Synthetic Media Processing Laboratory Information

SMPL is a young company founded by Koen Vos and Jonathan Christensen dedicated to making online conversations as reliable and natural as possible anywhere, on any device.  Today's leading audio / video software, with very few exceptions, is based on WebRTC open source. Most of that code is 10-15 years old now. There have been incremental tweaks along the way, but few fundamental improvements. We are applying the latest signal processing techniques along with new approaches that leverage deep neural networks -- with big breakthroughs that solve fundamental problems.  As users we encounter those sudden drops and robotic sounds (bursty packet loss), conversations with very poor audio and video (slow or intermittent connections). All of this is amplified by recent remote working trends (MiFi, satellite, outdated access networks). For service operators we are working to conserve constrained resources in cloud terminated services, eliminate the built-in algorithmic delay of current approaches, and dramatically reduce bandwidth consumption, while enabling new revenue and use cases like remote live entertainment and ultra-low-latency gaming. Stay tuned!