The first purpose of this phantom study was to verify whether a contrast-to-noise ratio (CNR)-driven exposure control (CEC) can maintain target CNR in angiography more precisely compared to a conventional detector dose-driven exposure control (DEC). The second purpose was to estimate the difference between incident air kerma produced by CEC and DEC when both exposure controls reach the same CNR. Methods: A standardized 3D-printed phantom with an iron foil and a cavity, filled with iodinated contrast material, was developed to measure CNR using different image acquisition settings. This phantom was placed into a stack of polymethylmethacrylate and aluminum plates, simulating a patient equivalent thickness (PET) of 2.5-40 cm. Images were acquired using fluoroscopy and digital radiography modes with CEC using one image quality level and four image quality gradients and DEC having three different detector dose levels. The spatial frequency weighted CNR and incident air kerma were determined. The differences in incident air kerma between DEC and CEC were estimated. Results: When using DEC, CNR decreased continuously with increasing attenuation, while CEC within physical limits maintained a predefined CNR level. Furthermore, CEC could be parameterized to deliver the CNR as a predefined function of PET. To provide a given CNR level, CEC used equal or lower air kerma than DEC. The mean estimated incident air kerma of CEC compared to DEC was between 3% (PET 20 cm) and 40% (PET 27.5 cm) lower in fluoroscopy and between 1% (PET 20 cm) and 55% (PET 2.5 cm) lower in digital radiography while maintaining CNR. Conclusion: Within physical and legislative limits, the CEC allows for a flexible adjustment of the CNR as a function of PET. Thus, the CEC enables taskdependent examination protocols with predefined image quality in order to easier achieve the as low as reasonably achievable principle. CEC required equal or lower incident air kerma than DEC to provide similar CNR, which allows for a substantial reduction of skin radiation dose in these situations.