KRANKY GEEK WEBRTC SHOW

Glad to have you here, we will review what happened with WebRTC in 2022 and how is that affecting our event today and the future for all of us.

With WebRTC entering a maturing stage and growing adoption of commercial products, is there still room for open source to survive? Some of the leading open source projects – Janus, Jitsi, mediasoup, and Pion – will discuss how open source had changed and what’s next in this exciting panel discussion.

Google, well known for its AI focus, will discuss its approach and learnings from adding AI features to Google Meet.

Machine Learning is great at noise suppression and cleaning up audio. The challenge: doing this in real-time, cost-effectively in the browser for WebRTC calling. Krisp will cover best practices for real-time audio pipelines leveraging the latest JavaScript APIs and WASM.

Collecting WebRTC statistics is a no brainer. Collecting more statistics about network conditions is a known problem as well. But what about figuring out based on such statistics what the issue is? In this session, we’ll go over our own processes in unraveling statistics to assist troubleshooting WebRTC connectivity and quality issues.

Google will provide insights on what they have been working on and will share their annual roadmap update.

Compositing in the cloud with native pipelines (not browsers) to scale streaming and recording with 1000s of participants

Hear how to replace RTMP and HLS with ultra-low-latency WebRTC streams leveraging these new protocols that are establishing a foothold in the broadcasting industry.

Game streaming is a demanding use-case that requires high-bitrate ultra-low-latency video without freezes. Forward error correction is a technique that allows NVIDIA’s GeForce NOW to deliver a smooth experience even when there is packet loss on the network. We will explain how this feature was implemented and measured.

When WebRTC launched in 2011, it came with a new audio codec called opus which remains a great choice today. We will review contenders like the machine-learning based Lyra and Satin codecs and reliability mechanisms like RED and FEC to see how they give even better quality.