Here, I'll do the little fiddling for you on the Tibet image. That will answer any question about compression artifacts and the blockies.
I adjust the levels first. Here is an enlarged, levels-adjust slice:
Here is an even more enlarged crop, with all the relevant artifacts marked:
Notice that pretty much the whole image is made up of little square or rectangular tiles. Some adjacent tiles have little patterns within them which
connect quite smoothly to each other. But the rectangular compression artifacts are everywhere. The image compression software thinks and processes in
rectangular portions of data, which can be represented by a matrix:
The two coordinates indicate a rectangular area which will be filled with a solid color, to further broken down into smaller rectangular portions
colored the same color. Some of these areas are a single pixel and would be represented by a single coordinate:
There is no simpler, more efficient, or effective way to compress images than using rectangles. OK, maybe there are better methods. One is by breaking
the image down into circles, which can be represented with three bits of data:
X,Y = Coordinate of center point, R = radius
You'll have to google yourself if you want to learn more, but it's really neat. I cannot remember the name of the method. Circles don't work well with
square pixels, though.
Another cool way to compress images which will not result in rectangular artifacts is
, but I don't really know anything about that.
The point is, four bits of data is a small amount of space for a compressed area of the image. This of course does not count the color of the image.
But it is very easy to process this type of compression, creates reasonably high quality images with a small filesize.
When there are changing colors from one pixel/area of the image to the next, you will see blocks. A photo of a sunset, for example, with all its
smooth color gradients, will have long strips of color horizontally when you toy with the image in photoshop. Normally there would be perhaps one
hundred different colors in a one hundred pixel thick strip of sunset. That would take up a lot of storage space, at least compared to the compressed
version. Those 100 different colors might only be a tiny bit different, and the computer will break them down into say 5 strips, each 20 pixels thick,
each a solid color.
Since I have nothing else to do and I've just had a nice smoke, let's do some math.
- Our sunset image is 640x480 pixels.
- One 100 px thick strip would be 640x100 pixels, area = 64000 px
- If each of those were a different color, there would be 2 integer coordinates (x,y) plus color data needed for each pixel. Each pixel would be an
individual rectangular unit of color (cause it's a square). That is a lot of data. This would be the uncompressed image.
Let's compress it, and say that each 640x1 px row within the 100px thick strip is all the same color. In a real photo, they would not be the same
color. It would vary slightly throughout. Anyways.
We'd have 100 individual rectangular units of color, each represented by 4 integer coordinates (x1,y1,x2,y2 - define the rectangle) and plus color
data (which is just one single color). This compressed strip of sunset requires ~1/320 the amount of data as the uncompressed sunset strip.
Further compression would make it even smaller. We're going to compress to the point where the image will be of bad quality. It will be totally
recognizable and still look pretty if it was a pretty sunset, but you will know it is low quality. Math:
- 640 x 100 px strip is divided into 5 strips of 20 px each
- We will have 5 individual rectangular color units (each having an area of 640x20=12800px), each represented by 4 integer coordinates and a single
color. That's a total of 20 integers and 5 colors needed to create the strip of sunset, versus 128000 integers and 64000 colors needed for the
uncompressed version which has a different color for each pixel.
If you see an image that ends in .jpg, it is compressed. Sometimes it will be compressed at a high quality and have a larger file size and hardly
discernible compression artifacts. But that's how it works, and nearly every digital image suffers from compression.
Hope that helps explain compression and compression artifacts and why they are there in every UFO photo, and every other photo too
edit: Compression artifacts are probably most commonly found surrounding objects which stand out against a relatively solid color background. Like
things in the air, or skylines. Good to know.
edit2: Oh yeah, why is it there on one image and not in the next? They are different images. It's different depending on the light that is received. I
could find you some compression artifacts in any compressed image, meaning any image. There are compression artifacts in the second one you posted, I
can see them without editing the image at all. Just look at your monitor from a different angle (if it's LCD that is).
edit on 10-1-2011 by
tetsuo because: (no reason given)