The Electric network frequency (ENF) signal is a unique signal for different parts of the world. It is captured by electric devices, and can be used in authentication and automatic synchronization of digital media recordings. In this paper we propose an algorithm to extract ENF from power and audio recordings, and use ENF criterion to identify the region-of-recording. We also propose a design of a circuit to record the electrical power grid. The ENF extraction was performed on one of the three highest peaking harmonics in the spectrum, using Discrete Fourier transform (DFT) in case of audio and Estimation of Signal Parameters via Rotational Invariance Techniques (ESPRIT) method in case of power recordings. The classification tool used was a set of binary support vector machines (SVM). Different combinations of features were used for each classification subgroup, resulting in 92% success rate on the Practice dataset, and 86% on cross-validation with the Training dataset, both containing recordings from 9 different grids. The sensing circuit, constructed with Hall effect sensor and Arduino UNO, was used to obtain over 10 hours of recordings of the Serbian power grid.