The existing tone correction procedure in the Effective Perceived Noise Level (EPNL) calculation procedure required for aircraft certification under Part 36 of the Federal Aviation Regulations was compared with other tone correction procedures, including the SAE Aerospace Recommended Practice 1071 and a multitone procedure due to Kryter and Pearsons. Different amounts of tone correction (level-weightings) and varying degrees of tone correction at different times in the flyover (time-weightings) were also explored. Also studied was a measure of spectral fluctuation, developed by NASA and known as spectral change. The research was limited to considering revisions within the framework of one-third octave, 0.5 second interval analysis, since such revisions can be quite easily implemented. The various tone correction noise metrics were tested against subjective judgements furnished by NASA of the noise from a range of propjet, turbojet, low and high bypass ratio turbofan, and supersonic commercial aircraft. It was found that a revision based on spectral change could, after further development, be a means to improve the accuracy of the EPNL metric. However, the success of the various other potential revisions depended on the characteristics of the data base tested. It was shown that research into improved metrics should be based on experimental plans which account for the correlations among the noise variables and the presence of any interactions. A separate, psychoacoustical pilot experiment was also performed into the effects of pseudotones on judged noisiness. (Pseudotones are low frequency tones introduced into a measured spectrum by ground reflections near the microphone).