Creating React for live streams - Insights on low-latency multimedia processing
Wojciech Barczyński @ Software Mansion
Over 1.5 years ago we started working on an open-source, low-latency video and audio compositor - LiveCompositor. Targeting use cases like WebRTC calls recording, converting them to HLS streams or composing streams for broadcasting came with a cost (and fun!) of dealing with many challenging problems. Synchronizing streams, avoiding premature decoding, reducing memory copies, or preventing UDP buffer overflows are just a few of them. I want to share a few solutions and tricks we learned that made our processing smoother, reduced latency and solved some tricky issues.
I’m a Software Engineer at Software Mansion with over 1.5 years of experience in building open-source multimedia processing and streaming tools.
My work is focus on low-latency, real-time video and audio processing. I enjoy high-level and low-level programming (everything from designing APIs and algorithms to aligning bits in GPU and network buffers)
I also pursuing BSc in Computer Science at AGH UST and frequently engage in multiple extracurricular programming projects (competitions, hackathons, ICPCs, etc.).