It looks like you're using an Ad Blocker.
Please white-list or disable AboveTopSecret.com in your ad-blocking tool.
Thank you.
Some features of ATS will be disabled while you continue to use an ad-blocker.
Originally posted by thiefriverfalls
reply to post by AmberLeaf
. i see a cylindrical rock casting a shadow. rock. ROOOOOOOOOOOOOOOCK. lol
Originally posted by senselessness
reply to post by AmberLeaf
I am a software engineer, and I have written custom interpolation algorithms, and also implemented well documented interpolation algorithms, and even created image averaging algorithms for custom applications.
None of the images you have shown in this topic were scaled using nearest-neighbor interpolation, and instead, were scaled using other forms of interpolation. That means the images you are looking at are completely computer generated. Any details you think you see in the images are potentially non-existent in reality.
Unless you have 10 or more images of the same area, stabilized them all so they are all aligned, and then averaged each pixel of every image together, you will never get more detail. Even then, with image averaging you don't get a whole lot of extra detail. Image averaging removes noise, and that reduced noise clarifies edges and certain details because it reduces the signal to noise ratio (SNR), but not by very much.
edit on 1-9-2012 by senselessness because: (no reason given)
I am trying to explain to you that the "things" you think you see in the image are not actually in the image. Not only is there visible compression in the RAW images, but when you "zoom / scale / resize" the image with the methods you are using, you are turning these compression artifacts into things they are not, and you are creating illusions of things that don't exist.
More sophisticated interpolation schemes try to make a better guess at what would be in those extra pixels if they had been present in the original image. This tends to involve deriving a mathematical function which matches the original image, and then using that function to fill in the blanks. Biiinear and bicubic interpolation are the most common. As it happens, Windows Picture and Fax Viewer does use interpolation. (At least it does on Windows XP - what are you using?) Not sure whether it's bilinear, bicubic or something else, but it's definitely not doing straight pixel replication.
Originally posted by senselessness
reply to post by FlySolo
Sorry, you have to be incredibly ignorant to think a dark area can only be a hole, and not just a shadow. You are looking at a greyscale 2D image, how do you know it is a hole and not a shadow? Oh that's right, you don't know.
edit on 1-9-2012 by senselessness because: (no reason given)
No. Interpolation begins as soon as you "zoom" beyond the native resolution of the image. The type of distortion of the original data depends on the type of interpolation. A pixel resize does not "add" any information, any other algorithm does.
But what you're not getting is, there is a limit before straight pixelation replication occurs and what ever your looking at becomes distorted beyond repair.
While yours and Phage's argument is, without a sophisticated interpolation algorithm, it's going to be junk. As it turns out, Windows picture viewer doesn't use straight pixelation but actually does work on a more sophisticated level.
Mars must be made of Swiss cheese.
Originally posted by FlySolo
reply to post by senselessness
No one is saying interpolation doesn't happen, just at what point?
Is it happening here?
Or is it happening here?
How about now?
Would you feel safe to say it's not? All three pictures have the same detail and hasn't changed at all.edit on 1-9-2012 by FlySolo because: (no reason given)
Using your "correct" scale I can still see the hole. Nothing has changed. Just more pixelation. But you would have to be blind or in denial not to see all the debatable points are still visible.