Dreamtonics
75+ AI voices · 6 languages
Get
Dreamtonics · Synthesizer V

Your virtual singer, in any language.

Synthesizer V Studio 2 Pro turns MIDI notes and lyrics into AI vocal performances. 75+ voices recorded with real, licensed singers. Cross-lingual synthesis across six languages. Polyphonic AI choirs. Built by Dreamtonics in Tokyo since 2018.

Get the studio → How it sings
75+ ethically-trained voices Cross-lingual · 6 languages Polyphonic AI choir
CH 01 · VOICE
Session at Twilight · F major
WANING
E
Etta · Powerful
VOCAL MODE · BELT
EN JA ZH
48 kHz · WAV · 32-BIT
Copyright-protected
Workflow

Three movements from silence to song.

You compose. The AI sings. The output is yours, copyright and all.

I

Write the melody.

Open Synthesizer V Studio 2 Pro and draw notes on the piano roll. Add your lyrics syllable by syllable. The interface is built for songwriters — a familiar MIDI workflow with vocal-specific tools layered in. Sketch fast; perfect later.

II

Choose a voice.

Pick from 75+ ethically-trained AI voices — English, Japanese, Mandarin, Cantonese, Spanish, Korean. Each voice comes with multiple Vocal Modes (Powerful, Breathy, Chest, Belt, Resonant) that you automate across the song for real expression.

III

Hear it sing.

Live Rendering visualizes the vocal waveform in real time as you tweak. Tune the expression, bounce a WAV stem, drop it into your DAW. Output is copyright-protected and commercially licensed out of the box — release, distribute, monetize.

Cross-lingual synthesis

One voice. Six languages. Same identity.

This is the technology that sets Synthesizer V apart. A single AI voice — English-native, Japanese-native, Chinese-native — can sing fluently across Japanese, English, Mandarin, Cantonese, Spanish and Korean. The voice's tonal identity stays consistent; only the phonetics adapt. Write a verse in English, a hook in Japanese, a bridge in Mandarin, all in the same singer's voice. No switching models. No mismatched timbre.

The honest trade-off: fluency varies slightly per voice and per language. A Japanese-native voice may carry a faint accent in Spanish. The team ships improvements regularly, and for professional production it is already at studio-ship quality across the core six.

VOICES AVAILABLE · MAI 2 · ETTA · DANNY · YI WEI · HXVOC · YUN HUA · LANG CHUAN · LIAM · SAKI · KOHARU RIKKA · YUMA · + 65 MORE
日本語
JAPANESE
English
ENGLISH
普通话
MANDARIN
廣東話
CANTONESE
Español
SPANISH
한국어
KOREAN
Built for

Anyone composing with a singer.

Synthesizer V is a producer's tool — full control, full copyright, full responsibility for the song.

C

Composers

Film, game, anime and TV score work. Render real-sounding vocals in any language without booking session singers per cue.

J

J-pop & anime

The home territory. Japanese, Mandarin and Cantonese voices built specifically for the idol, vocaloid and anime production traditions.

P

Producers

Sketch vocal demos in 20 minutes. Audition different singers on the same melody before committing to a session.

S

Songwriters

Hear your topline in a real voice while writing — not your scratch vocal, not a MIDI piano placeholder. The melody sells itself.

What's inside

The full vocal observatory, under one roof.

Synthesis is the headline. The rest of the kit is what makes it ship-ready.

Flagship

AI singing voice synthesis.

Enter MIDI notes and lyrics; the AI sings the part with real-singer naturalness. 75+ voices recorded under commercial licensing agreements. Industry-leading vocal expression with breath, dynamics and phrasing intact. Output is copyright-protected — release commercially without clearance.

Vocal Modes

Powerful, Breathy, Chest, Belt, Resonant, Bright, Dark, Scream — automate across a song.

Polyphonic AI Choir

Up to 16 simultaneous AI voices. Real choir performances, trained with consent on full choirs.

Live Rendering

Real-time waveform visualization as you tweak. Shorter idea-to-sound cycle, less ear fatigue.

Vocoflex add-on

Import, record, blend, morph, replace and transform vocals into any voice from the library.

Anywhere

Cross-platform desktop — Mac, Windows, Linux.

Synthesizer V Studio 2 Pro runs natively on all three desktop platforms. No mobile or iPad app — desktop is the only destination. Output bounces straight to WAV for Ableton, Logic, FL Studio, Pro Tools, Cubase, Studio One, Reaper, Bitwig. Free 7-14 day Pro trial to try the workflow before purchasing.

Compared

Where Synthesizer V wins. Where it doesn't.

The AI vocal market is splitting by workflow. Pick by the workflow, not the brand.

  Synthesizer V Vocaloid Suno / Udio ACE Studio
Workflow MIDI + lyrics MIDI + lyrics Text-to-song MIDI + lyrics
Vocal naturalness (2026) Industry-best Stylized / sampler Strong, full-mix Strong
Cross-lingual synthesis 6 languages Per-voice only Per-prompt Limited
Polyphonic AI choir Up to 16 voices No n/a No
Ethical training (consented) Yes — fully licensed Yes Disputed Yes
Output copyright-protected Yes Yes Disputed in some markets Yes
Anime / vocaloid culture Growing Dominant — original home n/a Limited
Text-to-song speed No No 30 seconds No
Voice library size 75+ voices 60+ voicebanks n/a (text prompts) Smaller library
Honest read: Synthesizer V is the wrong tool for one-button text-to-song — Suno and Udio do that, and they do it fast. It is also the wrong tool if you live inside the established vocaloid / anime idol fandom — Vocaloid still owns that scene's culture and brand. Synthesizer V's specific lane is studio-grade controllable AI vocal synthesis with the largest cross-lingual capability on the market and copyright-clean output. Different tools, different jobs.
From the studios

What composers say after a session.

Including the ones who're still wrestling with the learning curve.

★★★★★

For anime production I needed Japanese AND English vocals on the same character voice. Synthesizer V is the only tool where that question has a real answer. Mai 2 across both languages is unreal.

K
Kenji M.
Anime music supervisor, Osaka
★★★★★

Etta's belt mode is a session vocalist on a great day. We sketch hooks in twenty minutes that used to take a half-day booking. The copyright clarity is why we can actually release them.

M
Maja L.
Pop producer, Stockholm
★★★★★

Output quality is genuinely the best out there. But the learning curve is real — drawing every syllable, tuning Vocal Modes, getting cross-lingual fluency right takes hours. Don't expect Suno-fast. Once you're past it though, you can ship.

D
Diego R.
Songwriter, Mexico City
The story

A Tokyo lab, betting on artist-first AI.

Dreamtonics shipped the original Synthesizer V in 2018 from Tokyo with a thesis that has aged extremely well: AI vocal models should be built on consented studio recordings, not scraped data. The bet was unfashionable at the time. By 2026, with copyright fights swirling around scrape-based generators, it looks like the only sustainable model in the room.

Every official voice — from Mai 2 and Saki in Japanese, to Etta and Danny in English, to Yi Wei and Yun Hua in Chinese — is recorded under commercial licensing with the original singer, with months of session work and feedback. The January 2026 Polyphonic AI Choir update came after two full years of recording actual choirs in-house. That's the cadence.

The honest trade-offs sit in plain sight. The learning curve is real — drawing notes and syllables takes more time than typing a prompt. The cost stacks: roughly $195 for Studio 2 Pro plus $99-$199 per voice database, and serious users own several. Synthesizer V is desktop-only, no mobile. And cross-lingual fluency, while industry-best, varies between voices.

If you want a song from a sentence, this isn't your tool. If you want full control of a vocal performance you can ship under copyright, in six languages, from a singer who consented to be there — this is the only studio in this lane.

FAQ

Real questions, real answers.

What you wanted to know before downloading the trial.

What exactly is Synthesizer V?
Synthesizer V is an AI singing voice synthesis application built by Dreamtonics. You enter MIDI notes for the melody and lyrics for each note, choose a voice from the library, and the AI sings the part for you — complete with breath, expression and dynamics. It is built for composers, songwriters and producers who want a real-sounding vocal performance without booking a studio singer. The flagship product is Synthesizer V Studio 2 Pro; a free Basic version also exists for trying the workflow.
Is the AI ethically trained?
Yes — this is one of the platform's clearest advantages over scrape-based competitors. Every Synthesizer V voice database is trained on recordings made by real singers under commercial licensing agreements. Dreamtonics records full choirs and individual vocalists in studio sessions specifically for AI training. There is no scraped data, no unconsented use of streaming-platform vocals, and the company has been open about this practice since the 2018 launch.
Can I use Synthesizer V output commercially?
Yes. Synthesizer V Studio 2 Pro and all official Dreamtonics voice databases include commercial use licenses out of the box. You can release, distribute and monetize songs that include the AI vocals. Output is copyright-protected — unlike fully generative AI tools where some jurisdictions don't grant copyright to pure-AI output. This makes Synthesizer V workable for major-label release, sync placement and commercial production, not just hobbyist sketches.
What's the difference between Pro and Basic?
Synthesizer V Studio Basic is a free, restricted version that lets you experience the workflow with a limited subset of voices and features. Synthesizer V Studio 2 Pro is the full paid application — currently around $195 for a permanent license — and includes one complimentary voice to start. Additional voices are purchased separately from the Dreamtonics Store, typically $99-$199 each. Pro unlocks cross-lingual synthesis, the AI choir features, full export quality and the complete vocal mode toolkit.
How many voices are available?
75+ AI voice databases at this time, with new voices added regularly. The lineup includes English voices (Etta, Danny, Liam, HXVOC), Japanese voices (Mai 2, Saki, Koharu Rikka, Yuma), Chinese voices (Yi Wei, Yun Hua, Lang Chuan), and more across pop, rock, R&B, EDM, classical bel canto, anime and J-pop categories. Each voice ships with multiple Vocal Modes — Powerful, Breathy, Chest, Belt, Resonant, Bright, Dark, Scream — that you automate across a song for expression.
What does cross-lingual synthesis mean?
Synthesizer V Studio 2 Pro's cross-lingual feature lets a single voice perform lyrics in six languages: Japanese, English, Mandarin Chinese, Cantonese Chinese, Spanish and Korean. The voice keeps its tonal identity across languages — for example, you can have an English-native voice like Etta sing a Japanese verse and a Korean chorus without switching models. Fluency does vary slightly per voice and per language, but the system is good enough for professional cross-lingual production.
How does Synthesizer V compare to Suno or Udio?
Suno and Udio are text-to-song generators — type a prompt, get a finished audio track. Synthesizer V is the opposite workflow: you write the song yourself in MIDI and lyrics, the AI performs the vocal only, you keep everything else. That extra control is why pros use it for commercial release. The honest trade-off is speed: if you want a finished demo in 30 seconds, Suno wins. If you want a vocal you can fully edit and ship under copyright, Synthesizer V wins. Different lanes entirely.
Does it work as a DAW plugin?
Synthesizer V Studio 2 Pro runs as a standalone application on macOS, Windows and Linux. Plugin and tighter DAW-integration support has been expanding through Studio 2. For most users the workflow is to compose and render vocal stems inside the Synthesizer V app, then drag the WAV into Ableton, Logic, Cubase, Pro Tools, Studio One or Reaper. There is no mobile or iPad version — desktop is the only platform.
What are the honest limitations?
Three honest ones. First, the learning curve is real — drawing MIDI, entering lyrics syllable-by-syllable and tuning Vocal Modes takes more time than typing a prompt into Suno. It's a producer's tool, not a casual generator. Second, the cost adds up: $195 for Pro plus $99-$199 per voice database. Power users buy multiple voices. Third, cross-lingual fluency varies — voices native to one language sometimes have a slight accent in others. The team ships steady improvements but this is the trade-off for genuinely controllable AI vocals.
Press render

The studio is open after dark.

Free 7-14 day Pro trial. 75+ AI voices ready in the Product Manager. Six languages, copyright-clean, made in Tokyo.

Get the studio → Try free