Model caramel creams containing four levels of dairy fat and nine levels of xanthan gum as thickener were tasted and submitted to viscometric measurements. The panel of 15 tasters evaluated their liking of the consistency, the perceived creaminess and the thickness, using a magnitude estimation scoring method. Results were correlated with rheological parameters: viscosity and jlow behavior index. There is a quadratic relationship between rheological parameters and both liking consistency and perceived creaminess, with an optimum value. The optimum liking does not coincide with that for maximum creaminess; these values depend on the amount of the two ingredients: fat content and thickener. They can be obtainedfrom response suvaces. The perceived thickness is related to viscosity. Some modalities of changes in optimal values for different groups of people are considered.

Model cream soups prepared with different levels of several thickeners were evaluated by a panel who scored their liking of consistency, perceived creaminess and thickness. Their viscosities were measured with a Haake viscometer. The perceived thickness is linearly related to the logarithm of viscosity for all thickeners. Liking consistency shows a quadratic relation with viscosity. Perceived creaminess changes according to the type of thickener. The optimal values of viscosity and flow behaviour index for maximum creaminess and best liking consistency do not coincide. The viscosity and the flow behavior index affect the sensorial responses of liking consistency and creaminess. Response surfaces show that other rheological factors affect the model. The results for soups are compared with result for creams from a previous study. The viscosity for maximum creaminess is higher for creams than for soups while the flow behaviour is higher for soups than for creams.

A study of perceived texture properties was performed on six different gel systems. The results were correlated with eighteen mechanical parameters measured on these gels. The sensorial evaluation was carried out using quantitative descriptive analysis in two locations, England and Switzerland. Data were submitted to canonical analysis, principal component analysis and then compared using a Procrustes analysis. The results show to what extent the method was reproducible, sensitive and transposable. Correlations of sensorial analysis with physical data show how texture words are used, confused and related.

No abstract

No abstract