...you tell me to go study photography, implying that the picture still in the video at 10;11 is ok. But it is not. You can compress or unzip as often you like, you won`t see such an anomaly around a single object within a video, while having no similar patterns elsewhere.
This is an invalid apples-vs-images comparison. Compress and unzip are
lossless compression algorithms. What originally goes in is exactly what comes out later.
But compress and zip don't work too well (if at all) on pictures, so we use entirely different forms of compression on them.
Lossy compression -- what comes out is
not exactly what goes in, but only something that looks very much like it provided you don't look too closely. In exchange for this slight inaccuracy, you get rid of 90% or more of the original image data. Most people consider that acceptable, and that's why JPG is so popular.
JPG works by first dividing the picture up into little squares, usually 8x8 pixels, and analyzing each one independently. Its average brightness is measured with fairly high accuracy. But the detail in the square is a different story. It is first processed with a mathematical algorithm called the
discrete cosine transform, which converts it to a 2-dimensional frequency domain. Essentially it applies a set of reference patterns resembling tiny checkerboards of various sizes and shapes and says "how much does the picture look like
this one?" It provides a list of
coefficients that, when combined with the checkerboards, gives the original pixel values.
So far this isn't lossy. You could reverse all the steps and get exactly the original picture back. But what happens next
is lossy. The coefficients are quantized, that is, their accuracy is intentionally degraded by throwing away their least significant bits. Many of them go to zero. Then the remaining coefficents are efficiently packed into the JPG file for storage and transmission.
The decompressor reverses these steps, but because the coefficients have been quantized it will regenerate pixels that won't be exactly the same as the originals. Close, but not exact. If you look very closely at them, more closely than people normally look at pictures, you'll see the effects. You are especially likely to see sharp lines between the 8x8 squares, an artifact often called
blocking. With JPG this is especially common near and around sharp edges of objects -- just like the edge of Neil Armstrong's bright white suit and the black sky behind him.
So what you are seeing here are nothing more than normal, expected JPG compression artifacts. If you looked at the original digital data from the scanner, prior to JPG compression, or rescanned the original film yourself, you wouldn't see them. But the files would be huge, which is why the generally available versions of these pictures have all been JPG compressed.
