We describe a simple variation in the analysis of the transient photocurrent i(t) in semiconductors to determine localized state distributions g(E), which affords a remarkable improvement in the energy resolution available, without special computing requirements. The initial stage of the analysis uses numerical Fourier or Laplace transformation of i(t), and this results in an “ill-posed” integral equation for g(E), which has been hitherto approached by approximate low resolution techniques, or by specialized numerical methods. Our proposal allows solution of this implicit equation for a fine discrete version of g(E) using a simple least squares fitting algorithm with an exact model function. The technique is demonstrated by application to i(t) data, computer generated for several representative g(E) distributions. © 2000 American Institute of Physics.