Reviewing literature one will discover serveral approaches for 3D audio reproduction via loudspeakers. If we examine them more closely we will retrieve that eighter their development is related to the continuation of the stereo reproduction paradigma for a certain sound source direction or it is related to a more general holistic paradigma to reprocude the whole sound field in a certain restricted area. Another perspective will subdivide these approaches into channel-, object- and scene-based audio formats. Furthermore, we can identify that several approaches just mimic or cause certain perceptual cues that will be evoked by natural sound sources under real conditions. This might lead to a classification from functional towards a pure physical reproduction of real sound field scenarios. In practise, as self-awareness shows, the secret recipe seems to lie somewhere in between – try to meet the physical challenges and stay as functional as long as it sounds better. Optimal perceptual or at least subjective convincing results for various practical applications will be achieved with different approaches causing different effort. Nevertheless, this contribution reviews the basics of Ambisonics, outlines some successful examples and will address recent achievements.