Also known as: dibit
A dibit is a pair of bits represented by one transmitted symbol.1 In the
four-level modulations used by P25 (C4FM)
and DMR (4FSK), each of the four
symbol states maps to one of the four dibits 00, 01, 10, 11. The word is simply the
two-bit generalisation of “bit,” the way “byte” names eight — a compact name for the unit of
data one four-level symbol conveys.
How it works
Four distinct symbol states can be told apart at the receiver, and log₂4 = 2, so each state
labels a two-bit pattern. Which dibit maps to which level is not arbitrary: the standards use a
Gray-coded mapping, in which the dibits assigned to physically adjacent levels differ in only
one bit. In P25 C4FM the four frequency deviations +1800, +600, −600, −1800 Hz carry the dibits
01, 00, 10, 11 respectively — read down the levels and only a single bit changes between
neighbours. The payoff is error resilience: the most likely symbol error is a slip to an adjacent
level, and with Gray coding that slip corrupts just one of the two bits instead of both, roughly
halving the resulting bit-error rate.
The direct consequence for throughput is that a four-level signal’s bit rate is twice its symbol rate — the general bits-per-symbol relationship with two bits per symbol. That is why P25 Phase 1’s 4800-baud C4FM runs at 9600 bits per second, and DMR’s 4800-baud 4FSK likewise.
In practice
Grouping the bitstream two at a time is exactly what a four-level modulator does on transmit and what the demodulator undoes on receive: after symbol-timing recovery the slicer classifies each sample into one of the four levels, then a lookup turns that level back into its dibit, and the dibits are concatenated to rebuild the bitstream for framing and error correction. The two-bit alphabet is why P25/DMR frame structures, sync words, and interleaver spans are often described and counted in symbols (dibits) rather than raw bits. The same idea scales: a tribit is three bits per symbol (eight levels, as in 8-FSK or 8PSK) and a quadbit four bits (sixteen states, as in 16-QAM).
Relevance to SDR
For a scanner, “dibit” is the natural unit at the boundary between DSP and protocol logic: the demodulator’s job ends when it has classified each symbol into a dibit, and the framer’s job begins there. A constellation or eye diagram of a healthy four-level signal shows the four cleanly separated levels the slicer relies on. GopherTrunk’s C4FM/4FSK decode chain produces exactly this dibit stream from the demodulated 4800-baud symbols before handing it to P25 and DMR framing, where sync detection and forward error correction operate.