Hatsune Miku V6 Developer Wataru Sasaki × Professional Vocal-Tuner Bibi Special Dialogue: The Secret to Rich Vocal Expression Lies in “Love for the singing characters” and “Oshikatsu”!?

Hatsune Miku V6 Developer Wataru Sasaki × Professional Vocal-Tuner Bibi Special Dialogue

■Wataru Sasaki Profile
　Sound Team Manager at Crypton Future Media, INC. Chief Producer for the “Hatsune Miku” and singing voice synthesis-related projects. Born in Sapporo in 1979. Joined Crypton Future Media in 2005, and in 2007, led the planning and development of the singing voice synthesis software “Hatsune Miku”, which became a massive hit. Since then, he has managed the production and direction of the company’s singing voice synthesis products and projects. Additionally, he participates in the OngaACCEL project, which works on researching new music technologies, and writes liner notes for electronic music artists such as Aphex Twin, Arca, and Squarepusher.

■Bibi Profile
　A creator specializing in the vocal-tuning of singing voice synthesis software. Active since 2012, Bibi has handled vocal-tuning for songs produced by numerous Vocaloid producers such as Jin, *Luna, and Marasy. Furthermore, Bibi has tuned several virtual singer songs featured in the smartphone game “HATSUNE MIKU: COLORFUL STAGE!”.

I was deeply shocked, thinking, “This isn’t human, yet it’s singing!”

Sasaki　When did you first become aware of “Hatsune Miku,” and what was the trigger?

Bibi　I was originally interested in so-called “subcultures” like manga and anime, and through that connection, a friend introduced me to it, saying, “This is what’s popular right now!” The first song I heard back then was “Miku Miku ni Shite Ageru♪.” I remember being deeply shocked, thinking, “This isn’t human, yet it’s singing!” From there, I started listening to various Vocaloid songs. Seeing “utaites” (singers) and people create illustrations and videos based on those songs, I was quickly drawn into the culture centered around Vocaloid.

Sasaki　Besides being a culture that started on Niconico Douga, there was a background of internet culture like the anonymous textboard “2channel” at the time. I remember a sense of celebration, like, “Something interesting like this has appeared on the internet! Yay!” As a symbol of joy, people were waving leeks. You went through that phase too, didn’t you?

Bibi　Yes. Until then, the internet was just something I consumed, but since I originally drew illustrations, it stimulated a feeling of “I want to try making something too!”

Sasaki　It was a situation where everyone was listening because they wanted Hatsune Miku. After that, there was the impact of the song “Melt”, then a sense of family-likeness emerged with “Kagamine Rin/Len,” “Megurine Luka” could sing in English, and “Megpoid” felt a bit like a rival… Oh, and “Gackpoid” was shocking too. In those days, there was a thrill of not knowing what would happen or how everyone would “put their own spin on” the characters. When “Megpoid” came out, I was surprised by the charm of a singing voice completely different from Hatsune Miku. Which virtual singers or creators do you like, Bibi?

Bibi　My “Oshi” (favorite) is Kagamine Len (laughs). Because of that, I love Giga. Also, cosMo@Bousou-P, who can be called a pioneer of Vocaloid-like singing styles. The shock when I first heard “THE END OF HATSUNE MIKU” was incredible. I had never heard a song like that before, so I felt the immense possibility of Vocaloid. I still love it today.

Sasaki　I see. After all, you can’t forget the impact of something so edgy.

I tune with the desire to “see the moment the character shines”

Sasaki　Listening to the singing voices you’ve handled, I get the impression that the goals are clear and easy to understand, and that someone with a deep knowledge of singing and musical expression is doing the vocal-tuning. You mentioned “utaite” (singers) earlier, but what led you to become a “Vocal-Tuning Specialist”?

Bibi　I’ve always liked music and singing, but more than anything, I love the characters. Once I learned about the method of expression through covering songs with Vocaloid, my desire to “express my favorite songs with these characters!” became even stronger. Nowadays, I spend every day listening to human singing and thinking about “how to express this in the VOCALOID Editor.” I studied by looking at the project files released by Kannazuki-P until my eyes practically burned holes in them.

Sasaki　 I see. Your love for the characters—or rather, your desire to have the characters move dynamically—leads to the charm of the vocal expression you tune. I even feel a sense of obsession to realize the expressions you like.

Bibi　I think the desire to “see the moment the character shines” resulted in me specializing in vocal-tuning. If I liked music, there are other paths like composition, but I love voices, and I want to pursue how that character can shine the most. In other words, it’s “Oshikatsu” (supporting one’s favorite). Since I do vocal-tuning as a form of Oshikatsu, I think I’ll keep doing it even if I end up alone in the future (laughs).

Sasaki　That aspect of you might be what evokes sympathy from the listeners. The characters you like are characters everyone likes too. I feel like the nuances everyone wants to hear resonate and and are reflected on your thoughts. For example, in Marasy’s “Mugen no Ticket”, there’s a moment where I feel, this is the part where KAITO’s eyes snap open.” Marasy’s song is wonderful, of course, but I feel the dynamism where the character’s energy leaps through the voice.

Bibi　Thank you. However, my activity style is based on the premise that I am borrowing everything from others. That’s exactly why I want to cherish the songs and the characters. Perhaps that eventually connects to what everyone is seeking.

Sasaki　Your vocal-tuning feels lively even though it’s Vocaloid. Do you ever expand your image for vocal-tuning by singing yourself?

Bibi　I only sing as a hobby, but when I first started, I used to do vocal-tuning while singing. I would think about how to reproduce it by singing it myself, doing it with the feeling that we were singing together.

Sasaki　I might be hearing the feeling of “singing together” for the first time. Singing together with the character… that’s lovely. You really do love the characters, Bibi.

Bibi　Yes, I’ve come this far just on that (laughs).

Sasaki　On the other hand, are there points you focus on in vocal-tuning to bring out “Vocaloid-ness”?

Bibi　Not manipulating dynamics too boldly, I suppose. With humans, I think there’s quite a bit of fluctuation in vocal volume, but precisely because they aren’t human, there isn’t much of a difference. I think that has become part of the “Vocaloid-ness.”

Sasaki　You seem to like that “saturation of the voice” typical of Vocaloid, or that “overload feeling” where the volume difference at the transitions makes the Vocaloid’s voice sound louder.

Bibi　I love it. That’s why I always emphasize the attack. In fact, I like “Kagamine Rin/Len” act1 (the common name for the first version released) the best. To me, that is the image of Vocaloid.

Sasaki　For “Kagamine Rin/Len” act1, partly to differentiate from “Hatsune Miku”, we recorded the voices dynamically. We had Asami Shimoda give it her all until the very limit, which led to that sense of overload. It wasn’t an easy form to control, but it certainly had a dynamic quality typical of Vocaloid.

V6 gives the impression of evolving from the “Miku-chan” I felt in V2

Sasaki　You have a clear recognition of the V2 era “Hatsune Miku,” and you tried the new V6 this time. Do you feel the differences between each version from V2 to V6?

Bibi　I feel they are quite distinctly different. V2 felt like “Miku-chan,” but with V4X she became “Miku-san,” and I had the impression she was gradually becoming more of an “older sister.” However, there is always a sense of security that “Hatsune Miku” will always be “Hatsune Miku” for me. As synthetic voices evolve and get closer to humans, I strongly feel that the original voice is being cherished.

Sasaki　Did V6 feel like it was on an extension of “Miku-san” this time?

Bibi　The tone of V6 and NT feels more like an evolution from the V2 “Miku-chan.” It feels like the evolution was made by valuing the original voice quality of Hatsune Miku rather than just making the voice more human-like. I want to verify what happens if I re-tune songs I’ve tuned before with V6. I feel that my own direction of expression will be established by seeing what kind of differences emerge.

Sasaki　”Hatsune Miku V6” was developed with the image of the “Hatsune Miku” that everyone envisions as a virtual singer. We prepared “Original” and “Soft” as voice banks. For “Original,” we analyzed and mixed in a tiny bit of “LIGHT” from the V2 expansion voice library pack “Hatsune Miku Append,” and for “Soft,” we put in a bit of “Sweet” components. We aimed for “Hatsune Miku” by balancing the quirks of VOCALOID6 with the brightness or calmness of the voice. “LIGHT” was a bit tricky to position  back then as it wasn’t fully differentiated from the V2 Original.

Bibi　I loved “LIGHT” and “VIVID” from “Hatsune Miku Append.” I’m so happy!

Sasaki　At that time, we challenged the development of “Hatsune Miku Append” while thinking, “If there were another Hatsune Miku, what would she be like?” There was feedback like “Ultimately, the original is the best,” but it’s deeply moving for me that this is living on in V6 after all this time. Perhaps the image of Miku’s voice as a character is something that can be pursued infinitely. How about the difference between “Original” and “Soft?”

Bibi　”Original” was cheerful and easy to add expression to. Previous versions gave a somewhat quiet impression when untuned, but brightness has been added to “Original,” making it very fun to have her sing. “Soft” also blends well with songs and is easy to use. Voices with many breath components tend to become noisy, but that didn’t happen.

Sasaki　Due to the nature of Vocaloid, the voice can sometimes become muffled or weak and get buried in the track. I’ve always felt sorry for the creators about that and want to make it easy to hear with a good balance. However, on the other hand, there are opinions that a voice that seems to fade away was also one of the charms of Vocaloid. What do you think about the balance between ease of use and the character’s individuality?

Bibi　Certainly, previous Vocaloids had imperfections, but as someone who tunes, I did it thinking that those were all part of the individuality. I think the imperfect points were definitely a charm. As synthetic voices evolve—not just Vocaloid—I feel that uncertain elements have decreased. Previously, unexpected things would happen, like “if I write this phonetic symbol this way, this kind of sound comes out,” which was exciting (laughs). But because it has evolved, the hurdle has certainly lowered, and I think it’s easier for first-timers to touch. However, it’s a bit lonely that the room for experimentation with phonetic symbols is disappearing.

Sasaki　I see. Using AI, it would be great if we could include a gimmick that brings out vowel or expressive outliers while maintaining clarity. It might be interesting if we could generate new vocalizations from nonsense phonetic symbols or forcefully connect consonants and vowels.

Bibi　Yes. Actually, I find “Megurine Luka V4X” to be very fun. It has so many types of voices; it’s like a toy box.

Sasaki　That was my intention at the time. Thinking about ease of use, I remember it might have been a bit overdone… (laughs).

Bibi　 I somewhat felt that might be the case. But as a user, I was excited, thinking, “If there are this many choices, I’ll be exploring forever!”

Sasaki　Hearing you say that now makes me feel that life is worth living (laughs). I also hope for a world where characters sing richly, including the “weird” stuff. I’m sure you do vocal-tuning partly because you want to hear various kinds of songs.

Bibi　That’s right. I want to hear many kinds of songs. V6 has become very easy to use and intuitive, so I’d be happy if, utilizing AI and other tools, more variety of expression becomes possible in the future.

『Hatsune Miku V6』 Product Overview

『Hatsune Miku V6』 is virtual singer software compatible with VOCALOID6. It supports not only Japanese but also English, and support for Chinese is planned for a free update in 2027. Two formats are available: a starter pack (24,200 yen) that includes the VOCALOID6 Editor, and a voice bank (10,780 yen) for users who already own the editor. Both include Cubase LE DAW software. Compatible with Mac/Windows and operates in AU/VST3 formats.

Utilizing VOCALOID:AI, the singing voice synthesis engine has been optimized. Just by entering lyrics and melody, natural breathing can be added automatically, and fluent expression across multiple languages is possible. Additionally, since it has learned pitch transitions, volume changes, timing, and accents through machine learning, it excels in vocal expressions that feel warm and naturally fluctuating, close to a human voice. A new TAKE function has also been implemented to generate multiple vocals with different nuances from a single phrase, allowing for easy creation of harmony parts.

While natural singing can now be expressed even with default settings, fine adjustment of parameters remains possible as before. Users can adjust the intensity of the voice’s tension with the “Expression parameter” and intuitively edit delicate singing expressions such as “shakuri” (Bend Up), vibrato, and “tame” (deliberate delay) while looking at the pitch and volume curves on the screen.

■『Hatsune Miku V6』

『Hatsune Miku V6』