Comments on: How ‘bullet time’ will revolutionize exascale computing

By: Editor

Editor — Thu, 14 Feb 2013 11:54:21 +0000

It’s sampling in time and space (to reduce scientist analysis time — they can just analyze the samples, not the whole stream) rather than file compression, which would not retain the sample information. Like file compression, it also reduces file size. See the arXiv article.

By: Phillfrog

Phillfrog — Thu, 14 Feb 2013 10:58:03 +0000

Unless I’m missing something, this article is meaningless. We have been “compressing” data down for human consumption since computers were invented. Computing is all about taking data and computing it to get a result that has meaning to the user. This information is intrinsically smaller than the data it was derived from.

What this article seems to be implying is that the more data a system processes, the more needs to be communicated to the user. If this were true, installing a new hard drive would mean more data would start coming out of your monitor into your head! Or the bigger the Internet gets the more results have to be squeezed onto the first page of your google search.

Weird article.

By: eldras

eldras — Wed, 13 Feb 2013 15:27:43 +0000

You can reduce the data by compression maths…equations that inflate on demand ‘per piece’ as components.

A good analogy is a pocket map.

You have the world folded into a pocket sized book, but it can open at any area you wish.

On google maps you can see this is even better.

The data size problem is ENTIRELY solvable by components and library functions like the early Von Neuman machines.

Well I know what I mean anyhow:)

(The rule is: you dont calculate anything until you need it)

By: Snake Oil Baron

Snake Oil Baron — Tue, 12 Feb 2013 20:41:10 +0000

I missed something again. :-/
If the system is simulating something, why is there a need to simulate many camera angles? In a 3D video game simulations, the server doesn’t send what the environment looks like from every angle you may move to, does it? It sends you the info about the simulation (lots of which is a list of files on your oun computer) and your own computer figures out what it should show you based on how you move. I guess I am confused on where the calculation of the simulation–or data set or whatever–and the calculation of the point of view interface. Maybe I’ll just leave it to those who know about these things.

By: Gorden Russell

Gorden Russell — Tue, 12 Feb 2013 19:52:30 +0000

Right tim, they’re sure to do that, just as soon as they get the processing power.

By: WLGJR

WLGJR — Tue, 12 Feb 2013 18:36:51 +0000

IMO it would be great to apply what you mention in real life… we should find a way to create “super-multitasking” AIs or cyborgs (or special training methods to turn MOSHs into super-multitaskers) so that they can pay attention to and memorize the inputs from multiple cameras/earphones etc simultaneously.

This would greatly benefit surveillance/security, weather prediction and perhaps help with making STEM discoveries.

By: funkervogt

funkervogt — Tue, 12 Feb 2013 17:26:47 +0000

A lot of compression could be applied.

By: Walter Baltzley

Walter Baltzley — Tue, 12 Feb 2013 17:14:42 +0000

I believe that in the future, all movies are going to be shot in 3D…using multiple camera angles and “point clouds”…. allowing the viewer to move the camera around…and possibly view different areas in the movie…creating a multi-dimensional story experience.

Movie Directors could have a blast…filming each character separately, allowing the viewer to see the same story from each person’s perspective. Parts of the story could be taking place “behind the scenes” giving viewers reason to watch movies over and over again to catch little nuances they may have missed before.

By: DeBee Corley

DeBee Corley — Tue, 12 Feb 2013 17:10:58 +0000

This is nuts.
We are amassing hugh amounts of data, then writing software to mine the data. Of course, the miners are using preconceived notions to look for answers they expect.

Alas, the real data won’t be discovered until years later, by AI.

By: tim the realist

tim the realist — Tue, 12 Feb 2013 14:13:41 +0000

Need to combine this with the holodeck project for immersive computing