ΔEITP is Now Ratified as ITU-R BT.2124—Is the Industry Ready to Move on from ΔE2000?

Abstract

For many years, $\Delta E 2000$ has been considered the industry standard for estimating perceptual color differences and evaluating the color accuracy of display technologies. Recently, the International Telecommunication Union (ITU) released Recommendation ITU-R BT.2124, which defines a new metric, $\Delta E_{\mathrm {ITP}}$ , as the new standard for evaluating perceptual color differences. Given this new recommendation, it is important that we become familiar with how $\Delta E_{\mathrm {ITP}}$ performs in practice and what can be expected when using it in existing workflows. A key part in this process is having an understanding of the properties of the color space that the color difference metric is based on and how that relates to the intended application of the metric. In this paper, we explore this relationship as it pertains to the $\Delta E2000$ and $\Delta E_{\mathrm {ITP}}$ color difference metrics, specifically in the context of evaluating color reproduction on self-luminous displays. Simulated data and measured data from calibrated displays are used to compare the performance of the metrics. Given that $\Delta E 2000$ and $\Delta E_{\mathrm {ITP}}$ are based on significantly different derivation principles, there are notable differences that must be acknowledged between these two metrics, especially as we transition between them. It is highly likely that $\Delta E 2000$ has been underpredicting perceptual color differences, significantly so when used for the evaluation of high dynamic range (HDR) and wide color gamut (WCG) displays.