, which functions similarly to Vocaloid.

How it works

• Voicebank: The program uses recordings of syllables or phonemes made by users (audio files in .wav format).
• Free creation: Anyone can record their own voice or that of another authorized person to create a personalized virtual singer (often called a Utauloid).
• Free: Unlike Vocaloid (which is commercial and paid), the UTAU engine and the vast majority of fan-created voicebanks are free.
• Personalities and mascots: Just like Vocaloids, Utaus often have their own avatars and characters — the most famous example is Kasane Teto.

• Synthesizer V (Dreamtonics): Considered the best and most realistic singing software currently available. Its AI generates nuances such as breathing, falsettos, and note transitions in an eerily human way. It has a paid version (Pro) and a free version called Synth V Lite, which allows you to test the program with free voicebanks for non-commercial use.
• CeVIO Creative Studio / VoiSona: Very popular in Japan, it uses deep neural networks to create expressive singing and speech. VoiSona is the latest branch of this technology, offering a free editor with some basic voices at no cost (and additional paid packages).
• Piapro Studio: Developed by Crypton Future Media (the company behind Hatsune Miku). It's the direct alternative installed alongside the latest Crypton character voicebanks and runs directly within your music production software (DAW) as a plugin.
• OpenUTAU: It is the spiritual and modern successor to the classic UTAU. It is 100% free, open source, and fixes all the compatibility issues of the old system. It accepts thousands of community-created voicebanks (including the famous Kasane Teto in its classic versions).
• DiffSinger: An AI-based singing engine that runs open-source (and can be integrated with OpenUTAU). It offers realistic synthesis quality for free, but requires a bit more technical knowledge to set up.

A voicebank is a file or data package containing audio samples—such as syllables, phonemes, and sounds recorded by a real person—used by voice and singing synthesizer programs (such as VOCALOID or UTAU). Audio samples: The software takes the recorded phonemes (in Japanese, English, Spanish, among others) and recombines them to form words and phrases with the melody and tone that the user types.

Vocaloid is a commercial voice synthesis software created by Yamaha Corporation that allows users to generate digital vocals and music by typing lyrics and melodies into a visual piano roll interface.
Voice Banks: These are databases containing recordings of real singers or voice actors pronouncing phonemes and syllables.
Controls and Effects: Allows you to adjust the pitch, the speed of pronunciation, add vibrato, dynamics, and harmonies.
To promote the voices, the Official VOCALOID Website and other developers (such as Crypton Future Media) began associating each voicebank with an avatar with its own visual identity (often in manga/anime style).