Kanru Hua

CEO

Dreamtonics

About Me

Kanru Hua is the founder of Dreamtonics, the pioneering startup behind Synthesizer V, established in 2019. Under his leadership, the company has emerged as a global leader in vocal synthesis and voice conversion, developing breakthrough tools like Vocoflex, and since 2026, broadening its scope into musical instrument synthesis with Instrument X. A self‑taught programmer and researcher, Kanru bridged signal processing with state‑of‑the‑art generative models, delivering a versatile arsenal of music production tools for composers and musicians.

In recognition of his impact, he was named to Forbes JAPAN 30 Under 30 in 2022.

Sessions

  • Interface and Render Scheduler Design for Neural Audio Synthesizers: Part 2
    16:20 - 16:50 UTC | Tuesday 10th November 2026 | Bristol 3
    Beginner
    Intermediate
    Advanced

    We examine a new paradigm of virtual instruments based on locally-run deep neural networks, and discuss its implications on overall interface design. "Interface" carries multiple meanings here—the UI-engine interface, the user interface, and the host-plugin interface—all of which must be considered in tandem to deliver a smooth user experience. Part 2 of this talk presents a high-level solution to questions raised in Part 1. We discuss strategies to assign task priorities, optimize chunking to reduce overheads, and present ideas for future DAW-plugin interface design.

  • Interface and Render Scheduler Design for Neural Audio Synthesizers: Part 1
    12:20 - 12:50 UTC | Tuesday 10th November 2026 | Bristol 3
    Beginner
    Intermediate
    Advanced

    We examine a new paradigm of virtual instruments based on locally-run deep neural networks, and discuss its implications on overall interface design. "Interface" carries multiple meanings here—the UI-engine interface, the user interface, and the host-plugin interface—all of which must be considered in tandem to deliver a smooth user experience. Part 1 of this talk focuses on defining and analyzing the core problems. We introduce multi-stage rendering pipelines, caching strategies, and resource prioritization techniques designed to ensure responsiveness in heavy audio editing workflows.

  • Scripting Architecture for a DAW-like Plugin

    How we implemented Lua and JavaScript scripting for Synthesizer V Studio
    16:20 - 16:50 UTC | Tuesday 11th November 2025 | Bristol 2
    Intermediate
    Advanced

    Audio software at scale is often met with thousands of feature requests across diverse user groups with varying workflows. This talk presents scripting as a solution to bridge the gap between limited development resources and a growing backlog. This talk will guide you through the architecture of Synthesizer V Studio, a vocal synth with DAW-like infrastructure and strong scripting capabilities. We will cover the manipulation of all types of objects (Track/Group/Note/Phoneme) in the project, seamless undo/redo integration, designing an object-oriented interface that supports both Lua and JavaScript and the ability to allow users to create custom UIs within clearly defined […]

  • Responsible AI for Offline Plugins

    Tamper-Resistant Neural Audio Watermarking
    12:20 - 12:50 UTC | Tuesday 12th November 2024 | Bristol 2
    Intermediate
    Advanced

    In the era of rapid generative AI advancements, how do audio plugin developers harness the power of deep learning, while safeguarding against misuses and ensuring regulatory compliance? Existing audio watermarking algorithms are designed to run on the cloud, and when deployed offline, are trivial to bypass when faced against experienced hackers. In this talk, I introduce an innovative algorithm developed for Vocoflex, our AI powered voice transformation plugin. Instead of first generating the voice and then adding the watermark in a post-processing step, our approach involves an end-to-end trained neural network that generates voices with inaudible watermarks already embedded. Furthermore, […]