Analytical calibration, in its entirety, comprises both qualitative and quantitative analysis and has to be considered, therefore, from a multidimensional point of view. The relationships must be characterised between the presence of certain analytes and typical signal patterns (e.g. at certain wavelengths) on the one hand and between the amount of given analytes and signal intensities on the other. It is useful to take a three dimensional model as the basis for the classical quantitative analysis and calibration. The relationships between the signal intensity and the true concentration of reference materials used for calibration and the concentration to be estimated in real analytical samples spread a three dimensional function whose corresponding two dimensional projections represent the common calibration function, the analytical evaluation function and the recovery function, which is important for validation (testing the accuracy). Under certain conditions the three dimensional model can be simplified to the common two dimensional one. Multivariate calibration relates to multidimensionality in analytical problems and data sets in principle. Multicomponent analysis and multisignal evaluation are characterized and illustrated by special applications.