Methodology for a comparative quantitative analysis of film colour: a comparison between Cannes and box-office results

Abstract

This paper proposes a method for quantitative colour analysis aimed at determining colour variation between different sets of films with a common selection criterion. The process is exemplified by comparing films selected at the Cannes Film Festival and the highest-grossing films in the years 1988 and 2018. The method will try to answer several questions. Firstly, whether there are any differences between the films competing in Cannes Film Festival and those most watched by audiences. Secondly, if these differences exist in both the year 1988 and 2018. Thirdly, if the colour of films exhibited at Cannes and highest-grossing film have evolved during the 30-year difference between samples. To analyse these films, one image every 14 s is extracted and all frames are used to quantify the amount of each colour present. Values for each category undergo a Kolmogorov–Smirnov test to determine whether the distribution of each colour is different for both categories compared. The relative difference between colours is computed to find out if those colours where variations are present have increased or decreased their presence within those sets compared. This methodology enables film researchers to statistically prove that there are colour changes in those films included in their research.