Agreed, 10Mbps is likely where it'll be, probably less. If I were to guess, they are going to apply a lot of smarts to texture mapping to avoid needing a lot of bandwidth.
Doubtful, IMO. The fact that you need it to be both extremely low latency and with essentially zero compression artifacts (probably lossless, based on their goal). If the numbers listed here[1] are correct, then the most efficient lossy codec at that time was doing 100 minutes of 8k video in ~37GB of data. From that we can intuit that it was using an average of 50 Mbps for that 100 minutes of video. For the most efficient codec, and I'm not even sure from that whether it was using lossy or lossless numbers, because apparently HVEC can do both (but I would assume lossy, since it's about streaming video in that case).
You can't do weird texture mapping or lossy compression and expect people to really seem like they are there. Even if you don't notice that stuff normally watching a video, I think you'll notice it when you're interacting like someone is really in front of you, and that will throw off the immersion.
HVEC is no longer king of the hill when it comes to compression efficiency. AV1 and the upcoming VVC do better.
That said, my intuition is they aren't doing a pure video encoding solution. The fact that they talk about 3d modeling leads me to believe they are doing a combination of model + texture to get the realistic results. That would significantly decrease the amount of bandwidth and computational power needed. Over a low bandwidth situation you'd simply need to send model updates and do some smart interpolation to determine what things should look like.
Similar to the concept that playing a 3d game requires MB of resources but recording the same game at 8k would require a boatload more memory.
My assumption is they are using LIDAR to get a good model, high quality cameras to texture things, and a nice AI to stitch things together and interpolate when data isn't arriving fast enough.
I think this is what will be done in the future. You will interact with the camera for 15 minutes or so and it will create a custom compreasion algorithm for you.
It heavily depends on what encoder and config you use. In my experience, 20Mbps (or less) HEVC by Turing NVEnc realtime encoder is enough for 1080 60p. Also halving frame rate won't halve bitrate because of how video compressed by reference frame. Also video meeting won't move pictures as much as FPS gaming.
