Presentation Recorder

Overview

Presentation Recorder is a native macOS application built with SwiftUI and Apple’s ScreenCaptureKit framework. It captures screen content, system audio, microphone input, and camera footage simultaneously — producing timestamped MP4 recordings ideal for presentations, tutorials, and demos. The app features a live preview, real-time audio level metering, and configurable video quality presets ranging from 15 fps to 120 fps.

Features

• Simultaneous Screen + Camera Recording — Captures both a screen or window and the webcam feed at the same time, outputting separate timestamped MP4 files for flexible post-production
• Configurable Video Quality — Five presets (Low through Extreme) controlling frame rate (15–120 fps) and pixel scale (1x–3x), with HDR support via SCStreamConfiguration.Preset
• Real-Time Audio Metering — Uses Apple’s Accelerate framework (vDSP_rmsqv, vDSP_maxv) to compute RMS and peak power levels per channel, displayed as a live level indicator in the UI
• Microphone + System Audio Capture — Independent toggles for microphone and application audio, with the option to exclude the app’s own audio from the recording
• Content Sharing Picker — Integrates SCContentSharingPicker for selecting specific displays, windows, or applications to capture — including multi-window and multi-app modes
• Live Capture Preview — Renders incoming IOSurface frames via a CALayer-backed NSView with proper aspect ratio handling
• Transparent Window UI — The app window uses NSVisualEffectView with .ultraThinMaterial blending, giving it a modern translucent macOS look

How It Works

The architecture follows a clear separation: ScreenRecorder orchestrates the capture session, CaptureEngine wraps SCStream and yields frames via Swift’s AsyncThrowingStream, and CameraRecorder manages the AVCaptureSession independently.

The capture pipeline processes video and audio on dedicated dispatch queues:

private let videoSampleBufferQueue = DispatchQueue(label: "com.example.apple-samplecode.VideoSampleBufferQueue")
private let audioSampleBufferQueue = DispatchQueue(label: "com.example.apple-samplecode.AudioSampleBufferQueue")
private let micSampleBufferQueue   = DispatchQueue(label: "com.example.apple-samplecode.MicSampleBufferQueue")

The PowerMeter class handles multiple PCM buffer formats (float, int16, int32) with Accelerate-optimized DSP, converting raw samples to calibrated dB levels via a MeterTable lookup:

func process(buffer: AVAudioPCMBuffer) {
    // Handles float, int16, and int32 channel data
    // Uses vDSP_rmsqv for average power and vDSP_maxv for peak
}

Testing

The project includes XCTest-based unit tests in PresentationRecorderTests/. The PowerMeterTests suite covers key audio processing scenarios:

• Zero-buffer produces valid zero levels
• Float buffer processing yields correct non-zero average and peak values
• MeterTable correctly clamps power values at boundaries
• processSilence() resets accumulated values

Tech Stack

Platform	macOS (native)
Language	Swift
UI Framework	SwiftUI
Screen Capture	ScreenCaptureKit (SCStream, SCRecordingOutput)
Camera	AVFoundation (AVCaptureSession)
Audio DSP	Accelerate (vDSP)
Persistence	SwiftData
Concurrency	Swift async/await, AsyncThrowingStream
Testing	XCTest

Source Code

The source code is available on the project’s GitHub repository.