AI Video Compression for smoother video calls

This article is part of a series of blog posts that I wrote feeling inspired by my podcast conversation with Fabrizio Gramuglio about AI, VR and Brain-Computer-Interface (TCDP #28).

Nvidia is working on improving the quality of video calls using AI technology. The project is called Nvidia Maxine. The software creates a model of your fave using key points. On that basis it recreates your face. That way the video conference tool doesn’t have to aim at reproducing each pixel. The big advantage is that the bandwidth is lower – in the first example even much lower. Read below why this is interesting to me as a podcaster.

source video: AI-Powered Video Conferencing with NVIDIA Maxine

What I as a podcaster want to see: Reduced bandwidth for improved audio quality

In video calls that I record for my podcast often both the video and audio is not good but merely ok. So, in the post-production I have a lot of work to do in order to make a good podcast episode out of the audio and video material. I would love to have this choice in the settings of the video conference tool: make good audio quality a priority. I am willing to get worse video quality in order to get better audio.
The AI video compression software by Nvidia could help me with that – with the difference that both audio and video quality would be better.

Check out my conversation with Fabrizio Gramuglio about AI and more future technology:

Leave a Reply