Meaningful use of the autocorrelation in jet noise analysis is examined. The effect of peak frequency on the autocorrelation function width is removed through a temporal scaling prior to making comparisons between measurements or drawing conclusions about source characteristics. In addition, a Hilbert transform-based autocorrelation envelope helps to define consistent characteristic time scales. Application of these processes to correlation functions based on large and fine-scale similarity spectra reveal that the large-scale noise radiation from an F-22A deviates from the similarity spectrum model.