Descriptive analysis of complex odors: reality, model or illusion?

The primary sensory tool for specifying the characteristics of a complex aroma, fragrance, flavor or other odorous mixture of volatiles is descriptive analysis. Descriptive analysis uses a trained panel to specify the intensities of specific attributes, based on a psychophysical model for intensity scaling. However, the use of descriptive techniques for complex and well-blended aromas gives rise to several problems. The psychophysical intensity model based upon independent odor notes may be a poor way to characterize odor experience, bringing into question whether descriptive analysis is an adequate tool for sensory analysis of complex smells. These problems include the following: (1) disagreement among experts in the most prominent odor notes of a single product and other individual differences problems, (2) a correspondence between similarity scaling and intensity scaling, (3) the substitution of applicability measures for intensity, (4) the need to use mid-tier, general odor terms for profiling complex fragrances, and (5) blending and integration effects. Data will be presented on citrus-woody mixtures showing that ratings of similarity and intensity are highly correlated, suggesting a common underlying process for both ratings. A related issue concerns whether odors and their mixtures are perceived as unitary or analyzable percepts. With these same stimuli, the perception of singularity vs, mixed-ness of stimuli is difficult to predict. Sensory scientists should question the validity of descriptive data for such stimuli and avoid the simplistic mistake of equating data with perception. The use of simple and apparently independent intensity scales may produce the illusion that the odor experience is a collection of independent analyzable "notes" when it is not. (C) 1999 Elsevier Science Ltd. All rights reserved.