Speech Technology and Systems in Human

S peech technology and systems in human-machine communication have witnessed a steady and remarkable advancement over the last two decades. Fundamental changes have taken place from theoretical foundations to practical systems, from laboratory prototypes to commercial products, and from proprietary softwares to industrial standards. As the information age continues, research in speech technology is further accelerated by the advent of powerful computing devices, the data-driven pattern recognition methods, and the need to generate machine understandable metadata for Web contents and other information sources. Although various systems are built and applied to numerous applications, the full potential of speech technology still remains to be uncovered. This special section fills the need of a comprehensive review of new approaches and advances of speech technology under a broad perspective of intelligent humanmachine communication. Speech technology and systems touch upon many essential signal processing techniques and are in the core of multimodal/ multimedia communication research. We hope that such a systematic and upto-date overview of the field can bring the awareness and applications of speech technology closer to the general signal processing community. New research trends and directions in the field of speech technology and human-machine communication systems have been evolving rapidly in recent years due to the changing business environment and technology advances. With the Internet and the Web, an increasingly large amount of voice and speech data is made available. This, together with fast computing devices, leads to a new wave of advances on speech “document” understanding and multimedia/multimodal content search. New algorithms are being studied, many of which may not have been computationally feasible in the old days. Also, the large deployment of voice over IP (VoIP) has revitalized the research on noise-robust speech processing and recognition over the IP network, which is a very different environment than in the past. While scientific rigor remains the paramount selection criterion, an attempt is made to provide a balanced coverage of new research trends among articles selected for this special section. About one year ago, we put out the call for papers for articles about speech technology and human-machine communication. We received a large number of submissions, and we would like to thank all authors for their submission. After a peer-review process, nine articles were selected that provide a comprehensive overview of the landscape in speech technology and human-machine communication. A brief overview of the selected articles is provided below in the context of the general theme of this special issue—human-machine communication.