An eddy current system transfer function, which converts between the impedance output of simulation and the voltage output of commercial systems, is a common use of transfer functions in nondestructive testing. Normally, transfer functions are for specific settings on the commercial equipment. Any change to certain settings such as gain or drive voltage results in the transfer function no longer being valid. This creates an issue with NDT 4.0, with the integration of digital engineering and nondestructive testing. This issue becomes larger when trying to account for settings changes with normal regression techniques. This is because the internal circuitry of the commercial systems contains proprietary data processing and physical circuitry components. A deep learning approach was used to develop a system transfer function using basic MATLAB tools for bidirectional conversion between complex impedance and complex voltage that accounts for the change in settings. The settings accounted for are gain, drive, and frequency. Rotation can be accounted for in pre and post processing depending on the direction of the conversion. The system transfer functions showed high accuracy within trained frequencies with both trained and novel drive and gain values when converting from complex voltage to complex impedance. A simple linear interpolation of predicted values at novel frequencies achieves the same high accuracy within novel frequency settings. The inverse transfer function showed high accuracy within trained frequencies, drive and gain values, but struggles to predict commercial system measurements in untrained regions. This novel enhanced system transfer function has high potential impact in the application of model-assisted probability of detection, supplemental data-assisted probability of detection, settings optimization for enhanced probability of detection, digital twins, and other enhanced digitally integrated predictive capabilities in nondestructive testing.