Single-sideband modulation (SSB) carries information in one sideband of an amplitude-modulated signal. Radio voice operation normally uses a suppressed carrier. USB retains the upper sideband; LSB retains the lower sideband.
For a suppressed-carrier reference frequency f_ref and a positive audio frequency f_audio:
USB: f_RF = f_ref + f_audio LSB: f_RF = f_ref - f_audio
A 1500 Hz tone with a 27.265 MHz USB reference appears at 27.2665 MHz. In LSB it appears at 27.2635 MHz. The display reference is not necessarily an emitted carrier: carrier suppression is part of normal SSB operation.
Audio from 300 to 2700 Hz occupies about 2400 Hz of RF spectrum in an ideal sideband. The practical occupied bandwidth also depends on filtering, distortion and the chosen measurement criterion.
A transmitter can create SSB using filtering, phase-shift methods or digital signal processing. A receiver restores a carrier reference in a product detector. A tuning error shifts the recovered audio frequencies, which explains the unnatural pitch of a mistuned voice signal.
Suppressing the carrier and one sideband avoids spending transmitted power on those components. For speech, the RF envelope still changes continuously with the audio. Peak envelope power and average power therefore differ.
Many digital modes use audio tones passed through an SSB transmitter. The radio converts these tones to RF; it does not turn every tone-based mode into FM. The selected sideband, audio frequency and dial reference together determine the actual RF signal.
Excessive audio drive, clipping or unsuitable speech processing can widen the signal and impair decoding. Station settings must suit the waveform and the transmitter's continuous-duty capabilities.