Tag Archives: accuracy

Abstract In this video you can see how moq-encoder-player (web based media encoder) sends video using MOQTv14 and MOQ-MIv03 via moxygen relay to a native CISCO player. Resulting latency ~290ms Youtube link

Abstract In this talk we will be talking about the complex media processing pipelines behind media AI models (translations, lipsyncing, text to video, etc). AI Models are very picky on media specs (resolutions, frame rates, sample rates, colorspace, etc), also most of the times the AI tools that you see that, for instance, generates aContinue reading “Demuxed 2025: The media processing pipelines behind AI”

Abstract Short interview where Allen Helton, Ecosystem Engineer at Momento, sits down with Jordi Cenzano, Software Engineer at Meta. Jordi previews his session on Meta’s AI video restyle, a feature that allows users to transform existing video content with AI-powered presets that change backgrounds, outfits, lighting, or even turning people into statues or animals. YoutubeContinue reading “Buffer free Lounge: The media processing pipelines behind AI”

Abstract Short interview where Allen Helton, Ecosystem Engineer at Momento, sits down with Jordi Cenzano, Software Engineer at Meta. Jordi previews his session on Meta’s AI video restyle, a feature that allows users to transform existing video content with AI-powered presets that change backgrounds, outfits, lighting, or even turning people into statues or animals. YoutubeContinue reading “Buffer free Lounge: The media processing pipelines behind AI”

Abstract We will focus on the challenges we faced from a media processing / scaling point of view, such as: inference latency and scheduling, voice isolation, media timing/alignment, alternate tracks delivery, instrumentation, model evaluation, etc. Youtube link Streaming tech Sweden

Abstract In this talk we will show how we implemented a media processing pipeline to perform (autodub / lipsync) media inference at Meta scale. We will focus on the challenges we faced from a media processing / scaling point of view, such as: inference latency and scheduling, voice isolation, media timing/alignment, alternate tracks delivery, instrumentation,Continue reading “Video@Scale 2024: Scaling AI translations at Meta”

Abstract Media over QUIC (MOQ) is an IETF proposal for a media transport protocol that involves several big players of online video space (Meta, Youtube, Cisco, Akamai, etc). Ideally MOQ can accommodate most the online media use cases we see today: To start proving some of the novel ideas of the MOQ working group weContinue reading “Demuxed 2023: MOQ”

Abstract Demo of an open sourced proof of concept we built to test MOQ ideas. Recorded demo (in a more reliable environtment): Demo video (4m)  Those are the links to the repos:

Abstract Demo of an open sourced proof of concept we built to test MOQ ideas. Recorded demo (in a more reliable environtment): Demo video (4m)  Those are the links to the repos:

Abstract At Meta we implemented RTMP Go Away, this is a new mechanism that allows the live RTMP server to send a signal to the client indicating that it needs to reconnect. This allows the client to create a new connection at a logical media boundary, incurring zero data loss in your live streams. ThisContinue reading “Sydney Video Nov 2022: RTMP GoAway (at Meta)”