Piapro Studio DTM Station interview: Hatsune Miku NT and VOCALOID 6

[MagicalMiku]

DTM Station interview to Crypton Future Media about Hatsune Miku NT [Part 1]
(translated from the original article in japanese, here):
https://www.dtmstation.com/archives/74903.html

Piapro Studio NT2 and VOCALOID 6 data compatibility have been realized on November 19th. The NT project accelerates with the release of Kagamine Rin and Len NT.
"Hatsune Miku NT", released in November 2020, underwent a major update in March 2025 to a version equipped with a new in-house developed engine ("Hatsune Miku NT Ver. 2"). Approximately six months later, on October 15th, " Kagamine Rin & Len NT " was released, greatly accelerating the development speed of the NT project.

Then, at 15:00 on November 19th, it was announced that data compatibility between Piapro Studio NT2, the vocal editor included with NT products, and VOCALOID 6 will be realized. Piapro Studio can now read VOCALOID 6 .vpr files, and conversely, VOCALOID 6 can read Piapro Studio .ppsf files. This will enable data exchange between the Piapro Studio NT series and the VOCALOID 6 series, providing creators with a more flexible production environment. We spoke with Koizumi Takamichi, Sasaki Wataru, and Okamoto Kengo of Crypton Future Media, Inc., to learn more about this compatibility and Kagamine Rin & Len NT.

Following Hatsune Miku NT, Kagamine Rin & Len NT was also released.

table of contents:
-"Kagamine Rin & Len NT" has arrived with increased speed;
-Why didn't Kagamine Rin and Len appear in the old NT?
-Equipped with the same new engine as "Hatsune Miku NT (Ver.2)";
-What are the benefits of in-house development?
-Optimized exclusively for "Kagamine Rin & Len";
-New vocal morphing function added;
-The evolution of "Piapro Studio NT2";
-On November 19th, Piapro Studio and VOCALOID 6 became compatible;
-VOCALOID Hatsune Miku development status.

"Kagamine Rin & Len NT" has arrived with increased speed
--Kagamine Rin & Len NT was released on October 15th. First, please tell us about the events leading up to this release.
Koizumi: Kagamine Rin & Len NT was released on October 15th, but it was announced on August 31st as part of a stage event called "Hatsune Miku Magical Mirai 2025," and pre-orders began on that day. The NT series began with the development of Hatsune Miku NT, which was released in November 2020. Kagamine Rin & Len NT was released about six months after the major update to Hatsune Miku NT in March of this year. Considering that it took nearly five years for Hatsune Miku NT to be updated to a new engine, this is a significant speedup.

--What's behind this speedup?
Koizumi: A big part of it is the strengthening of our development structure. With producer Sasaki at the center and Okamoto and I involved in direction, we've been able to take on management aspects such as the overall progress schedule, mid-term plans, task management for on-site members, and progress management of the development situation. I think we'll be able to continue releasing products at a similar pace in the future.
Sasaki: Actually, the foundation of the new NT engine was already in place when we released "Hatsune Miku NT" in November 2020, so we were able to gradually speed up from there. The development of "Hatsune Miku NT" created what could be considered the foundation of the NT series, so we were able to release "Kagamine Rin & Len NT" with a little more leeway, adding features as we went.

From left: Koizumi Takamichi, Sasaki Wataru, Okamoto Kengo

Why didn't Kagamine Rin and Len appear in the old NT?
--Kagamine Rin and Len with the old NT engine were never released, were they? Why didn't you release them then?
Okamoto: Simply put, it was because we hadn't decided yet on a policy that would allow us to turn NT into a series. At the time, we were still carefully considering the design aspects of how to release the latest VOCALOID versions of each virtual singer and NT products.
Koizumi: VOCALOID naturally has its merits. However, in Yamaha's recent development plan, VOCALOID itself has evolved into an engine that, to put it simply, aims for a more human-like or realistic feel, while Hatsune Miku is different and unique with her own style. So, we first commercialized Hatsune Miku with the old NT engine, while exploring the direction of each product.

--I see, so VOCALOID is moving in the direction of pursuing human-likeness.
Koizumi: Yes. However, Hatsune Miku is not human, so as the VOCALOID engine evolves in that direction, we need to be creative in order to maintain the "Hatsune Miku-ness" of the character.
Sasaki: I'll be able to talk more about the details of the VOCALOID engine's innovations when "Hatsune Miku V6" is released, but unlike typical AI voicebanks, we're carefully deciding how much to maintain mechanical voice movement and how much to update to more organic patterns, processing elements not found in human voices, and clarifying the voice to make it more Hatsune Miku-like. While the NT series maintains a style of classical vocal synthesis and synthesizer-like aspects, VOCALOID6 is moving in a direction that includes English singing.
Koizumi: With NT, we're developing each virtual singer's synthesizer and instrumental voice while firmly preserving their traditional sound—that is, the "characteristics" of the character—and ensuring that our engine can guarantee that.

--So, with the path to releasing VOCALOID 6 now in sight, the direction of NT has also become clear.
Koizumi: That's right. So, even though we had been in the process of careful consideration and hadn't been able to release "Kagamine Rin & Len NT" until now, we were able to decide to release it after sorting out the situation and our vision for the future. That being said, we will continue to pursue the "identity" of NT, so it's not like it's over once the product is released.

--What exactly do you mean by "it's not over once it's released"?
Okamoto: One piece of evidence is that on November 5th, one month after the release on October 15th, we updated the voice library and pitch shift for "Kagamine Rin & Len NT." This corrects the low-frequency range, particularly to make Kagamine Len easier to listen to. Rather than completely changing the singing voice itself, we're making appropriate improvements, and we'll continue making these adjustments.
Sasaki: In addition, we plan to strengthen the synthesizer-like aspects. We hope to meet user needs with VOCALOID6, while also being able to make important updates to NT.

 

[MagicalMiku]

DTM Station interview to Crypton Future Media about Hatsune Miku NT [Part 2]
(translated from the original article in japanese, here):
https://www.dtmstation.com/archives/74903.html

Equipped with the same new engine as "Hatsune Miku NT (Ver.2)"
――I understand that "Kagamine Rin & Len NT" uses the same engine as "Hatsune Miku NT (Ver. 2)." Can you tell us again about the features of this new engine?
Okamoto: The synthesis method of the new engine is very different from the old NT series engine. The old engine used something similar to AI synthesis. So, while it was convenient for editing and making changes, it had the problem of losing a lot of the Hatsune Miku-like quality.
With this new in-house developed engine, we wanted to create a sound that was close to the image of "VOCALOID Hatsune Miku" that is fundamentally present in the minds of our users, so we completely changed the synthesis method. We developed a new synthesis method that is closer to VOCALOID, not often used in other companies' synthesis engines. This has allowed us to create a new engine in the NT series that effortlessly reproduces the "VOCALOID-like voice direction" in a high-quality manner while staying close to the original sound.

--When you say something close to VOCALOID, you don't mean the current VOCALOID 6 engine, but the original VOCALOID.
Okamoto: It means something that is close to the lineage of the technology used between V2 and V4. It's not exactly the same, but it's just that it's in the same broad category.
Sasaki: The people in charge of the engine are continuing to work on it so that it can reproduce the strong, crisp feel of the voices that are characteristic of Hatsune Miku and Kagamine Rin and Len, and updates are expected to continue.

--It's a different type of engine from the old one, which is why it's easier to bring out the Hatsune Miku feel.
Okamoto: Yes, that's right.
Koizumi: There's also still room for improvement in the synthesis method itself, and there are requests within the company for ways in which they would like things to be done, so we switched to an in-house version that was easy to change and improve.
Sasaki: This is true for the Voice Color library, so we can update and change the concept as a whole, and we've also taken future development potential into consideration.

What are the benefits of in-house development?
--If that's the case, wouldn't it be fine to just leave it as is with the original VOCALOID? That may raise a question, but what are the differences between this and the original VOCALOID?
Okamoto: As you know, VOCALOID is a Yamaha product, so there are limits to the adjustments we can make. Also, it was difficult to see how the sound was changing internally, so we had some concerns about how difficult it was to incorporate the library adjustments we wanted.
So, we developed an engine in-house that was similar to the technology used between V2 and V4, as I mentioned earlier, and created an environment where we could easily try out new approaches. This has made it possible to integrate it with computer-generated parameters, such as Automatic Control, which is included in the new NT engine.

Automatic Control function of Piapro Studio NT2

Optimized exclusively for "Kagamine Rin & Len"
--When you were talking about "Hatsune Miku NT (Ver.2)," the previous engineers and database were geared towards Hatsune Miku, but this time it's Kagamine Rin and Len, so are there any other differences? In terms of differences from "Hatsune Miku NT (Ver.2)"
Okamoto: As you said, the new engine for "Hatsune Miku NT (Ver.2)" was optimized for Hatsune Miku, so we made various adjustments to achieve the "likeness" of Kagamine Rin and Kagamine Len.

--I see. In terms of functionality, can I understand that the basics are the same because they use the same engine?
Okamoto: That's right. We've made some adjustments, but the basic functionality of the engine is the same as "Hatsune Miku NT (Ver.2)".
Sasaki: We've made updates such as optimizing the pitch shift algorithm so that we can pursue a more "boys-like" voice for Kagamine Len, while still maintaining the girlish voices of Hatsune Miku and Kagamine Rin. We'll continue to adjust to various voice qualities, and in the future, I hope we'll be able to challenge voice qualities that we weren't able to fully develop in the VOCALOID version.

 

[MagicalMiku]

DTM Station interview to Crypton Future Media about Hatsune Miku NT [Part 3]
(translated from the original article in japanese, here):
https://www.dtmstation.com/archives/74903.html

New vocal morphing function added
--With the release of "Kagamine Rin & Len NT," a new feature called vocal morphing has been added.
Okamoto: Yes. This is similar to the cross synthesis feature in VOCALOID, which allows you to mix different databases of the same singer, in other words, mix sound quality, and create intermediate expressions.

--What else? In terms of functionality...
Okamoto: Going into the details, the pitch line estimation in Automatic Control also uses a different model from that of Hatsune Miku, and is optimized for Kagamine Rin and Len.

The evolution of "Piapro Studio NT2"
--Could you also tell us about the evolution of "Piapro Studio NT2" itself? In the recent article about "Hatsune Miku NT (Ver. 2)," we focused on the detailed editor features, but what about this time?
Okamoto: As I mentioned earlier, the Automatic Control section not only controls pitch, but also vibrato and intonation. These parameters are automatically estimated using the built-in engine, and we've made adjustments to them specifically for "Kagamine Rin & Len NT." Rather than simply reusing the adjustments made for Hatsune Miku, we've created a new data set specifically for Kagamine Rin & Len and combined them accordingly.
Koizumi: I'd also like to talk about the major features added in the updates to Piapro Studio itself since March and April. First, there are improvements in speed. The September update reduced memory usage by 25% and increased editing speed by about 1.5 times.

The greatly evolved Piapro Studio NT2

--Is it correct to understand that the same changes have been made to the features included with "Hatsune Miku NT"?
Koizumi: Yes, that's right. The GUI of "Piapro Studio NT2" itself has been sped up. The process of transferring pitch curve data has also been changed and sped up. That's the editing speed increase I just mentioned, about 1.5 times faster.

Koizumi: The speed of computer data analysis is also increasing, more than doubling, and in some cases even ten times faster. Automatic Control is also more than ten times faster.

--I heard that improvements were made to the bass range immediately after the release of "Hatsune Miku NT (Ver.2)."
Okamoto: Yes. We worked on improving the bass range in preparation for the future development of other virtual singers. As a result of reviewing the calculation formula in the synthesis algorithm, the synthesis feel of the so-called low-frequency synthesizer has improved, making it easier to listen to. By adjusting the phase and increasing its precision, we have improved not only the bass range but also the overall sound quality.
Sasaki: This has also had an impact on the improvement of libraries such as "KAITO" and "Megurine Luka," and it is an area where we are consciously making adjustments.

--What other features have been added?
Okamoto: We've added a keyboard shortcut editing function. Another major addition is the planned implementation of a multi-output function.

Keyboard shortcut function included

--When you say "multi-output", does that mean outputting each track separately?
Okamoto: Yes, that's right.

--So they'll all be output as WAV files, without mixing. That's what is planned.
Okamoto: Speaking of other small details, we've also added screen scrolling behavior. There are various ways for the screen to switch interface and follow movement and display size, such as changing around the seek bar, or going all the way to the edge of the screen. The VOCALOID editor was the latter, so this is designed to accommodate that.

--So you will now be able to switch between the two scrolling interfaces (with Piapro Studio NT2)?
Okamoto: Yes, that's right.

 

[MagicalMiku]

DTM Station interview to Crypton Future Media about Hatsune Miku NT [Part 4]
(translated from the original article in japanese, here):
https://www.dtmstation.com/archives/74903.html

On November 19th, Piapro Studio and VOCALOID 6 became compatible.
--And then there was a big announcement on November 19th.
Koizumi: Yes. At 15:00 on November 19th, Yamaha's VOCALOID 6 .vpr file format became readable by Piapro Studio. Conversely, VOCALOID 6 can now read Piapro Studio .ppsf files.

Piapro Studio NT can now read VOCALOID .vsqx, .vpr, and .vsq files.

--This is a big change. What exactly will it be possible to do?
Okamoto: Piapro Studio will be able to load tracks, lyrics, and notes from .vpr files. On the VOCALOID 6 side, there are VOCALOID tracks and AI tracks, and VOCALOID tracks can load parameters up to V4. AI tracks can load lyrics and notes.

VOCALOID 6 Editor can now also read .ppsf files.

--What about loading .ppsf files into VOCALOID 6?
Okamoto: Similarly, you can load tracks, lyrics, and notes. However, while VOCALOID 6 and V5 allow you to change the language and lyrics for each part, .ppsf doesn't support that, so you'll need to separate the tracks and singers. There are other parameters added in V5, but they aren't supported.

--Is it possible to export?
Okamoto: Piapro Studio will be able to read .vpr files, but there is no export function. VOCALOID 6 will be able to read .ppsf files.

--I see. As the NT series and VOCALOID series become able to coexist, it would be easier if there was no need to retype lyrics at least.
Koizumi: That's right. You'll be working on both NT and VOCALOID 6, and we've developed engine modules and Automatic Control modules in-house and have them as separate units, which allows us to flexibly support software and plug-ins. We're now in an age where AI is being used not only in generative AI but also in DAWs, so we have our own engine as a technology that connects us to these latest trends.

VOCALOID Hatsune Miku development status
--Finally, could you tell us about the development status of the VOCALOID version of Hatsune Miku?
Koizumi: We first unveiled the audio on August 31st as part of a stage event, and on October 31st we released a singing demo video with orchestral instruments and a teaser page. We hope to provide further updates on this teaser page.
Sasaki: It's been a long journey so far... We've put in a lot of effort to ensure that Miku's voice sounds beautiful in the core part of the VOCALOID6 engine, where it meshes with human voices. We hope to have an opportunity to explain this together with Yamaha at some point.

--Please tell us about the upcoming schedule.
Koizumi: In December, we will release a beta version called the Early Access version to users of the VOCALOID version of Hatsune Miku. We plan to provide a Japanese database that can be used for 39 days.
Sasaki: As happened with Hatsune Miku V4X, we will monitor the reaction and incorporate any feedback we get, and we are considering a release in the first half of 2026.

--What will the name be then?
Koizumi: It will be "Hatsune Miku V6". The tentative name announced in August 2024 was "Hatsune Miku V6 AI," but we decided not to call it AI. It is AI (for Vocaloid 6 engine), but it was created through trial and error to achieve the Hatsune Miku original feel.
Sasaki: The AI is a method to make Miku's voice more expressive, and is not being used as a way to replace Miku with a human voice. It took a while to work out the details, but we hope to make the most of the good points of both VOCALOID and NT.

--thank you very much.

The author of this article: Ken Fujimoto
A writer who writes mainly about DTM, digital recording, and digital audio. He has been serializing his Digital Audio Laboratory in Impress's AV Watch magazine since 2001. He has also written many books, including "Complete Cubase Operation Guide" (Rittor Music Japan) and "Vocaloid Technology Theory" (Yamaha Music Media). His hobby is solar power generation, and he has been using solar power for his home electricity since 2004. He is also the manager of three power plants he operates.