Dreamtonics
Tokyo studio Dreamtonics makes two tools: Synthesizer V Studio 2 Pro, an AI singing voice synthesizer, and Vocoflex, a real-time voice morpher. Voices come from licensed recordings of real singers. Both are one-time perpetual licences, with a 14-day free trial.
What is Dreamtonics?
Dreamtonics is a Tokyo-based startup specialising in electronic music and voice technology. Its two flagship products are Synthesizer V Studio 2 Pro, an AI singing voice synthesizer, and Vocoflex, a real-time voice morpher.
Synthesizer V Studio 2 Pro behaves like a sequencer, not like a prompt box. You enter notes, type lyrics onto them, choose a voice, shape the expression, and the software renders a sung performance. Underneath sits a deep neural network engine that works from real vocals, recorded and licensed from real singers. Dreamtonics reports a human-level naturalness rating from a blind test run with 50 participating musicians. Cross-Lingual Synthesis lets any voice sing in six languages - English, Japanese, Korean, Mandarin, Cantonese and Spanish - whether or not that singer recorded them.
Control is where version 2 concentrates its effort. AI Retakes regenerate pitch, timbre or timing on a phrase that did not land. Vocal Modes switch between chest, belt and breathy delivery, alongside Tension, Breathiness, Voicing, Gender and Tone Shift parameters. Pitch editing is backed by Smart Pitch Controls, a Mouth Opening parameter and a dedicated phoneme timing panel, while Live Rendering redraws the waveform as you work. Voice-to-MIDI turns a sung or spoken take into notes you can re-render with another voice. Offline rendering is up to 300% faster than in version 1.
Vocoflex is the other half of the catalogue. Import, record, blend, morph or replace a voice; move an X/Y pad for gender, randomise the tone, share a result as a hexadecimal code. Forty licensed voice presets ship with it. It runs live from a microphone or as a plugin inside a DAW, with 105 ms of latency at maximum quality and 35 ms in minimum-latency mode.
Both run standalone and as VST3, AU, AAX and ARA plugins, with verified support for ABILITY Pro, Ableton Live, Cakewalk, Cubase, FL Studio, Logic Pro, GarageBand, Pro Tools, REAPER and Studio One. Synthesis executes locally on your own machine: once a voice has been bought and activated, creative content is not uploaded. Third-party voices come from partners including AHS Co., Ltd., Eclipsed Sounds, AUDIOLOGIE, INTERNET Co., Ltd., E-CAPSULE, Quadimension and Studio ENTRE, and a separate Synthesizer V Engine SDK is offered to businesses. The company was incorporated in February 2019 and its representative is Kanru Hua.
What it does
- Write a vocal melody and hear it sung back immediately
- Put vocals on a demo without bringing a singer into the room
- Regenerate a take you are not happy with - pitch, timbre or timing - using AI Retakes
- Make a voice sing in a language it never recorded, across English, Japanese, Korean, Mandarin, Cantonese and Spanish
- Convert your own voice to MIDI, then re-render the same performance with a different voice
- Build harmonies, doubles and choir parts from a single written line
- Reshape the timbre of an existing vocal in real time with Vocoflex
When to use Dreamtonics / When not to
A quick filter to help you decide if Dreamtonics is the right fit.
When to use Dreamtonics
- Music producers and composers who need to demo a vocal line without booking a singing session
- Songwriters who want to hear a melody and its lyrics sung back while the song is still being written
- Film and TV composers and sound designers - Charlie Fisher, Jordan Rudess and Alan Branch are named among users
- Music teachers and students: educational licences are half price, and the software is documented in higher education use
- Studios that need processing to stay on the machine, and clients who need commercial use with no revenue cap - the licence allows both
When not to use Dreamtonics
- Anyone expecting to type a prompt and get a finished song - this is not a track generator, you enter notes and lyrics yourself
- Buyers looking for a cheap monthly subscription - licences are perpetual and run from 89 to 249 USD
- Anyone wanting to clone an identifiable person's voice, or to push distorted audio, choirs or backing music through Vocoflex - the first is forbidden by the EULA outside strictly private use, the second simply does not work
- Linux users, and anyone who needs a mobile or web app - version 2 is Windows and macOS only, as a desktop app and plugin, where version 1 covered Linux
- Teams that require a signed DPA or a documented GDPR framework - nothing of the kind is published
How to use Dreamtonics
A typical end-to-end flow, from setup to results.
- Download the installer from the Free Trials page - a .exe for Windows, a .pkg for macOS
- Run the software standalone, or load the plugin into your DAW as VST3, AU, AAX or ARA
- Open the sample project supplied with the trial to see how a finished track is put together
- Enter notes in the piano roll, or import a MIDI file
- Type the lyrics onto the notes and choose a voice
- Shape the performance: expressions, vocal modes, pitch curves and phoneme timing
- Use AI Retakes to regenerate a take, adjusting pitch, timbre or timing
- With Voice-to-MIDI, convert your own voice or a vocal sample to MIDI, then re-render it with a different voice
- For Vocoflex, import a target voice sample - ten seconds is enough - or generate a voice, then sing or speak into the microphone
- Buy and activate through the Store and the Product Manager; each voice carries its own 7-day trial, and the user manuals live at svdocs.dreamtonics.com and vfxdocs.dreamtonics.com
Pros & Cons
Pros
- Voices are built from recordings of real singers, licensed with informed consent claimed by the publisher
- Processing stays local: once a voice is activated, creative content is not uploaded
- One-time purchase rather than a subscription, with free updates inside the same major version
- Unusually fine control over pitch, timbre, timing, phonemes and vocal modes
- Broad DAW and plugin-format coverage, including ARA
- Six sung languages, with cross-lingual synthesis across all of them
- Educational licences at 50% off, volume licences on request, and a substantial specialist press record (Sound On Sound Gear of the Year 2025, MusicRadar, Future Music, Music Tech)
Cons
- The GDPR is not mentioned anywhere on the site
- No data processing agreement, no sub-processor list, and no documented way to opt out of training
- No data hosting country is stated
- The entry cost adds up: the software alone is not enough, and each additional voice costs 79 USD
- Version 2 drops Linux, which version 1 supported
- Vocoflex requires KYC identity verification and embeds an inaudible watermark identifying the user - and the EULA lets Dreamtonics disclose that identification without the user's consent
- The EULA can be amended without notice, the Japanese version prevails and jurisdiction is Tokyo; support runs through a Zendesk help centre and a forum, with no support email address
Pricing & Plans
There is no free plan. What is offered instead is a 14-day free trial of the software and a 7-day trial for each voice, with no subscription and no automatic charge when a trial ends. All prices are quoted in US dollars and every licence is a one-time perpetual purchase. The entry point is Synthesizer V Studio Pro, the version 1 product, at USD 89.00. Synthesizer V Studio 2 Pro costs USD 99.00 and includes one voice of your choice, while each additional Dreamtonics Synthesizer V voice is USD 79.00. Vocoflex is USD 199.00. Two bundles lower the total: Synthesizer V Studio 2 Pro with an extra voice at USD 149.00, and Synthesizer V Studio 2 Pro with Vocoflex at USD 249.00. Qualified schools, teachers and students receive a 50% discount, and volume licences for organisations, studios and educational institutions are available on request through the Contact page.
- standard offline rendering
- pitch and timbre retakes
- manual mode
- Windows
- Linux and macOS
- one voice included
- rendering up to 300% faster
- pitch
- timbre and timing retakes
- Smart Pitch Controls
- Mouth Opening
- phoneme timing panel
- Windows and macOS
- USD 79.00 each
- perpetual licence
- 40 presets
- VST3
- AU and AAX plugins plus standalone
- USD 149.00
- USD 249.00
Data, GDPR & hosting
A consolidated view of how Dreamtonics handles your data.
GDPR overview
There is no mention of the GDPR anywhere on the site. Neither GDPR nor General Data Protection Regulation appears in the pages collected, in the visible text or in the raw HTML. Dreamtonics does publish a privacy policy, last updated on 16 April 2024, and it sets out rights that look GDPR-shaped - access, copy, rectification, erasure, objection and restriction - but it never ties them to the Regulation. There is no Article 27 EU representative and no named data protection officer. No data processing agreement is published or offered, and no list of sub-processors exists. No hosting country or region is stated. Governing law is Japanese, disputes go to the Tokyo District Court, and the Japanese version of the EULA prevails. Read the position as undocumented rather than refused.
Who owns the data?
Dreamtonics Co., Ltd. and its affiliates describe themselves as the personal data processor for your personal information. They hold account data (name, email, phone, address), payment details, device data (make, model, hardware serial number, IP address, microphone and audio peripherals), geolocation with consent, usage and crash diagnostics, and support correspondence. Creative content differs: once a voice is activated, synthesis runs locally and your work is not uploaded. Stated rights are access, copy, completion, correction and deletion, plus refusing or restricting use and asking how processing rules are applied; write to privacy-policy@dreamtonics.com, answered within 15 business days. Under the Vocoflex EULA, a watermark identifies the user in any continuous ten-second segment.
Reuse rights
Dreamtonics lists its own purposes plainly: providing support, processing transactions and performing contracts, communicating about your account and about changes, personalised recommendations for products and services including third-party ones, security and fraud prevention, and legal compliance. Any other purpose requires your consent. International transfers are announced - to affiliates in other countries and to third parties, with consent - and the safeguards cited are standard contractual clauses, consent and anonymisation. Session and persistent cookies handle preferences and audience measurement. Nothing anywhere states whether customer data is used to train models, so treat that question as unanswered rather than settled. On your side of the deal, reuse is generous: the licence covers commercial use with no revenue limit, and rendered vocals are yours to release. Three things are ruled out - selling the output as sample packs, training machine learning models on the generated audio, and redistributing the software under another name. The Vocoflex EULA repeats the point: voices processed by the software must not be used for training machine learning models.
Data retention & training
Hosting summary
No hosting country and no hosting region are stated anywhere on the pages published by Dreamtonics. The privacy policy announces cross-border transfers, both to affiliates in other countries and to third parties, without naming a single destination, and cites standard contractual clauses, consent and anonymisation as the safeguards. The practical mitigation here is architectural rather than contractual: once a voice has been bought and activated, synthesis runs locally on your own workstation, so the creative content itself never leaves it. Trial voices are the exception - data travels to Dreamtonics servers for licence verification, and offline mode is unavailable while a trial voice is loaded. The website is served behind Cloudflare on anycast address 104.26.0.147, which says nothing about where customer data is actually stored. Jurisdiction is the only firm anchor: Japanese law, with the Tokyo District Court named for disputes. If your procurement process needs a named hosting location, you will have to ask for one, because none is published.
Things to keep in mind
Risks and trade-offs to weigh before adopting Dreamtonics.
- Vocoflex requires identity verification (KYC) through a third-party provider before you can use it
- An inaudible watermark is embedded in every voice the software processes, identifying the user from any continuous ten-second segment, and it survives mixing and phone transmission
- The EULA allows Dreamtonics to disclose that identification without the user's consent, to complainants or judicial authorities, free of charge or for payment
- Reproducing an identifiable person's voice without permission is forbidden outside strictly private use - clearing that is your responsibility, not the software's
- The EULA can be amended without notice, the Japanese version prevails, and disputes go to the Tokyo District Court, a long way from most buyers
- The licence is tied to one person and one device and cannot be transferred to anyone else
- No GDPR mention, no data processing agreement and no sub-processor list: if your work touches client data or regulated material, there is nothing to hand a legal team
Setup & Integrations
Technical difficulty
Installation is ordinary: one installer per system (.exe for Windows, .pkg for macOS), nothing to deploy, no cloud account for synthesis - only an internet connection for licence verification. Requirements are modest: Windows 10/11 or macOS 11+, Intel Core i5, AMD Ryzen or Apple Silicon M1+, 4 GB of RAM, 1 GB of disk (200 MB for Vocoflex). The learning curve is musical, not technical: piano roll, expression parameters and phoneme editing take practice. Ten official English video tutorials, online manuals and a forum help. ARA needs a recent DAW: Cubase 12+, REAPER 6+, Pro Tools 2024.10+, Studio One 5+.
Deployment
Integrations
Supported languages
Behind Dreamtonics
Social
Resources
All the official URLs gathered for verification and reference.
Frequently asked questions
Is this a song generator?
Do I need a subscription?
Can I try it before buying?
Can I use the results commercially?
Does my audio leave my computer?
Which languages can the voices sing?
Which DAWs and systems does it support?
Is there a minimum age?
Is the GDPR covered?
How do I contact support?
Should you pick Dreamtonics?
Dreamtonics reads as a mature operation rather than a launch: a company identified in Tokyo since February 2019, with a named representative and a postal address, three separate contractual documents, products documented down to hardware specifications and verified DAW lists, and steady updates and specialist coverage through to 2026.
The positioning it claims is ethical vocal AI: voices built from recordings licensed from real singers, with their consent. Whatever weight you give that, the commercial model is refreshingly legible - one-time purchase, free trial, no subscription, no automatic charge when the trial ends. For a producer who needs a vocal on a demo tomorrow morning, the path from download to first render is short.
Two reservations deserve reading before you buy. The first is compliance documentation, which is very thin: no mention of the GDPR anywhere on the site, no data processing agreement, no sub-processor list, no hosting country. A privacy policy exists and was last updated in April 2024, but it never connects itself to the Regulation. Inside a procurement process, that gap will be noticed.
The second is specific to Vocoflex: identity verification before use, an inaudible watermark that identifies you from ten continuous seconds of audio, and an EULA clause allowing Dreamtonics to disclose that identification without your consent. It is a deliberate anti-abuse design, stated openly, but it is a trade to make knowingly.
Judged as a music tool, Synthesizer V Studio 2 Pro is strong and unusually controllable. Judged as a vendor whose paperwork you have to sign, Dreamtonics still has work to do.
- Choosing a selection results in a full page refresh.
- Opens in a new window.