The company behind Synthesizer V models strings, woodwinds and brass with neural networks instead of sample libraries, completing a three-product trilogy.
You might have seen Dreamtonics tease that something new is coming over the past couple of weeks. Now the secret’s out.
Dreamtonics has spent years teaching software to sing, and Instrument X brings the same approach to a full orchestra – one that generates strings, woodwinds and brass from your notation, rather than playing back recordings of them.
Best known for the vocal software Synthesizer V and Vocoflex, the Tokyo company has built Instrument X as the third piece in what founder and CEO Kanru Hua calls a trilogy. The note-level control developed for synthetic vocals now applies to an orchestral score, with each part generated as you write rather than pulled from a library.
Instrument X at a glance
• Officially launched now via the Dreamtonics Store
• Fifteen instrument expansions available across strings, woodwinds and brass
• Compatible a s a standalone app and VST3, AU and AAX plugin formats
Catch up on all the latest news here.
How it works
Instrument X uses Neural Acoustics Modeling in place of traditional sample playback. You write the notation, and the software generates the bow, breath and key articulation of every note, responding to pitch, duration and phrasing dynamically inside your DAW. Since nothing is being streamed off disk, techniques that a sample library keeps in separate patches can be stacked on a single note, a col legno or a stem-muted brass note layered onto a tremolo. The whole thing runs on a standard laptop, with no internet connection or dedicated GPU required.
Each instrument is captured with up to 11 microphone feeds, from close spots through a Decca tree to room ambience, so a part has the depth of a real session instead of a flat, dry sample. That many feeds usually creates phase problems for this kind of synthesis, but Dreamtonics has a patent-pending fix: it measures a studio’s acoustics with a custom dodecahedral speaker array, then uses that to separate the room’s sound from the dry instrument.
“It completely reshapes the orchestral mockup workflow,” says film composer Federico Ricci. “The level of expression you get, without lugging around 100 GB of samples, gives this tech incredible potential for the future.”
What it does
The Unified Articulation Editor sets playing techniques for you automatically, or hands over control note by note. Predictive Dynamics Lane sets how hard each part plays from the notes themselves, and moving it changes the tone of the instrument rather than just the level – the way a real player will play dynamically. Overlapping notes ring out as separate voices in one track instead of cutting each other off, which helps with dense chords. A Microphone Mixer Panel balances the feeds to shape the sense of space.
If you want a different read of a phrase, AI Performance Retakes generate alternative takes, either across a whole section or on a single element like pitch or intensity, without redoing the take from scratch.
The launch lineup spans three collections: Strings (violin, viola, cello and contrbass), Woodwinds (flute, piccolo, oboe, clarinet in A and bassoon) and Brass (trumpet, trombone, horn and tuba). Each instrument comes with its own set of techniques and five performance styles in the Synthesizer V mould, so you can create a more dynamic, natural sound.
The Dreamtonics setup
Instrument X works as a plugin alongside Synthesizer V Studio 2 Pro and Vocoflex, so backing instruments and synthetic vocals run much the same way. It also loads standalone and in VST3, AU and AAX formats. It’s available now through Dreamtonics, with the core editor is free to download with any Instrument Expansion. Each expansion comes with a 14-day trial.
For more information, head to Dreamtonics.
