Morse Decoder AI vs PhosphorCW

The Morse Decoder AI is an application for decoding Morse code signals using artificial intelligence. The application is designed for use by amateur radio enthusiasts for educational purposes. The Morse decoder listens to the audio stream through the microphone or line input and, upon detecting Morse code signals, decodes them into text. The neural network of the application is trained to decode signals with a speed ranging from 10 to 40 words per minute within a frequency range of 200 Hz to 900 Hz. The application supports two modes of operation: direct (default) and tone filtering mode. In direct mode, the neural network will attempt to decipher Morse code signals in the audio range of 250 Hz to 900 Hz. This mode is suitable for confident reception of Morse code at levels of 7db-9db on the S-meter. The tone filtering mode is ideal for decoding Morse code from noisy weak signals in the presence of radio interference. The audio input signal is first filtered using band-pass filters before being passed to the neural network for decoding. This mode allows for the decoding of faint signals, but it requires the accurate specification of the signal's tone frequency. Each radio amateur selects their own CW tone frequency in the transceiver settings, and it is important to tune precisely to the carrier frequency using the ZIN/SPOT button in YAESU transceivers, or the AUTOTUNE button in ICOM transceivers. There are 3 band-pass filter options available: 25Hz, 50Hz, 150Hz. If you can accurately determine the CW signal tone frequency, using the 25Hz filter, you can decode very faint Morse code signals. The application offers two neural network options: A and B, which can be easily switched in the interface. Network A is recommended for use with stable signal transmission with a constant duration of dots and dashes, while network B is recommended when using a straight key, where the duration of dots and dashes may vary. You have the ability to switch between these networks in real time and observe how each network hears and decodes Morse code. It is important to monitor the level of the incoming audio signal, for which the application provides a sound level indicator. Ensure that the signal is not too quiet or too loud. It is recommended to maintain the signal around -7db, which is sufficient for decoding. Keep in mind that higher audio frequencies are quieter than lower frequencies. Additionally, the application provides various color themes, allowing radio enthusiasts to customize the appearance of the application for comfortable use.

PhosphorCW — Morse Code Decoder Decode CW (Morse code) signals in real time using your iPhone or iPad's microphone. PhosphorCW listens to radio audio and prints decoded text on screen with the warm glow of a vintage phosphor terminal. REAL-TIME DECODING Point your device at a radio speaker or feed audio via a line-in adapter. PhosphorCW's DSP engine continuously processes the audio stream, detecting CW tones and decoding them into readable text as they arrive. Characters appear on a teleprinter-style display that scrolls automatically, just like the machines that once clattered in radio rooms and newsrooms worldwide. AUTOMATIC SIGNAL TRACKING • Auto Frequency (AFC) — Automatically finds and locks onto the CW sidetone using FFT peak detection. No need to know the exact tone frequency. The tracker adapts as conditions change. • Auto Speed — Measures dit and dah timing to determine sending speed in words per minute. Adapts continuously from around 5 to 60 WPM as the operator's speed varies. Both can be switched to manual mode for fixed-frequency or fixed-speed operation. ON-DEVICE DSP ENGINE All signal processing happens locally on your device. Nothing is recorded, transmitted, or stored externally. The decoding pipeline includes: Bandpass filtering centred on the detected tone frequency Goertzel-based tone detection with adaptive noise floor tracking Envelope extraction with slow-decay peak tracking for reliable squelch Adaptive timing estimation for dit/dah discrimination and character/word gap detection Full ITU Morse code table including prosigns SIGNAL ANALYSIS A dedicated signal view provides: Live FFT spectrum display showing the audio frequency content Scrolling waterfall display for visualising signal activity over time Tone level indicator Real-time WPM, tone frequency, and SNR readout Lock status indicator showing when AFC has acquired a signal DISPLAY Choose from multiple phosphor colour themes — green, amber, white, and blue — inspired by the CRT terminals of the Cold War era. Adjustable decoder font size. Text selection enabled for copying decoded output. SESSION LOG Each decoding session is automatically logged with timestamp, duration, character count, WPM, and tone frequency. Review past sessions and their decoded text at any time. CONFIGURATION Input gain control (0.1x to 20x) Auto/manual frequency detection (300–1500 Hz range) Auto/manual speed tracking (5–60 WPM) Colour theme selection Font size adjustment USE CASES Monitor amateur radio CW transmissions Practice copying Morse code from on-air signals Decode CW practice recordings and training files Analyse CW signal characteristics Learn Morse code by watching the decoder work ABOUT CW Continuous Wave — the original digital mode. Samuel Morse's code has been the universal language of radio since the 1840s. From clipper ships to Cold War submarines, from Bletchley Park to amateur radio contests, CW remains the most efficient and reliable communications mode ever devised. This app decodes it from your microphone or line input. Requires microphone access for audio input. iOS 18.0 or later.