Extracting spatial information from an audio recording is a necessary step for upmixing stereo tracks to be played on surround systems. One important spatial feature is the perceived direction of the different audio sources in the recording, which determines how to remix the different sources in the surround system. The focus of this paper is the separation of two types of audio sources: primary (direct) and ambient (surrounding) sources. Several approaches have been proposed to solve the problem, based mainly on the correlation between the two channels in the stereo recording. In this paper, we propose a new approach based on training a neural network to determine and extract the two sources from a stereo track. By performing a subjective and objective evaluation between the proposed method and common methods from the literature, the proposed approach shows improvement in the separation accuracy, while being computationally attractive for real-time applications.