Issue link: https://hub.radisys.com/i/1048301
4 www.radisys.com RADISYS SOLUTION BRIEF | Integrated Speech Recognition Solution Powers Next-Gen Service Providers Radisys Engage Media Server Embedded "In-Call" Keyword Detection Reduces Cost and Improves User Experience The Radisys Engage Media Server, an industry-leading virtual media server platform for real-time communication services (such as VoLTE, Unified Communications, Conferencing, Contact Center / Customer Service and WebRTC), combines the ability to process speech ranging from a limited vocabulary of keywords and commands to natural language interaction in the context of voice and video calls. With integrated speech recognition, Engage Media Server offers a much more cost-effective alternative to traditional ASR approaches and lowers the cost and complexity of deploying and maintaining external network elements by as much as 90%. Also, essential to continued growth in the utility of the technology are the quality of experience, media optimization and on-the-fly transcoding of the user-generated content. These are pivotal for both—the users of mission-critical vertical market applications such as healthcare, law enforcement, senior living care, automotive, translation, and other similar services; as well as for more common mass-market consumer applications. Radisys Engage Media Server provides an extensive set of features—Integrated Keyword Detection system, Transcoding, Conferencing, Voice Quality Enhancements (VQE – including noise reduction, echo cancellation, gain control, and more), Announcements, and Recording - with simplified deployment and management. Embedded VQE capability helps reduce recognition errors and the need for post-processing human intervention. User-B MRCP Enable Keyword Detection Identify Specific Keywords Keyword Detected Optional Post Keyword Detection Advanced Analytics Application Server ASR Server Audio Decoding Voice Quality Enhancement Keyword Detection Engage Media Server • Enables Service Providers to Offer Differentiated Mass Market "In-Call" Speech Recognition Services at Significantly Lower Cost without Depending on 3rd Party Solutions • Embedded always-on "in-call" speech processing at a fraction of the cost • Embedded Voice Quality Enhancement ( VQE) enhances the accuracy of speech recognition under difficult and varied network conditions • Optimizes performance by avoiding latency, complexity and extra processing • Optimizes cost by invoking advanced ASR, transcription and analytics engines "As Needed" based on business policies • Enables new services without requiring additional software downloads on user devices and deploying expensive NLP solutions for all calls • Enables CSPs to create a unique branded persona • Leverages CSP networks for guaranteed SLAs and privacy Key Benefits