A novel approach has been developed that allows for very high scale load testing of WebRTC media without the need for 100's of servers as a result of the limited scalability of real browsers being used for load testing of WebRTC media.
Why wait until your customers complain that calls stopped working? Learn how to automate WebRTC calls and continuously monitor audio quality! See how to test on various browsers, how to inject/record media and how to extract RTCP statistics to visualise media metrics
Creating an effective media server can be a challenging task. In this presentation, I will share my experience in developing a video streaming solution that balances a simple API with advanced capabilities. We'll be discussing how we have taken advantage of Elixir’s features and our approach towards API design.
500ms is the magic number for fluid voice interactions, but most voicebots fall short, leading to frustrating lag. We've solved the latency puzzle by using every optimization possible to combine latest LLMs, speech-to-text, and text-to-speech while seamlessly handling interruptions and backchanneling as appropriate.
This talk presents the first testbed comparing LL-DASH and MOQT's performances and usability under different network conditions. To that effect, we provide several results on the latency and quality achieved. We also demonstrate how MOQT's prioritization capabilities can further improve delivery performance at low latencies.
We finally have WebTransport and WebCodecs in browsers* so the dream is closer to being a reality. This talk will discuss how to use these tools to build something close enough to WebRTC, even without raw UDP sockets. This is the basis of a new IETF standard being developed: Media over QUIC.
WHIP and WHEP have dramatically changed what is possible with WebRTC. WebRTC has found its way into ffmpeg, GStreamer and OBS. Use cases like robotics, broadcasting and security cameras are maturing. This talk reflects on what is happening, and what will happen because of standardized signaling. If you are looking to build something impactful/useful now is the time to get involved, the opportunities have never been greater in WebRTC.
Large Language Models are capable of multi-turn conversations that are useful, interesting, and fun. This talk will provide an overview of combining WebRTC with LLMs for voice-to-voice interactions.