Merranvo: Always a strange trope when a game "in the future" has worse "in game" digital rendering than the game itself. I think the effect looks neat, but I can't help but laugh at the idea that rendering technology is so horrible in 2077 that you actually see pixels when editing video.

wingchild: I'd argue it's because "editor mode" detaches you from the video source (the BD rig on whoever's doing the filming). Arguably the things directly in their focus could still be sharp and pretty, but all the things out of direct line of sight would wind up invisible. Maybe going pixelated is the happy middle ground they settled on for the visuals.

Merranvo: Always a strange trope when a game "in the future" has worse "in game" digital rendering than the game itself. I think the effect looks neat, but I can't help but laugh at the idea that rendering technology is so horrible in 2077 that you actually see pixels when editing video.

Archonsod: It's not a video ...

Merranvo: It's just the trope to me, I mean even some of the hollow displays use a similar tactic. The devs want to make a visual distinction between "reality" and "virtual" and pixelation tends to be a favourite.

Of course, I think it would have been awesome if they used depth of field in this situation, instead of pixelating everything out it blurrs out. This would also allow some optically interesting effects, such as how what the BD recorder is focusing on would likely have more permanance than what you as the editor focus on. So even after the BD recorder moves on, she might be thinking about an object she was looking at before and that would remain in focus for a while longer.

I mean, I get it, but you could definately have some better fun with this than just pixelation.,

BlackKnightSix: The braindances are supposed to be depth information pulled from a single source (the person).

Similar to how Tesla cars' cameras can generate an idea of what is around them in 3D space since each pixel is judging depth.

youtu. be/HM23sjhtk4Q?t=156

BlackKnightSix: The braindances are supposed to be depth information pulled from a single source (the person).

Similar to how Tesla cars' cameras can generate an idea of what is around them in 3D space since each pixel is judging depth.

youtu. be/HM23sjhtk4Q?t=156

Thunderbringer: Oh... car autopilots... Those things are SCREAMING "accidents will happen". Because it doesn't matter how many radars, lidars, sonars, cameras, or any other possible sensors will you have - thing is software won't be able to react to a lot of unpredictable stuff. Like - a crash is unavoidable but bumping in one car means a few pizzas worth of damage while bumping into other means serious accident which can possibly kill a passenger or someone else, maybe even creating a chain of crashes. A human driver with some experience can see that and make a decision in less than 1/10 of a second.
Or when a kid gets almost right in front of your wheels and the only way to avoid killing him will be to crash into other car with severe damage but in that case everyone will be alive. Even making a computer to tell a water puddle from some super slippery oil or something isn't that easy. How to program a computer to drive dangerously and speed up instead of braking when a human sees easily that it's the best option to avoid collision or car spinning out of control on some ice or whatever. No computer can get even close to human intuition based on years of experience when it comes to dealing with unpredictable chaotic stuff.

And even for sensors - cities got a LOT of interferences. Even GPS signal can bounce off tall buildings sometimes and create echo.

And then, if you'll miraculously manage to make some super complex AI that will properly react to all of that without getting your car price into jet airliner range - oops... a simple error in the code. Calculating trajectory to Mars for an automatic spacecraft is waay easier than making a car to drive around a busy city. The code for the whole mission is waaay simplier. And yet... how many if those things failed because of simple error in software or data, like just trajectory data miscalculation or race condition or something, failing completely or partially?