CW Companion

CW Companion is a macOS application designed for Morse Code (CW) enthusiasts. It provides tools for both receiving (decoding) and transmitting (encoding) Morse code, featuring real-time system audio capture and advanced signal processing.

Key Features

1. Morse Code Decoder (Receive)

• Live Audio Capture: Utilizes ScreenCaptureKit to capture high-quality audio directly from specific active windows on your Mac. This allows you to decode CW from SDR software, web-based radios, or videos without using a microphone.
• Signal Processing Pipeline:
  • Bandpass Filter: Implements a custom Digital Biquad Filter centered at 600Hz (Q=5.0). This pre-processing step attenuates background noise and isolates the CW signal for cleaner detection.
  • Envelope Follower: Converts the filtered AC audio signal into a DC amplitude envelope with fast attack and calculated decay, enabling robust signal detection even with varying amplitudes.
  • Adaptive Thresholding: Dynamically detects Signal ON/OFF states based on audio levels.
• Algorithmic Decoding:
  • Uses a time-based state machine to interpret signal duration.
  • Automatically distinguishes between Dots, Dashes, Intra-character gaps, Inter-character gaps, and Word spaces.
  • Adaptive WPM: Capable of estimating the sender's speed (Words Per Minute) to adjust timing tolerances.
• File Input: Supports loading and decoding existing .wav audio files.

2. Morse Code Encoder (Transmit)

• Text-to-CW: Converts typed text into standard International Morse Code timing.
• Audio Synthesis: Generates clean 600Hz sine wave audio.
• Envelope Shaping: Applies a 5ms linear rise and fall time (ramp) to every key-down event. This essentially eliminates "key clicks" (spectral splatter) common in simple square-wave generators.
• WAV Export: Saves the generated Morse code audio to a broadcast-quality .wav file for sharing or practice.

3. Cloud Receiver (FT8 & Waterfall) 📡

• Public SDR Integration: Connects to public KiwiSDR receivers (e.g., KPH San Francisco) to monitor radio bands without local hardware.

• Locally Generated Waterfall:

  • Metal-Accelerated: Uses Accelerate (vDSP) for high-performance FFT and Metal for GPU rendering, creating a smooth 60fps spectrogram.
  • Passband Zoom: Automatically focuses on the active 0-3kHz audio band.
  • Calibrated Visualization: High-contrast color mapping for easy signal spotting.

• FT8 Decoder:

  • Live Decoding: Integrated C-based FT8 decoding engine processes 15-second transmission cycles in real-time.
  • Grid Extraction: Automatically parses Maidenhead Grid Locators (e.g., "PL02") from decoded messages.

• Global Map Plotter:

  • Live Visualization: Plots decoded stations on an interactive Apple Map using the extracted grid squares.
  • Split-View Dashboard: Professional monitoring layout with Map on the left and Waterfall/Data on the right.

  

Technical Architecture

• Platform: macOS 12.3+ (Required for ScreenCaptureKit).
• Language: Swift 5.
• UI Framework: SwiftUI.
• Audio Engine: AVFoundation for low-level audio buffer manipulation/playback and ScreenCaptureKit for system audio streams.
• DSP Engine: Apple Accelerate (vDSP) for FFT calculations.
• Rendering: Apple Metal for high-performance waterfall visualization.
• Concurrency:
  • Uses Swift structured concurrency (async/await) for managing live stream sessions.
  • Utilizes DispatchQueue and Combine pipelines to offload heavy DSP tasks.

Usage

Receive Mode

1. Go to the Receive tab.
2. Switch the input mode to Live Window.
3. Select the target application (e.g., your SDR program or web browser) from the list.
4. The app will immediately start capturing audio, filtering for 600Hz tones, and streaming decoded text to the window.

Transmit Mode

1. Go to the Transmit tab.
2. Enter your text message.
3. Adjust the WPM (Speed) and Frequency if desired.
4. Press Play to preview or Save WAV to export the audio file.

Cloud Receiver Mode

1. Go to the Cloud tab.
2. Select a receiver from the dropdown (e.g., "KPH San Francisco").
3. Click Connect. The waterfall will begin scrolling.
4. Wait for the 15-second FT8 cycle.
5. Decoded messages will appear in the list, and pins will drop on the map corresponding to the transmitter's location.

Development

• AudioProcessing.swift: core logic for DSP, including the BiquadFilter, AudioModel state management, and the processLiveBuffer loop.
• AudioCaptureManager.swift: Handles the ScreenCaptureKit stream lifecycle and permission handling.
• MorseDecoder.swift: Logic for translating time intervals into characters.
• FT8Engine.swift: Wrapper around the C-based FT8 library (ft8_lib).
• WaterfallRenderer.swift: Metal-based renderer for the audio spectrogram.
• MaidenheadLocator.swift: Utility to convert Grid Squares to Lat/Long coordinates.