AI Pre Amp Post Processing

What is AI pre / post processing ?

Get raw data into the system; push real time data back out. Output from the raw video into projects.

Use cases: StreamYard? - running AI processing at scale large amounts of video. Client side upscaling Use AI to upscale and provide better content to the user. Nvidia works for facial transforms. Deep fake prevention and content tracing. Audio transcription Facetime 720p max size, Upscale content on the receiving side for WebRTC?.

Challenges: it's expensive to run even whisper on large volumes of video. AI upscale challenges with DRM Decoded the asset; Maybe the browser vendors will add it directly.

LivePeer? - distributed compute Gateways on the network for AI processing. Risk data ownership & privacy. The new Macs can run big LLMs?; how will segmenting work?

Using AI to do translation; loses things in translation if someone is not aware of the nuances. Maybe good for an initial pass.

Adobe is making some progress on content identification Deep fake prevention. Cryptographic ways to deal with content hashing. Voice cloning over dub people, New Stations streaming with webRTC ; source material authenticity. Dealing with user generated content and controlling that. A lot of similarities to malware battles.

Action items: Continue to see if there is adoption of AI upscaling; then worry about DRM after Library of LLMs? to test against for deep fake preventions