D-STAR is a digital voice and data system for amateur radio, developed in Japan and built mainly into Icom radios, with some support from other makers. Your voice is turned into bits, sent with a header that names the stations involved, and can be routed across the world by call sign.
How digital voice works
An ordinary FM radio sends the sound wave itself. A digital voice radio instead uses a vocoder to analyse speech and describe it with a small stream of numbers, then sends only that. The receiver rebuilds a voice from them. Error-correcting bits are added so a few damaged bits can be repaired.
The vocoder is the heart of digital voice. Different systems use different vocoders and different ways of sending the bits.
The result is clean, noise-free audio until the signal gets too weak, then it turns robotic and drops out rather than fading into hiss. Vocoders are tuned for speech, so music and strong background noise come out poorly. See Digital modulation: FSK, PSK and QAM and Error detection and correction.
Routing by call sign
Every D-STAR transmission starts with a header. Four fields matter to the operator:
The radio fills these in from its menu. Routing is done by call sign, not by frequency.
D-STAR repeaters connect to a gateway that is tied to a network. A station registers its call sign once; the network then notes where it was last heard and routes calls to it. Reflectors are shared rooms: many repeaters and hotspots link to one, and everyone connected hears everyone else, much like a conference call.
In practice
Program your call sign into the radio before transmitting. It is the one thing it cannot work without.
Use a repeater memory (frequency, offset, name) as you would for FM, then set UR as needed.
Joining a reflector usually means sending a short, specific command through the UR field.
A hotspot gives you a personal, low-power gateway at home; see Hotspots and gateways.
Analog FM and D-STAR radios cannot hear each other. A D-STAR signal sounds like a buzz to an FM receiver.