The time until a media decoder is refreshed to produce correct output samples is minimized if an IDR picture is the first picture in a decoding order in each MPE-FEC frame. A system enabling minimization of tune-in delay is described. Aspects of the system include a content encoder for receiving a source signal including first and second bit streams and for encoding the first and second bit streams, the encoded first bit stream including a decoder refresh bit stream and the encoded second bit stream including a media bit stream. A server receives the encoded first and second bit streams and encapsulates the encoded first and second bit streams into packets. An encapsulator then receives the packets and composes a time-sliced multi-protocol encapsultation (MPE) stream by replacing a spliceable picture from the encoded second bit stream.