Invented by Australian company IRT in the 1980s, the basic concept of VIMCAS is to transmit two channels of PCM-encoded (i.e. digital) audio during the vertical blanking interval of a composite video signal.
[4][2] To fit into the available bandwidth, the audio signal would first be companded and limited before being sampled for PCM encoding.
[1][3] The encoded signal would be transmitted in the six scanlines in time compressed form, i.e. much faster than its actual speed.
[1] Decoding was simply the reverse process, with 100ms of audio (at a time) stored in the transmitted digital form into a digital memory and played out from that memory at original speed through a digital-to-analogue converter, with appropriate timing circuits to synchronize this playout with the accompanying video.
[1][3] A reduced version, using just one scan line instead of six and thus providing narrower bandwidth, was called VISCAS (Vertical Interval Single Channel Audio System), which was good enough for talkback between the studio and the OB or foldback.