Drift 🌊

A high-performance, real-time audio synthesis engine for focus and relaxation.

Drift is a native iOS application that generates Binaural Beats and procedural atmospheric soundscapes in real-time. Unlike typical meditation apps that play pre-recorded MP3s, Drift synthesizes audio samples on the fly using AVAudioSourceNode, ensuring zero looping artifacts and minimal memory footprint.

---

🛠 Engineering Highlights

1. Real-Time Digital Signal Processing (DSP)

• Zero Samples: No audio files are used for the main tones. All sound is mathematically generated.
• C-Interop: Utilizes UnsafeMutablePointer and C-based audio rendering callbacks (AVAudioSourceNodeRenderBlock) for lock-free, high-performance audio generation on the real-time thread.
• Procedural Noise: Implements a Linear Congruential Generator (LCG) to synthesize Brown and White noise procedurally, completely eliminating large audio assets.

2. Modern Swift 6 Concurrency

• Observation Framework: Migrated from Combine/ObservableObject to the new Swift 6 @Observable macro for performant, fine-grained state tracking.
• Thread Safety: strict isolation between the UI (Main Actor) and the high-priority Audio Thread.

3. High-Performance SwiftUI

• Canvas Drawing: The Wave visualization uses immediate-mode Canvas drawing rather than Shape paths for optimal performance.
• Animatable Protocol: Custom conformance to Animatable enables smooth interpolation of strictly mathematical values (Frequency/Amplitude) during UI transitions.
• TimelineView: decouples the animation loop from the update loop, ensuring silky smooth 60/120Hz rendering.

---

🏗 Tech Stack

• Language: Swift 6
• UI: SwiftUI (Canvas, TimelineView, matchedGeometryEffect)
• Audio: AVAudioEngine, AVAudioSourceNode, AVAudioMixerNode
• State Management: Swift Observation framework (@Observable, @Bindable)
• Architecture: MVVM + Singleton Audio Controller
• Tools: Xcode 16, Git

---

🧠 Technical Deep Dive: The Audio Engine

The heart of Drift is the AudioController. Instead of playing files, it attaches a "Tap" to the audio engine using AVAudioSourceNode.

// Simplified logic of the Render Block
sourceNode = AVAudioSourceNode { _, _, frameCount, _ -> OSStatus in
    for frame in 0..<Int(frameCount) {
        // 1. Calculate Sine Wave for Left/Right Ears
        let valL = sin(phaseL)
        let valR = sin(phaseR)
        
        // 2. Advance phases by different increments to create Binaural Beat
        // Left: 200Hz | Right: 206Hz -> Result: 6Hz interference pattern (Theta Wave)
        phaseL += incrementL
        phaseR += incrementR
        
        // 3. Write directly to the output buffer
        bufferL[frame] = valL
        bufferR[frame] = valR
    }
    return noErr
}

This approach allows for instantaneous frequency shifting and "glitch-free" transitions between different brainwave states.

---

📱 Features

• Dynamic Binaural Beats:
  • Delta (2.0 Hz): Deep Sleep
  • Theta (6.0 Hz): Meditation
  • Alpha (10.0 Hz): Relaxation
  • Beta (20.0 Hz): Focus
• Audio Mixer: Multitrack control for synthesis, rain (brown noise), and white noise layers.
• Visual Entrainment: A math-driven sine wave animation that perfectly matches the audio frequency.

---

📜 License

Created by Jonni Akesson. Open for educational use.