Ultrasonic Doppler techniques are well established and allow qualitative and quantitative flow analysis. However, due to inherent limitations of the imaging process, the actual flow dynamics and the ultrasound (US) image do not always correspond. To investigate the performance of ultrasonic flow imaging methods, computational fluid dynamics (CFD) can play an important role. CFD simulations can be directly processed to mimic ultrasonic images or can be further coupled to ultrasound simulation models. We studied both approaches in the clinically relevant setting of a carotid artery using color flow images (CFI). The first order approach consisted of producing ultrasound images by color-coding CFD-simulations. For the second order approach, CFI was simulated using an ultrasound simulator, which models blood as a collection of point scatterers moving according to the CFD velocity fields. Color flow images were also measured in an experimental setup of the same carotid geometry for comparison. Results showed that during dynamic stages of the cardiac cycle, realistic ultrasound data can only be achieved when incorporating both the dynamic image formation and the measurement statistics into the simulations.