Cryptochords Framework: Multi-Layered Symbolic Audio Steganography for Covert Communications

Abstract

Conventional audio steganography methods that operate in the physical domain (LSB modification, phase encoding, or echo hiding) are fragile under lossy compression (MP3, AAC, Ogg Vorbis) and acoustic air-gap re-recording. This paper presents the formal specification of Cryptochords, a symbolic audio steganography framework that embeds encrypted payloads into higher-level musical abstractions—harmonic voicing matrices and micro-timing offsets—rather than the raw waveform. By targeting pitch-class profiles in harmonically dense genres (Bossa Nova, Jazz Fusion, Shibuya-kei), the payload remains recoverable after multiple transcoding cycles and analog re-recording. We define a 16-symbol harmonic codebook (4 bits/chord) constrained by a first-order Markov model trained on genre corpora, a micro-timing channel with δ = 15 ms modulation depth, and a dual-layer Reed-Solomon code over GF(28 ). Realistic capacity estimates range from approximately 8 to 22 bytes per minute depending on tempo and harmonic density, with zero-knowledge properties for performing musicians. The framework is intended as a low-bandwidth, high-deniability covert channel for strategic signaling rather than bulk data transfer.