Amateur radio has not one digital voice mode but a whole family of them, each born in a different decade for a different reason — and each one convinced it does something the others don't. This is a brief, plain-language tour of all seven: where each came from, what codec it speaks, its signature trait, and where you'll hear it today. Every mode gets one simple diagram showing the thing that makes it itself.
D-Star is where amateur digital voice began. Developed in Japan around the turn of the millennium as an open standard by JARL (the Japanese national amateur radio society) and brought to market almost entirely by Icom, it predates everything else on this page. It uses the original AMBE codec, and its voice channel carries your callsign in every transmission — which enables its signature trait: callsign routing. A D-Star gateway can route your call to a specific station or, far more commonly today, link your repeater or hotspot to a reflector (REF, XRF, DCS, and the modern XLX) where many systems meet in one conversation.
You'll hear D-Star on Icom radios (and a handful of others), on thousands of gateways worldwide, and across the reflector ecosystem. It can sound slightly more "robotic" than newer modes — the price of being first — but its routing-and-reflector architecture set the template every later mode borrowed.
DMR (Digital Mobile Radio) is a European commercial standard, published by ETSI in the mid-2000s for business and professional radio. Hams adopted it for a simple reason: the radios are everywhere and inexpensive, from surplus commercial gear to a flood of budget handhelds. It speaks AMBE+2, and its signature trait is TDMA: two timeslots alternating on one frequency, so a single repeater carries two independent conversations at once.
DMR is today's most popular amateur digital voice mode, with worldwide networks — BrandMeister, TGIF, FreeDMR, and others — organized around talkgroups. (For the full story, see How a DMR Network Works.)
System Fusion is Yaesu's digital mode, launched in 2013, transmitting C4FM (four-level FSK) and speaking AMBE+2 like DMR. Its signature trait is AMS — Automatic Mode Select: a Fusion repeater or radio can receive analog FM and digital C4FM on the same channel and automatically answer in kind. That made it the gentlest upgrade path ever offered — a club could replace its analog repeater with a Fusion machine and the analog old-timers never lost access.
Fusion's networking comes in two flavors: Yaesu's own Wires-X system, and the open YSF reflector world that the hotspot community built around the protocol. DG-IDs (digital group IDs) act as a light routing layer for selecting destinations through a single gateway.
APCO Project 25 is the North American public-safety digital standard — the mode of police, fire, and EMS systems across the continent. The widely used Phase 1 speaks the IMBE codec on FDMA: one 12.5 kHz channel, one conversation, no timeslots. Hams run P25 almost entirely on surplus commercial gear — used Motorola and other public-safety radios that flow onto the used market as agencies upgrade — which gives the mode its particular flavor: rugged, professional hardware and a strong crossover with the scanner-listening world.
Amateur P25 lives on conventional (non-trunked) repeaters and hotspots linked to P25 reflectors. It's a smaller community than DMR or Fusion, but a devoted one — for many, the appeal is operating the same equipment the professionals carry.
NXDN is a commercial standard developed jointly by Icom (as IDAS) and Kenwood (as NEXEDGE) in the mid-2000s, speaking AMBE+2. Its signature trait is bandwidth: NXDN's narrow mode fits a voice channel into just 6.25 kHz — a quarter of a classic 25 kHz analog channel, and half of everything else on this page. In the commercial world that meant more channels in the same spectrum; in the amateur world it makes NXDN the quiet efficiency champion.
Like P25, amateur NXDN runs mostly on surplus commercial radios, with a smaller but loyal community connected through NXDN reflectors. Instead of color codes or DG-IDs, NXDN uses a RAN (Radio Access Number) for repeater access.
M17 is the youngest mode here and the only one designed by hams, for hams, from scratch — an open-source project started around 2020 by Wojciech Kaczmarski (SP5WWP) and a growing community. Its signature trait is openness at every layer: the protocol specification is public, the reference implementations are open source, and — the part that matters most — it uses Codec 2, a free and open vocoder, instead of the patented AMBE family every other mode on this page depends on. No license fees, no proprietary chips, nothing you can't study or improve.
M17 networking runs through mrefd reflectors (modules A–Z, much like the XLX world), and the hardware ecosystem is growing: MMDVM hotspots speak it, open firmware projects bring it to existing radios, and purpose-built open hardware is appearing. It's the smallest community on this page and the one moving the fastest.
FreeDV is the outlier on this page, and deliberately so: it's open-source digital voice for the HF bands, created by David Rowe (VK5DGR) and a small team of radio amateurs. There's no special radio and no network — FreeDV is software. Your SSB transceiver's audio runs through a computer (or a small dedicated device) that does the encoding and decoding, using Codec 2 — which was born here; M17 adopted it later. Its modes are engineered for the realities of shortwave: narrow bandwidth, fading, and noise, with the newest generation even using machine-learning techniques to stay intelligible in conditions where SSB gets hard to copy.
The signature trait is what's missing: no repeaters, no reflectors, no internet. FreeDV is point-to-point ionosphere work — digital voice the way SSB is voice. Two stations, one band, and physics in between.
| Mode | Origin | Codec | Signature trait |
|---|---|---|---|
| D-Star | JARL / Icom, ~2000 | AMBE | Callsign routing; the original reflector ecosystem |
| DMR | ETSI commercial standard, mid-2000s | AMBE+2 | Two timeslots on one frequency; huge networks, cheap radios |
| C4FM / Fusion | Yaesu, 2013 | AMBE+2 | AMS: analog and digital share one channel |
| P25 | APCO / public safety, 1990s | IMBE (Phase 1) | The professionals' mode, on surplus pro gear |
| NXDN | Icom + Kenwood, mid-2000s | AMBE+2 | Ultra-narrow 6.25 kHz channels |
| M17 | Ham community, ~2020 | Codec 2 (open) | Open source at every layer |
| FreeDV | Ham community (VK5DGR et al.) | Codec 2 (open) | HF point-to-point; no infrastructure at all |
Why the codec column matters
Look down that column and the whole transcoding story falls out of it. DMR, Fusion, and NXDN all speak AMBE+2 — same language, so they can share audio freely on a multi-protocol reflector. D-Star speaks the older AMBE and P25 speaks IMBE — different languages, so bridging them to anything else requires a transcoder. M17 and FreeDV speak open Codec 2 — free of patents, but still a different language needing translation to reach the AMBE world. When you read about reflectors needing vocoder hardware to be "fully transcoded," this column is the reason. (See Understanding XLX Reflectors and URF Reflectors for that story.)
A noncommercial hobby reference compiled by N6JET, gathered from public sources and shared freely for anyone interested in amateur digital voice.