Research and development of high-frequency (HF) communication systems is gaining renewed vigour despite advances in satellite communication systems and undersea optical fibre. Civilians primarily use the HF band for communications. Government organisations monitor the HF spectrum to control and enforce licensing, to respond to distress calls, and for communications. Defence agencies often play a role in these latter applications, but also have the task of spectrum surveillance for subversive communications. Modern HF systems for performing these activities require capabilities for locating the sources of transmissions, separating valid signals from interference and noise, and recognising signal modulation. The goal of this investigation is to develop robust algorithms for recognising the modulation of real HF signals, that is, signals propagating via multiple ionospheric modes with co-channel signals and non- Gaussian noise. It specifically addresses a new application of the entropic distance parameter to HF signals and demonstrates that entropic distance can be used to separate real HF signals.