How this instrument works
Baud rate counts symbols, not bits. A symbol is one discrete state the communication line can take on — a specific voltage level, phase, or frequency — and it simply counts how many of those state changes happen each second. Bit rate, by contrast, counts bits. The two are equal only when each symbol carries exactly one bit, which is why the terms get used interchangeably in casual conversation about simple binary links, even though they measure genuinely different things.
The relationship is baud = bit rate / bits per symbol. When a modulation scheme packs more than one bit into each symbol — QPSK packs 2, 8-level PAM packs 3, 256-QAM packs 8 — the baud figure drops below the bps figure even as that bps count stays the same or grows, because fewer such changes are needed to carry the same volume of bits. This is exactly how modem technology increased throughput over the 20th century without requiring proportionally faster line signaling: by encoding more bits into each symbol rather than raising the raw signaling speed.
Baud rate is named after Émile Baudot, a pioneer of early telegraph coding, and it remains the figure that actually describes a channel's raw signaling speed — the limit set by the physical medium and modulation hardware. A channel's bps figure can be pushed higher than its baud figure by using denser modulation, but that baud figure itself is bounded by the channel's bandwidth and noise characteristics in ways the bps number alone doesn't reveal.
- Enter Bit rate (bits per second) — the data rate you already know, for example 9600 for a classic 9600 bps serial link.
- Enter Bits per symbol — how many bits each individual symbol encodes; use 1 for simple binary (on/off) signaling.
- Read Baud rate (symbols per second) directly — this is the line's actual signaling rate, distinct from its bit rate whenever bits per symbol is greater than 1.
- Bit rate must be zero or greater, and bits per symbol must be greater than zero — the instrument will prompt you if either falls outside that range.
Worked example — 9600 bps at 1 bit per symbol
Enter Bit rate (bits per second) = 9600 and Bits per symbol = 1 — simple binary signaling, the kind a classic RS-232 serial connection or an early modem used. Dividing gives 9600 / 1 = 9600. Baud rate reads 9600.
Because each symbol here carries exactly one bit, baud and bps come out identical — 9600 baud and 9600 bps describe the same link. That equality only holds at 1 bit per symbol; a modem using a denser modulation scheme to reach the same 9600 bps of throughput would show a lower baud figure, since fewer such changes per second would be needed to carry the same bits.
Questions
Is baud rate the same thing as bit rate?
Only when each symbol carries exactly one bit. Baud rate counts state changes per second; bit rate counts bits per second. The two numbers are identical for simple binary signaling, but the moment a modulation scheme packs more than one bit into each symbol — as QPSK, PAM, or QAM do — bps rises above baud for the same signaling speed.
Why would a modem have a lower baud rate than its bit rate?
Because it's using a modulation scheme that encodes multiple bits per symbol. A classic example is a 2400 bps modem using QPSK (2 bits per symbol), which only needs 2400 / 2 = 1200 baud of actual line signaling to deliver that 2400 bps of throughput — denser encoding lets bps outpace baud without demanding faster raw signaling.
Why is baud rate named after Émile Baudot?
Baudot developed an early telegraph code and multiplexing system in the 1870s, and 'baud' was adopted as the unit for symbol speed in his honor. The unit predates modern digital modems by decades, but the underlying concept — counting discrete signal-state changes per second, independent of how many bits each one encodes — is exactly what today's baud rate measures.
What does 'bits per symbol' actually mean physically?
It's how many distinct states a single symbol can take, expressed as bits: 1 bit per symbol means 2 possible states (on/off), 2 bits per symbol means 4 possible states, and so on, since n bits per symbol allows 2^n distinguishable symbol states. A modulation scheme with more possible states per symbol packs more bits into each such change, raising bps without raising baud.
Why can't bits per symbol be zero or negative?
A symbol has to encode at least some information to be meaningful — zero bits per symbol would mean each such change carries no information at all, and dividing by zero to find the baud rate is undefined regardless. Negative bits per symbol has no physical meaning, since a symbol's state count (2 raised to that power) can't be less than one distinguishable state.