Blog - Product

May 5, 2020 | Read time 3 min

New real-time container available for efficient, scalable transcription

Speechmatics now offers real-time transcription as a container as well as a virtual appliance. Containers offer even better scalability and efficiency.

Organizations from a range of sectors and verticals are looking to utilize new technologies and machine learning to power their solutions through the inclusion of products like automatic speech recognition (ASR).

The capabilities of ASR help to optimize products and solutions, deliver new and differentiated capabilities and maintain pace with or lead their markets. Innovating with voice and utilizing the data that exists within organizations enables a better understand of customers and make vital improvements to their businesses. When it comes to the application of ASR there are two modes.

Batch transcription Submitting recorded media files (audio or video) to the transcription engine and receiving the full text transcription of each asset as a separate file. Real-time transcription Streaming media (audio or video) to the transcription engine which delivers text output immediately as it happens. Speechmatics’ any-context speech recognition engine works for both batch and real-time transcription, giving businesses the flexibility to build innovative applications at scale. With these applications, organizations can extract information and key insights to uplift their knowledge. With this knowledge, companies can assist other operations and people within their organization to adapt in real-time. Two Speechmatics reports on The State of Voice in the Media and Broadcast Industry and The State of Voice in the Contact Centre Industry found that real-time use of speech recognition provides more value (53% & 45%, respectively) than pre-recorded (27% & 22%). However, some 20% & 33% identified both modes to be equally valuable to their companies. Virtual appliances: Speechmatics’ real-time speech-to-text offering Speechmatics has supported real-time transcription through a real-time virtual appliance (RTA) since 2018. The virtual appliance is a small footprint piece of software able to be deployed wherever the customer wants. Hosted in a secure customer site or in a public/private cloud, the Speechmatics RTA provides best-in-class transcription, as it happens. The RTA can be deployed anywhere, offering deployment flexibility to customers as well as data control. However, we understand that virtual appliances may not be ideal for all customers.

Containers: Now available in Speechmatics’ real-time speech-to-text offering Organizations that want to build highly scalable services based around a microservices architecture are more suited to container-based solutions. This is because they do not have the memory and CPU overheads that virtual appliances do. Containers are also faster to scale up and down than virtual appliances. Containers, therefore, provide an effective solution for real-time use cases which often deal with dynamic and spiky traffic. For flexible workloads, containers offer an accelerated and more elastic capability for unpredictable resource requirements. The real-time container enables customers to deploy Speechmatics’ real-time transcription within their solution stack. It increases operational simplicity and customers can use their existing orchestration layer to manage both their own and Speechmatics’ containers to build flexibly scalable services. The real-time container features a new licensing model which lets customers upgrade and update rapidly. Speak to an expert Want to know more about Speechmatics’ real-time containers? Why not speak to one of our experts to see how we can help you and your organization to innovate with voice?