Persons who have lost their larynx and thus speech functionality need to use a substitution voice to regain speech. The electro-larynx (EL) is a widely used device but is known for its unnatural and monotonic speech quality. Previous research has addressed these problems, but until now no significant improvements could be reported. Moreover, the importance of human-centered computing and co-adaptation was only considered marginally. The EL speech system is a complex system including hardware (artificial excitation source or sound transducer), software (control and generation of the artificial excitation signal) and most important of all, the human user. Taking one isolated problem into consideration is not an adequate approach here, since all aspects of the EL speech system need to be integrated into a holistic approach. We have responded by pushing forward the boundaries of the conventional EL device toward a new bionic electro-larynx speech system. In this paper, we review existing approaches and summarize the most important findings. Subsequently, we piece together an overall speech system, which integrates several parts to significantly improve EL speech: we study electromyographic signals to analyze their suitability for hands-free on/off control of the EL speech system and investigate learning effects in the training of naive users. We propose a new transducer for the EL speech system based on electromagnetic mechanisms. The technical properties of the new transducer show significant advantages over the conventional electro-dynamic transducer. Furthermore, we propose an algorithm to estimate a changing fundamental frequency from the speech spectral envelope and an optimized waveform shape for the excitation signal. Listening tests show that a changing fundamental frequency improves the perceived naturalness of EL speech. Listening tests serve as a proof of concept for the resulting EL speech system, which confirm that the proposed system is very promising.