I wouldn't say Skipper isn't in it for the notoreity. Just because he isn't as well connected as Richard Hoagland doesn't mean attention isn't what he seeks.
"The government has been deceiving us, and I alone have the knowledge to reveal this," is a well-used theme in pseudoscience. On its face that sets up a remarkably high burden of proof, and Skipper simply hasn't met it. Claims of highly evolved infrastructure on Mars are fantastic and require much more than "kinda looks like" lines of reasoning, especially when the material looks very much to trained experts like ordinary digital photo artifacts.
In addition to JPEG (i.e., DCT) compression effects, there are also error-diffusion effects visible. Error diffusion is a method of spreading quantization errors over the spatial frequency domain in order to minimize their effects -- in a word: dithering.
Returning images by radio from space means reducing the number of bits needed to express each image. Home computer users understand a similar problem by setting up their displays to 16-bit, 24-bit, or 32-bit color. They see that the smaller numbers of bits (per pixel) can produce visible banding and other effects. Quantization of incoming analog information is always problematic, but being able to diffuse the "error" in such a system over more than one pixel helps.
In the error-diffusion method, the "error" (i.e, the difference between the pixel's true intensity in the camera and the closest intensity available in the bit representation) propagates to nearby pixels, often preferentially down and to the right -- in scan order. Stated differently, the measured intensity of a pixel is combined with the errors in adjacent pixels, which often pushes the represented intensity up or down to the next value. The outcome is that the total error for some patch of pixels tends to average to zero.
But there are still discontinuities in the data. Error-diffusion methods produced discontinuous boundaries that are not rectangular or checkerboarded, as DCT discontinuity boundaries often are. They can be completely random because they depend solely on the whims of the incoming analog intensity.
The problem with Skipper's method doesn't really depend on how the discontinuities come about. We mention the different sources because each one leaves its signature on the quantized data. Skipper's error is made when he artificially expands the contrast between adjacent zones of quantization, and the draws conclusions and makes interpretations that require the arbitrarily-manipulated degree of contrast to be the natural degree. You can't argue there's a shadow there when you created the shadow. The most important thing to realize when using image-processing tools is when your effect has been revealed by the tool versus created.
So Skipper invents a purely conjectural scenario in which his haphazard manipulation of the contrast slider has "undone" some unknown, unevidenced concealment strategy he speculates was introduced by "The Government" to hide the detail. And because most readers don't understand how digital imaging works, they don't recognize or appreciate the far more parsimonious and evident conclusion that Skipper -- like so many other charlatans -- is simply exposing the ordinary properties of digital photography. Those artifacts are supposed to be there.
The inventors of those digital photographic techniques did not intend for them to survive gross molestation in Photoshop. JPEG compression and error diffision work because they ordinarily produce very tiny discontinuities that escape unaided observation. Expanding those to absurd and easily-noticed proportions does not all of a sudden create farm fields and factories any more than electron microscopy actually turns the surface of a compact disc into a waste treatment plant.
|