Audio systems are used to create two-dimensional (2D) and three-dimensional (3D) audio effects which involve the ability to localize sound within a multi-dimensional space. Multi-dimensional audio systems could be used to imitate moving sounds in applications such as home theaters, video games or headphones. When two or more equidistant speakers produce the same sound, the observer perceives the sound to be localized at a single point. The blending of sound from equidistant speakers is called the virtual sound and is perceived to originate from a virtual source. For two speakers, the virtual source is located on a circular arc between the speakers and for three speakers, the virtual source is located on a spherical cone defined by the speakers. For the observer to perceive one sound from multiple sources, the sounds must arrive at the observer at the same time and the sounds must be the same. By calculating the individual speaker gains using the method of vector-based amplitude panning (VBAP), the audio from all the speakers can be manipulated such that the observer perceives the sound to be originating from a single point. The objective of this project was to develop an algorithm that can place an audio tone in a desired location by calculating and controlling the gain factors of each speaker. In this thesis, the results of simulating in MATLAB and testing in the lab, two-dimensional (2D) and three-dimensional (3D) audio systems with multiple speakers placed in testing positions equidistant to the observer are presented. It is envisioned that this research will lead to a better understanding of localization of sound and to a better understanding of how accurately sound is perceived by the human ear.