Speech analytics

Abstract: We present a state-of-the-art speech recognition system developed using end-to-end deep learning. Our architecture is significantly simpler than traditional speech systems, which rely on laboriously engineered processing pipelines; these traditional systems also tend to perform poorly when used in noisy environments. In contrast, our system does not need hand-designed components to model background noise, reverberation, or speaker variation, but instead directly learns a function that is robust to such effects. We do not need a phoneme dictionary, nor even the concept of a "phoneme." Key to our approach is a well-optimized RNN training system that uses multiple GPUs, as well as a set of novel data synthesis techniques that allow us to efficiently obtain a large amount of varied data for training. Our system, called Deep Speech, outperforms previously published results on the widely studied Switchboard Hub5'00, achieving 16.0% error on the full test set. Deep Speech also handles challenging noisy environments better than widely used, state-of-the-art commercial speech systems.

Abstract: From the Publisher: New advances in spoken language processing: theory and practice In-depth coverage of speech processing, speech recognition, speech synthesis, spoken language understanding, and speech interface design Many case studies from state-of-the-art systems, including examples from Microsoft's advanced research labs Spoken Language Processing draws on the latest advances and techniques from multiple fields: computer science, electrical engineering, acoustics, linguistics, mathematics, psychology, and beyond. Starting with the fundamentals, it presents all this and more: Essential background on speech production and perception, probability and information theory, and pattern recognition Extracting information from the speech signal: useful representations and practical compression solutions Modern speech recognition techniques: hidden Markov models, acoustic and language modeling, improving resistance to environmental noises, search algorithms, and large vocabulary speech recognition Text-to-speech: analyzing documents, pitch and duration controls; trainable synthesis, and more Spoken language understanding: dialog management, spoken language applications, and multimodal interfaces To illustrate the book's methods, the authors present detailed case studies based on state-of-the-art systems, including Microsoft's Whisper speech recognizer, Whistler text-to-speech system, Dr. Who dialog system, and the MiPad handheld device. Whether you're planning, designing, building, or purchasing spoken language technology, this is the state of the art—fromalgorithms through business productivity.

Abstract: An audio-visual corpus has been collected to support the use of common material in speech perception and automatic speech recognition studies. The corpus consists of high-quality audio and video recordings of 1000 sentences spoken by each of 34 talkers. Sentences are simple, syntactically identical phrases such as "place green at B 4 now". Intelligibility tests using the audio signals suggest that the material is easily identifiable in quiet and low levels of stationary noise. The annotated corpus is available on the web for research use.