Enghouse EspialTV enables TV accessibility with Amazon Polly
This is a guest post by Mick McCluskey, the VP of Product Management at Enghouse EspialTV. Enghouse provides software solutions that power digital transformation for communications service operators. EspialTV is an Enghouse SaaS solution that transforms the delivery of TV services for these operators across Set Top Boxes (STBs), media players, and mobile devices.
A large audience of consumers use TV services, and several of these groups may have disabilities that make it more difficult for them to access these services. To ensure that TV services are accessible to the broadest possible audience, we need to consider accessibility as a key element of the user experience (UX) for the service. Additionally, because TV is viewed as a key service by governments, it’s often subject to regulatory requirements for accessibility, including talking interfaces for the visually impaired. In the US, the Twenty-First Century Communications and Video Accessibility Act (CVAA) mandates improved accessibility for visual interfaces for users with limited hearing and vision in the US. The CVAA ensures accessibility laws from the 1980s and 1990s are brought up to date with modern technologies, including new digital, broadband, and mobile innovations.
This post describes how Enghouse uses Amazon Polly to significantly improve accessibility for EspialTV through talking interactive menu guides for visually impaired users while meeting regulatory requirements.
Challenges
A key challenge for visually impaired users is navigating TV menus to find the content they want to view. Most TV menus are designed for a 10-foot viewing experience, meaning that a consumer sitting 10 feet from the screen can easily see the menu items. For the visually impaired, these menu items aren’t easy to see and are therefore hard to navigate. To improve our UX for subscribers with limited vision, we sought to develop a mechanism to provide audible descriptions of the menu, allowing easier navigation of key functions such as the following:
- Channel and program selection
- Channel and program information
- Setup configuration, closed-caption control and options, and video description control
- Configuration information
- Playback
Overview of the AWS talking menu solution
Hosted on AWS, EspialTV is offered to communications service providers in a software as a service (SaaS) model. It was important for Enghouse to have a solution that not only supported the navigation currently offered at the time of launch, but was highly flexible to support changes and enhancements over time. This way, the voice assistance continuously evolved and improved to accommodate new capabilities as new services and features were added to the menu. For this reason, the solution had to be driven by real-time APIs calls as opposed to hardcoded text-to-speech menu configurations.
To ensure CVAA compliance and accelerate deployment, Enghouse chose to use Amazon Polly to implement this talking menu solution for the following reasons:
- We wanted a reliable and robust solution within minimal operational and management overhead
- It permitted faster time to market by using ready-made text-to-speech APIs
- The real-time API approach offered greater flexibility as we evolved the service over time
The following diagram illustrates the architecture of the talking menu solution.
Using the Amazon Polly text-to-speech API allowed us to build a simple solution that integrated with our current infrastructure and followed this flow:
- Steps 1 and 2 – When TV users open the menu guide service, the client software running on the Set Top Box (STB) makes a call via the internet or Data Over Cable Service Interface Specification (DOCSIS) cable modem, which is routed through the cable operators headend server to the Espial Guide service running on the AWS Cloud.
- Step 3 – As TV users interact with the menu guide on the STBs, the client software running on the STBs sends the string containing the specific menu description highlighted by the customer.
- Step 4 – The cable operators headend server routes the request to a local cache to verify whether the requested string’s text-to-speech is cached locally. If it is, the corresponding text-to-speech is sent back to the STB to be read out loud to the TV user.
- Step 5 – Each unique cable operator has a local cache. If the requested string isn’t cached locally in the cable operator’s environment, the requested string is sent to the EspialTV service in AWS, where it’s met by a secondary caching server to respond to the request. This secondary layer of caching hosted in the Espial environment ensures high availability and increases cache hit rates. For example, if the caching servers on the cable operator environment is unavailable, the cache request can be resolved by the secondary caching system hosted in the Espial environment.
- Steps 6 and 7 – If the requested string isn’t found in the caching server in the EspialTV service, it’s routed to the Amazon Polly API to be converted to text-to-speech, which is routed back to the cable operator headend server and then to the TV user’s STB to be read out loud to the user.
This architecture has several key considerations. Firstly, there are several layers of caching implemented to minimize latency for the end user. This also supports the spikey nature of this workload to ensure that only requests not found in the respective caches are made to Amazon Polly.
The ready-made text-to-speech APIs provided by Amazon Polly enables us able to implement the service with just one engineer. We also reduced the expected delivery time by 75% compared to our estimates for building an in-house custom solution. The Amazon Polly documentation was very clear, and the ramp-up time was limited. Since implementation, this solution is reliably supporting 40 cable operators, which each have between 1,000–100,000 STBs.
Conclusion
EspialTV offers operators a TV solution that provides fast time to revenue, low startup costs, and scalability from small to very large operators. EspialTV offers providers and consumers a compelling and always relevant experience for their TV services. With Amazon Polly, we have ensured operators can offer a TV service to the broadest possible range of consumers and align with regulatory requirements for accessibility. To learn more about Amazon Polly, visit the product page.
The content and opinions in this post are those of the third-party author and AWS is not responsible for the content or accuracy of this post.
About the Author
Mick McCluskey is VP of Product Management at Enghouse, a leading provider of software solutions helping operators use digital transformation to drive profitability in fast-changing and emerging markets. In the area of video solutions, Mick has been pivotal in creating the EspialTV solution—a truly disruptive TVaaS solution run on the AWS Cloud that permits pay TV operators to manage transition while maintaining profitability in a rapidly changing market. He is currently working on solutions that help operators take advantage of key technology and industry trends like OTT video, IoT, and 5G. In addition to delivering cloud-based solutions, he continues his journey of learning how to play golf.