Seven color separation criteria are evaluated and compared for multi-ink printer characterization, which is a union of five 3-ink and six 4-ink cellular Yule-Nielsen spectral Neugebauer (CYNSN) submodels. In the experimental stage, testing reflectances from printing samples and Munsell color samples are employed. The results show that the prediction and actual accuracy of all the seven selection criteria are approximately the same for printing samples. As for the Munsell samples, the combE00 &MI outperforms other selection criteria because it combines one color difference with two metamerism indices, which are important for the heterogeneous samples during prediction.

Conventional printer characterisation models are generally based on the assumption that the densities of primary colours are additive. However, additivity failure frequently occurs in practice. We propose a novel grey component replacement (GCR) method based on the spectral density sub-additivity equations in this letter for spectral characterisation of a 4-ink colour printer. The method effectively correct the error caused by additivity failure. Real high-quality hardcopy samples are produced as evidence of the feasibility of the proposed method and to evaluate the model performance. Finally, the GCR model for characterising colour printer with high spectral and colorimetric prediction accuracy is established.

A spectral-based 8-ink characterization model is developed to accurately predict the recipe for a multi-ink printer. The 8-ink color separation method is a union of five 3-ink and six 4-ink combinations based on the cellular Yule-Nielsen spectral Neugebauer model with a recipe selection strategy. The performance levels of the forward and backward models are evaluated for individual ink combinations using printed testing samples. Furthermore, the spectral-based method performs better compared with the XYZ-based approach. On the basis of the backward model performance, a novel fast recipe selection strategy is proposed and estimated.