Study On Hi-Fi Color Reproduction Technology Based On Spectra

Along with the social economy development and the progress of printing technology, shortcomings of the traditional four-color printing could not satisfy the quality requirements from high-end products and brands customers, multitudinous printing enterprises pay attention to high fidelity color reproduction technology for the research direction, and its scope has not only confined to improve color gamut, but also effectively improve the color reproduction precision. Reproduction color method based on spectra to improve the precision and reduce the metamerism phenomenon, is a mainstream direction to realize the generalized high fidelity color reproduction method.This paper analyzes the three-primary color reproduction theory, many-primary color reproduction theory and spectral color reproduction theory, then summarizes the spectral prediction models of the development process which contains Murray-Davies model for a monochrome printer, Neugebauer model suitable for multi-color printing, add n value optimization Yule-Nielsen spectral reflectance model, Cellular Neugebauer Model, Cellular-Yule-Nielsen Model, Yule-Nielsen Spectral Neugebauer Model and Cellular Yule-Nielsen Spectral Neugebauer Model, and then analyze the characteristics of models. Based on light reflection and scattering caused optical dot gain and model accuracy improvement and the actual situation of the printing consideration, this paper selects the CYNSN spectral prediction model as the main research object.CYNSN model contains four unknown parameters, it is respectively:16Neugebauer primary effective area coverage, Neugebauer primary spectrum reflectance, Yule-Nielsen coefficient n and the cell quantity. This paper introduces the experiment, measurement and obtain the large number of training sample. On the basis of analyzing performance of the printer and the color measurement system, use testing samples to evaluate the spectral prediction accuracy of CYNSN model with parameters RRMS and GFC.The experiment results indicate that the optimum parameters of CYNSN model are best n value2.8and cell number6. After the parameter is setting the best value, the average RRMS is0.0002, the max RRMS is only0.0005, and the GFC value close to1, that is to say, measured spectrum reflectance and predicted spectrum reflectance is consistent and the spectral CYNSN model can achieve the good prediction accuracy.