Abstract

Calculation of the average of repeated scans of spectral distributions shows that the root-mean-square noise cumulates as the square root of the number N of the scans only when the correlation time of the noise is less than repetition time, whereas it is independent of N when the correlation time is longer than the total measuring time. We also examine an optimum linear filter that takes into account the additive noise and the multiplicative noise, which can be generated by flux-density fluctuations during the scanning. We discuss the effect of the noise and filtering on the restoration of the spectra, in the case of a synthesis spectrometer, and we find the conditions for which the synthesis spectrometer is advantageous.

© 1974 Optical Society of America

Maximum-likelihood restoration of nonstationary imagery

Martin J. Lahart
J. Opt. Soc. Am. 64(1) 17-22 (1974)

Optimizing precision of rotating-analyzer ellipsometers

D. E. Aspnes
J. Opt. Soc. Am. 64(5) 639-646 (1974)

Optical pulses in atmospheric turbulence

C. S. Gardner and M. A. Plonus
J. Opt. Soc. Am. 64(1) 68-77 (1974)

Internal thermal noise in optical frequency converters

Lee E. Estes, Robert F. Lucy, John Gunter, and Kenneth Duval
J. Opt. Soc. Am. 64(3) 295-300 (1974)

Dichroic microspectrophotometer: A computer-assisted, rapid, wavelength-scanning photometer for measuring linear dichroism in single cells*

Ferenc I. Hárosi and Edward F. MacNichol
J. Opt. Soc. Am. 64(7) 903-918 (1974)