The source coding theorem for stationary sources describes the optimal perf
ormance theoretically achievable by fixed- and variable-rate block quantize
rs. The source coding theorem may be generalized by considering the problem
of multiresolution or successive refinement source coding, which is the to
pic of this work. Given a distortion vector (D-1, ..., D-L), this work desc
ribes the family of achievable rate vectors (R-1, ..., R-L) for describing
a stationary source at L resolutions, where the description at the first re
solution is given at rate R-1 and achieves an expected distortion no greate
r than D-1, the description at the second resolution includes both the firs
t description and a refining description of rate R-2 and achieves expected
distortion no greater than D-2, and so on. The work includes performance bo
unds for both fixed- and variable-rate source codes on discrete-time statio
nary ergodic sources and discrete-time stationary nonergodic sources for an
y integer number of resolutions L greater than or equal to 1, For L = 1, th
e source coding theorems for stationary sources result, For L > 1, the resu
lts extend previous theorems for discrete-alphabet memoryless sources.