We present a quantitative comparison between the high-pressure fuel spray images obtained experimentally using classical imaging with coherent and incoherent ultrafast illuminations recorded using a compatible CMOS camera. The ultrafast, incoherent illumination source was extracted from the supercontinuum generated by tightly focusing the femtosecond laser pulses in water. The average velocity maps computed using time-correlated image-pairs and spray edge complexity computed using the average curvature scale space maps are compared for the spray images obtained with the two illumination techniques and also for the numerically simulated spray using the coupled volume of fluid and level set method for interface tracking (direct numerical simulation or DNS). The spray images obtained with supercontinuum-derived, incoherent, ultrafast illumination are clearer, since the artifacts arising due to laser speckles and multiple diffraction effects are largely reduced and show a better correlation with the DNS results. Introduction The enhancement in the efficiency and reduction in the emitted pollutants by the liquid or gaseous fuel based combustion engines requires a detailed understanding of various processes leading up to the combustion of the fuel, like injection, atomization, vaporization, etc. Among these processes fuel injection, leading to its atomization, plays a very crucial role since it is one of the initial steps and determines the drop size distribution, which in turn influences the liquid evaporation rate and the fuel-air mixture efficiency inside the combustion chamber and hence, the efficiency of the whole combustion process largely depends on it. To this end, several optical diagnostic tools, most being non-intrusive in nature, have been proposed to understand each of these processes individually [1]. The general trend in Diesel applications so far has been to increase the injection pressure (up to 2000 bars) and to decrease the orifice diameter (down to 100 µm), as these two parameters seem to improve the engine efficiency, but still the breakup of the spray is not fully understood. The main reason is that the form and distribution of liquid structures in the spray is a result of a lot of interdependent and complex processes such as hydrodynamic instability, cavitation, turbulence, etc. [2], which complicate the analysis of atomization phenomena and prevent direct control of the spray formation. This work aims at improving the classical imaging methods particularly ultrafast shadow imaging for the high-pressure fuel sprays [3] by replacing the coherent ultrafast laser source by an incoherent ultrafast source, derived from the supercontinuum generated using a femtosecond laser. Since the light source used for imaging is now incoherent, the artifacts arising due to speckles and multiple diffraction effects are reduced. For quantitative comparison, interface velocity maps and complexity are computed from the time-correlated image-pairs and average curvature scale space (CSS) map respectively, extracted from a large set of images obtained using these two kinds of illumination sources and the results are compared with the simulated version of the spray computed using coupled volume of fluid and level set method for interface tracking, i.e. by direct numerical simulation (DNS). The incompressible Navier-Stokes equations are solved following a projection method and coupled to a transport equation for the level set function on a Cartesian meshing [4, 5].