Skip to content

Introduction to VoiceEngine

VoiceEngine is a comprehensive 3D positional voice chat and audio streaming system for Minecraft servers (Paper standalone and Velocity networks). Unlike traditional voice chat solutions, it completely eliminates client-side mod requirements (Fabric, Forge, NeoForge).

Players connect to your server using the official vanilla Minecraft Java Edition client, type /voice in chat, and click a secure connection link to open the high-performance web client in any modern desktop or mobile browser. Audio is transported via WebRTC and routed through a mediasoup Selective Forwarding Unit (SFU).


Key Features

3D Spatial Audio & Immersion

  • Binaural 3D Audio: High-fidelity positional audio rendered with the browser’s native Web Audio API and Head-Related Transfer Function (HRTF) algorithms.
  • Atmospheric Air Absorption: Dynamic low-pass frequency rolloff simulating acoustic air dampening (20 kHz down to 3.5 kHz) over physical block distance.
  • Underwater Acoustic Damping: Automatic 600 Hz low-pass biquad filter applied when either interlocutor is submerged in water.
  • Sneak Whispering: Crouching (Shift) dynamically compresses the audible voice radius from 30 blocks down to 8 blocks.
  • Inter-Dimensional Isolation: Strict room partitioning across Overworld, Nether, and The End.
  • Spectator Modes: Configurable spectator behavior (listen-only, isolated, all) tailored for minigames, events, and tournaments.

Client-Side Microphone DSP

  • Neural Noise Suppression (RNNoise): WebAssembly AudioWorklet with SIMD detection eliminating keyboard clicks, computer fans, and room reverb.
  • 80 Hz High-Pass Filter: Removes desk rumbles, breath pops, and sub-bass vibrations.
  • Soft-Knee Dynamic Compressor & Limiter: Automatically controls volume peaks to prevent digital distortion and clipping.
  • Unthrottled Background VAD: Dedicated inline Web Worker timer (40 Hz) driving Voice Activity Detection, completely immune to browser tab throttling when Minecraft is running in fullscreen.

In-World Mechanics & Media Emitters

  • Speech Particle Indicators: Green musical note particles appear above speaking players’ heads in real time.
  • Physical Speaker Blocks: World blocks (such as jukeboxes or note blocks) registered as amplified broadcast speakers with custom radiuses and Redstone signal toggling.
  • Dynamic 3D Audio Emitters: Positional music playback, one-shot sound effects, or 2D global server broadcasts initiated via commands.
  • LRU Media Caching: Automatic downloading and disk caching for remote media streams and audio files.

Network Scalability & Moderation

  • Native Velocity Proxy Support: Single authentication upon joining the network; players stay in active calls seamlessly while switching backend Paper servers.
  • Embedded SQLite Moderation Engine (WAL): Persistent sanctions (kick, mute, deafen, ban) with UUID and IP tracking to prevent ban evasion.
  • Zero-Latency SFU Sync: Staff actions push immediately via WebSocket, muting or disconnecting the browser in milliseconds.
  • Proximity Radar & Picture-in-Picture (PiP): Live circular radar canvas visualizing nearby players and floating HUD overlay pinned over Minecraft with M (Mute) and D (Deafen) hotkeys.

Connection Architecture

┌─────────────────┐ Token /ws/plugin ┌──────────────────────┐
│ Minecraft │ ───────────────────────────> │ Voice Server (SFU) │
│ Paper/Velocity │ <─────────────────────────── │ Node.js + mediasoup │
└─────────────────┘ Sanctions / Telemetry └──────────────────────┘
│ │
/voice Link WebRTC Audio Tracks
│ & WebSocket /ws/client
▼ │
┌────────────────────────────────────────────────────────────▼──┐
│ Player Web Browser (Web Client) │
│ - WebRTC Audio Consumer/Producer │
│ - DSP: RNNoise WASM + High-Pass Filter + Dynamic Compressor │
│ - Web Audio API: PannerNode HRTF 3D + Acoustic Air Rolloff │
└───────────────────────────────────────────────────────────────┘