We investigate experimental limitations in the accuracy of Fourier-transfor
m spectral interferometry, a widely used technique for determining the spec
tral phase difference between two light beams consisting of, for example, f
emtosecond light pulses. We demonstrate that the spectrometer's finite spec
tral resolution, pixel aliasing, and frequency-interpolation error can play
an important role, and we provide a new and more accurate recipe for recov
ering the spectral phase from the experimental data. (C) 2000 Optical Socie
ty of America.