Selective transmission of prerecorded voice signals to subscribers of

Abstract

A dynamic memory UR containing n different message words, carried on as many endless signal tracks, works into a set of m temporal voice channels through a TDM switching matrix RC enabling any channel to be periodically connected, in its recurrent time slot, to a selected signal track. Messages synthesized from up to k words may provide information on the charge of a toll call, for example, as determined in a processor of a central office, the processor translating that information into data codes identifying corresponding signal tracks in response to a request from an inquiring subscriber. A first static memory M.sub.1, with m stages individually allocated to the several PCM channels, thereupon receives the data codes in respective sections of the stage associated with the inquiring subscriber and emits a start signal in the corresponding time slot informing a logic network LC of the presence of data stored therein, the logic network then monitoring selected outputs of the dynamic memory UR for the appearance of synchronizing signals prefacing their prerecorded message words. Coincidence of the first synchronizing signal with the start signal triggers the logic network LC into a sequence of operations in which a second static memory M.sub.2, connected in a loop circuit with that network, receives and re-emits bit combinations of progressively increasing numerical value identifying successive words of a message to be read out from the selected signal tracks to the inquiring subscriber. The readout of the message, ending after k words or upon the appearance of a terminal code from the processor entered in one or more sections of the allocated stage of memory M.sub.1, may be repeated at least once. A counter CO responsive to the start signal arrests the dynamic memory UR whenever no such signal appears within a scanning cycle of m time slots.