The performance of a Time Assignement Speech Interpolation (TASI) syst
em is usually characterized by the fraction of the freeze-out talkspur
ts of talkers. In the present investigation, we analyze the process of
information transmission in a TASI system in details and introduce a
new performance characteristic of such systems. Explicit analytical ex
pressions are obtained under the assumption that input subscribers are
continuously occupied, talkspurts and silence intervals are exponenti
ally distributed.