Sharpening the Image and Blurring the Lines

Nvidia recently released a video on their YouTube channel highlighting the global illumination.  To illustrate what a great advancement global illumination is Nvidia used it to validate the moon landing.  One complaint from conspiracy theorist about the legitimacy of the moon landing is this photo.  In the photo Buzz Aldrin is climbing down the ladder of the moon lander in its shadow.  Conspiracy theorists claim there is a second light  source, a stage light, illuminating Buzz Aldrin.  Nvidia shows through the use of global illumination lighting in the photo is valid, even with only one light source.  Global illumination is a technique that allows light to be reflected off objects in the field.  In a traditional light engine the simulated light would stop at the target it hits.  A render with a traditional light engine would produce the image where Buzz Aldrin would be covered in shadow and not be visible.  Introducing global illumination allows the simulated light to bounce off reflective surfaces such as the Moons surface.  Through this technique Nvidia was able to recreate a nearly identical image with just one light source, proving that the photo of Buzz Aldrin was legitimate.

Computer graphics have been around less than a century and they can already create a highly realistic light field. I am curious, will computer graphics ever advance so far that the human eye could not tell the difference from the real world in real time.  There are already computer-generated images that look 100% real because they have been tampered with. Could this be done a three-dimensional world, like a video game. With a realistic light field already accomplished two other major obstacles is the resolution and the number of polygons creating the objects in an image.

The reolution of the human eye is difficult to define because the eye does’nt really see in a resolution. According to this video by Vsauce, in order to fool the human eye an image covering the entirety of the field of view would require just 8 megapixels.  This image would have to completely envelope the field of view and be able to track the motion of the eye. But now the real trick is filing that image.

In computer graphics objects are made up of small two-dimensional shapes called polygons.  These polygons fit together to create a three-dimensional image.  The higher the number of polygons the smoother and more realistic the object will look.  So how many polygons does it take to make an object look like the real deal?  This is hard because like the eye not seeing a resolution real life objects are not made of small polygons.  Actually maybe they are, at the molecular level. If a computer could render every single molecule the object would appear real.  This is obviously the upper bound on the required number of polygons.  The actual number is most likely much lower, but this shows that this is possible.

We can only speculate about these things but I believe within this century computers will be able create a generated world that is indistinguishable from the real thing.  I think I will explore this bulling of worlds in my next blog.