During the past five years, many approaches to vocoder synthesis have been proposed. The use of the spectral-flattening technique in the voice-excited vocoder represents the first attempt to apply a synthesis process capable of operation in other vocoder systems, such as the channel vocoder. The use of this technique represents the first major variation in the basic channel-vocoder concept and certainly is worthy of a detailed theoretical and practical evaluation. Conflicts concerning the theoretical value of this technique are now prevalent. A detailed study of the use of spectral flattening (or dual-filter synthesis) in the channel vocoder has been completed and evaluations of various implementation of this technique have been performed. As a result of these studies and evaluations, some definite answers and comments on the theoretical and practical value of spectral flattening have been derived. Conclusions reached and theoretical considerations are presented. A tape demonstration follows. [A portion of this study supported under contract No. AF19(628)-586.]