In this thesis, the study and design of Voice Change and Voice Conversion systems are presented. Particularly, a voice change system manipulates a speaker’s voice to be perceived as it is not spoken by this speaker; and voice conversion system modifies a speaker’s voice, such that it is perceived as being spoken by a target speaker. This thesis mainly includes two sub-parts. The first part is to develop a low latency and low complexity voice change system (i.e. includes frequency/pitch scale modification and formant scale modification algorithms), which can be executed on the smartphones in 2012 with very limited computational capability. Although some low-complexity voice change algorithms have been proposed and studied, the real-time implementations are very rare. According to the experimental results, the proposed voice change system achieves the same quality as the baseline approach but requires much less computational complexity and satisfies the requirement of real-time. Moreover, the proposed system has been implemented in C language and was released as a commercial software application. The second part of this thesis is to investigate a novel low-complexity voice conversion system (i.e. from a source speaker A to a target speaker B) that improves the perceptual quality and identity without introducing large processing latencies. The proposed scheme directly manipulates the spectrum using an effective and physically motivated method – Continuous Frequency Warping and Magnitude Scaling (CFWMS) to guarantee high perceptual naturalness and quality. In addition, a trajectory limitation strategy is proposed to prevent the frame-by-frame discontinuity to further enhance the speech quality. The experimental results show that the proposed method outperforms the conventional baseline solutions in terms of either objective tests or subjective tests.