Note that, in the above-described example, the gradation values for which the spectral reflectances are obtained regarding the reference color and the prediction target gradation values coincide with each other. However, such a case is also assumed in which the prediction target gradation values include a gradation value other than the gradation values for which the spectral reflectances are obtained regarding the reference color. For the gradation value other than the gradation values for which the spectral reflectances are obtained regarding the reference color, spectral reflectances thereof cannot be directly obtained from the second relational equation. Accordingly, the spectral reflectances of such a gradation value are obtained by performing spline interpolation (for example, cubic spline interpolation) based on data of the spectral reflectances directly obtained from the second relational equations. In such a case, the above-described Step S160 is composed of Step S162 and Step S164 as illustrated in FIG. 10. In Step S162, the prediction values of the spectral reflectances of the gradation values for which the spectral reflectances are obtained regarding the reference color among the plurality of gradation values serving as the prediction target gradation values are obtained by applying the spectral reflectances of the maximum gradation values for the predict on target color to the corresponding second relational equations. Thereafter, in Step S169, the prediction values of the spectral reflectances of the gradation value for which the spectral reflectances are not obtained regarding the reference color among the plurality of gradation values serving as the prediction target gradation values are obtained by spline interpolation based on the prediction values obtained in Step S162. Note that, though linear interpolation can also be used in place of the spline interpolation, the prediction values are obtained more accurately by the spline interpolation than the linear interpolation.