Skip to Content

Features

Virtual Video Director automates multi-camera switching by listening to who’s speaking and triggering your video switcher, PTZ cameras, and broadcast infrastructure in real time. From a two-camera podcast to a 128-channel broadcast, VVD handles the cuts so you can focus on the show.

Mills Level Sensor

The Mills Level Sensor is a proprietary ballistic algorithm designed by Toby Mills. Rather than looking at absolute audio levels, it analyses the relative differences between all audio signals in real time to determine who is speaking.

Under the hood, the sensor uses moving statistics with adaptive threshold detection. It maintains a rolling window of audio level data across every channel and identifies the active speaker by comparing each signal against the group — regardless of volume, distance from the microphone, or gain settings.

Zero audio calibration is required. The algorithm adapts automatically to any voice, any microphone, and any room. A configurable attack/decay envelope controls how quickly channels activate and deactivate, giving you smooth, natural switching behaviour.

For meter-based audio sources such as Wheatstone Blade consoles, ClearOne DSPs, and vMix Direct Connect, the sensor includes a dedicated fallback path that applies the same relative-difference logic to pre-metered data.

Mills Smart Sensor (AI Voice Activity Detection)

The Mills Smart Sensor adds an on-device neural network layer to voice detection. It runs directly on your machine — no cloud, no latency, no privacy concerns.

The model processes audio at 16 kHz sampling in 512-sample windows (32 ms chunks) through a 2-layer RNN with a 128-unit hidden state. Each channel maintains its own state tracking with context persistence, so the network builds up a per-speaker understanding over time.

DirectML GPU acceleration is supported across NVIDIA, AMD, and Intel GPUs, with automatic CPU fallback when no compatible GPU is available.

The system scales to 40+ channels using a multi-threaded worker architecture, distributing inference across available hardware. 

8 Audio Source Types

VVD can receive audio from eight different source types — and you can mix and match them simultaneously in a single project.

  • WASAPI (Windows Audio) — Standard Windows audio devices. Simple setups, no extra drivers needed.
  • ASIO — Low-latency professional audio drivers for studio and broadcast environments.
  • NDI — NewTek network audio over IP. Available from Standard tier and above.
  • OMT (Open Media Transport) — Open-source NDI alternative by vMix. Available from Standard tier and above.
  • Wheatstone Blade — Direct integration with Wheatstone broadcast consoles. Professional tier and above.
  • ClearOne Converge Pro 2 — Conference room DSP integration. Professional tier and above.
  • vMix Direct Connect — Audio levels received via the vMix TCP API. Professional tier and above.
  • Televic Conference — Conference microphone on/off status. Professional tier and above.

30+ Trigger Protocols

When VVD detects a speaker change, it fires triggers across one or more protocols. Each channel supports 8 trigger slots, all of which fire simultaneously with optional per-slot delays.

Video Switchers

  • vMix — Full API integration including Fade, Cut, Preview, Merge, Overlay, and Mix outputs.
  • Blackmagic ATEM — Program, Preview, Cut, Auto, AUX, and multi-M/E support.
  • OBS Studio — WebSocket v5 protocol.
  • TriCaster — Fade, Cut, Macro, and PTZ control.
  • mimoLive — Layer activation and control.
  • Roland Pro AV — V600UHD, VR400UHD, V160HD, and more.

PTZ Cameras

  • Panasonic AW Series — Preset recall and control.
  • PTZOptics — HTTP-based preset recall.
  • NDI PTZ — PTZ control over NDI protocol.
  • VISCA / Sony — Serial and IP-based PTZ.
  • Canon XC — Canon remote camera control.

Network Protocols

  • HTTP / Webhook — GET/POST to any URL on trigger.
  • Custom HTTP — Full control over headers and request body.
  • TCP — Send raw data over TCP connections.
  • UDP — Fire-and-forget datagrams.
  • OSC (Open Sound Control) — Standard show-control messaging.
  • MIDI — Note and CC messages.
  • Art-Net (DMX512) — Lighting and stage control.

Broadcast Infrastructure

  • Pathfinder Core Pro — Telos Alliance routing control.
  • Wheatstone Blade ACI — Salvos, utility mixers, SLIO GPIO, and ducking.
  • Wheatstone Mixer — Channel control, fader automation, and TAKEPROG.
  • Televic Camera Control — Direct conference camera integration.
  • Lawo VSM (Ember+) — Broadcast control system integration.
  • Allen & Heath AHM — Digital mixer preset and zone control.

Scripting & AI

  • JavaScript — Custom scripts executed via the Jint engine.
  • PowerShell — Windows automation scripts.
  • Super Triggers (AI) — Describe your automation in plain English and VVD’s AI generates compiled C# code that runs as a native trigger.

Control Inputs & GPIO

VVD can be controlled by external hardware and broadcast control systems, enabling integration into larger production workflows.

  • Skaarhoj ETH-GPI Link — Raw Panel TCP protocol with 8 HWC (Hardware Component) inputs for physical button control.
  • Lawo VSM / Ember+ — Full Ember+ protocol support for broadcast control system integration.
  • ControlByWeb — Support for X-400, X-440, X-500, and X-600M web-enabled I/O modules.
  • Wheatstone SLIO/LIO — Up to 192 GPIO pins with edge detection for precise state-change triggering.
  • Allen & Heath AHM — Preset monitoring to trigger actions based on mixer state changes.

Smart Switching Features

VVD supports up to 128 channels (4, 8, 64, or 128 depending on your licence tier), each with 8 trigger slots that all fire simultaneously with optional per-slot delays.

Switching Modes

  • Overview Mode — Automatically cuts to a wide shot when multiple people are speaking at once.
  • Silence Detection — Switches to a designated fallback camera when no one is speaking.
  • Realism Simulator — Cuts to reaction shots of recent speakers, adding visual variety and a more natural broadcast feel.
  • Fixed Timer — Periodic triggers for establishing shots or cutaways at timed intervals.
  • Minimum Duration — Prevents rapid switching by enforcing a minimum time between cuts.

Audio Processing

  • Attack/Decay Controls — Smooth transitions with configurable envelope shaping.
  • Per-Channel Gain — Adjust input levels independently for each channel.
  • Gate — Noise gate to suppress low-level background noise.
  • High-Pass Filter (120 Hz) — Remove rumble and low-frequency interference.

Production Management

  • Broadcast States — On Air, Off Air, Music Break, and Ad Break modes with per-state trigger behaviour.
  • Scene/Show Management — Save and recall complete configurations for quick switching during live production.

API & External Control

VVD exposes three network APIs for external control and integration with third-party systems.

  • HTTP REST API (port 8088) — V1 and V2 endpoints with CORS enabled. JSON request and response format. Query channel state, mute/unmute channels, fire triggers, and read configuration.
  • TCP API (port 8089) — Persistent connections for real-time bidirectional control. Ideal for applications that need to maintain a continuous link.
  • UDP API (port 8090) — Fire-and-forget, low-latency commands. Best for time-critical trigger firing where connection overhead is unacceptable.

Available in 15 Languages

VVD’s interface is fully localised in 15 languages: English, Danish, German, Spanish, French, Italian, Japanese, Korean, Dutch, Norwegian, Polish, Portuguese (Brazil), Russian, Swedish, and Chinese (Simplified).

System Requirements

  • Operating System: Windows 10 or Windows 11 (64-bit)
  • Runtime: .NET 8.0 Runtime
  • CPU: Multi-core recommended (4+ cores for 40+ channels)
  • GPU: Optional — NVIDIA, AMD, or Intel GPU for DirectML acceleration of AI voice detection
  • RAM: 4 GB or more recommended
  • Audio Input: WASAPI, ASIO, or NDI-capable audio source

Ready to Automate Your Production?

Download VVD for free and start switching cameras automatically in minutes. Upgrade to Standard, Professional, or Broadcast for more channels and integrations.