Standard time and frequency signals (STFS) for HF broadcasts (e.g. ATA, JJY, WWV etc.) usually consist of an audio tone, a burst of an audio signal every second, a voice announcement at some regular interval of time, etc. The authors describe a generator of such a signal, which is based on microprocessor technology. In this system an audio sinusoidal signal is produced through a digital technique which facilitates the maintenance of the phase relation of the zero crossing of the audio sine wave with respect to that of the second pulse from a standard clock. The burst width is basically controlled by a microprocessor and the basic clock is also run by software. The voice announcement is effected by playing a pre-recorded audio tape. The new scheme, with a dual track-recording system combined with microprocessor controlled management, eliminates the possibility of an incorrect voice announcement caused by tape elongation. The authors describe the functioning of the format generator and discuss its salient features. The generation of the ATA format has been taken as a typical case without loss of generality.