Live Video on Web and mobile

Integrating real-time video into Web and mobile services

We live in a society where we consume, learn and live through what we see. That is why it is essential when communicating to incorporate audiovisual content. Integrating this type of content on the Web and mobility has been a reality for years. However, doing it with content in real time and in a scalable way is still a challenge.

Currently the incorporation of web video conferencing is a reality that is allowing us to develop and test our video transmission infrastructures with a real case and verify their performance, how to manage the quality they provide and how they scale. But the reality is that there are many services that involve real-time video that go far beyond simple video conferencing. Integrating different means of communication, incorporating use cases such as simultaneous translation, mass video distribution, are real needs that NAeva Tec is addressing in the solutions we offer.

WebRTC

For 8 years, WebRTC has been the standard that governs interactive communications in real time. Building on Internet standards, WebRTC has made high-quality cross-browser video calls a reality. Its success attests to the fact that it is supported in practically all web browsers both on desktop and mobile. What at the same time facilitates and allows that WebRTC services can be performed that will be accessible in practically all devices.

WebRTC allows the construction of services in three levels to be considered beyond the point-to-point call between browsers, for that an additional element is needed which is the Media Server, in charge of managing media flows in real time, switching them between origin and destination and also to implement associated value-added services: multiconferencing, transcoding,

At Naeva Tec we started the Kurento project in 2011, which quickly became a benchmark in terms of open source WebRTC Media Servers, and which is currently still active.

More recently the Kurento infrastructure has been improved with the introduction of OpenVidu.

And on top of all these infrastructures, and other commercial ones such as Twilio, AWS Chime, etc. At Naeva Tec we create custom applications that implement specific use cases not covered by the most common videoconferencing applications: online translator support, integration with telephone services, specific tele-teaching applications, etc.

Non-interactive real-time video

Of course WebRTC is the technology to use when we talk about interactivity, but there are also services that involve live video, but do not require interactivity. In those cases, the technology extends beyond WebRTC to include other more common technologies in the web world: HLS, RTMP, etc.

But to integrate live video with these technologies, it is still necessary to have Media Server elements that are capable of collecting live video from its original sources and adapting it to the characteristics necessary for its distribution with low latency: HLS, multi-resolution, RTMP. These tasks involve low-level tools such as ffmpeg or GStreamer that are used to make the operations necessary for this adaptation as efficiently as possible, which are often expensive operations in terms of computing resources and require detailed study and optimizations to to be able to execute them in real time.

We have successfully implemented this type of service in several clients using AWS as an execution platform, even in serverless environments.

The future: extract value from content in real time

There is a future beyond video conferencing for real-time video services, interactive or not. Actually, video streams can contain very valuable information that can be exploited in real time. Both the information itself is valuable, such as the example of surveillance cameras and often the context they provide to other information. In an emergency call, being able to have video can provide essential information for the requested emergency service: the location of the person sending the call can be inferred from what is seen around him, information to evaluate the type of situation. it can also infer from what is observed in the call, etc. In many cases, a human observer is necessary to evaluate this context information, but there is already technology that can be integrated with the systems we have for transporting video in real time that allow us to start taking advantage of these functionalities in real cases.