•Balanced focus on speech science, perception, and speech processing/engineering. •Comprehensive survey of speech production modalities and assessment/equalization techniques for increased robustness of speech systems in real-world applications. •All topics are accompanied with critical comments and original research by the authors. •The text introduces readers to closely-related subjects that rarely appear together in a single book. The bookprovides a survey on the variability of speech production due to stress, emotions, and Lombard effect, and its impacts on speech-enabled systems. The book is structured in 2 parts: speech analysis and speech systems. In the speech analysis section, both aspects of speech production and perception are studied, with the emphasis on significant trends of interest to speech scientists and phenomena that help engineers design robust automated speech systems. Attention is given to stress modalities underrepresented in the literature, such as physical task stress and cognitive load in real world scenarios. The authors comment on experimental design and confront studies conducted on simulated and real data. In the speech systems section, the authors focus on automated speech assessment, speech recognition, and speaker identification, with the emphasis on the design of the key system stages – front-end feature extraction and acoustic modeling. The text presents state-of-the-art techniques that increase system robustness when exposed to varying speech modalities.