A 10GB file is compressed to 3GB.No data is removed. No Quality lostWhat is actually being removed?
Imagine this text OP is a nigger nigger faggot nigger faggotNow an algorithm replaces nigger with 1aCompressed -> OP is 1a 1a faggot 1aA decompression algorithm knows to replace 1a with nigger In a nutshell that's how compression works
>>110180386Winrar removes the air molecules between the bytes
>>110180386>>110180455On the flip side, not all data can be losslessly compressed. Think of the compressed version a shorter description of the original data. If the shortest possible description is the data itself, it's incompressible.Random data can't be compressed, because there is no predictable pattern to it. If there was a pattern, it wouldn't be random.
>>110180514Even with random data, you could have a scheme which counted run lengths (successive zeros, ones, or any other pattern), and if the book keeping for these were smaller than what they represented, there would be savings. [Compressed Data Length][Compressed Data][Book keeping]
>>110180514>Random data can't be compressedWrong. Imagine a random number generator that outputs either 0 or 1. You can still run length compress its output unless it never gets runs of 2 or more, in which case it's not random.
>>110180386lol dude damn bean in gta?? fuuuuuck wow haha
>>110180386scientific notation
>>110180386Redundant information.
Think of it like 4chan. If you remove this thread, no data is removed and no quality is lost.
>>110180386Imagine a string of AAAAAAAAAAA that goes on for 1000 characters. A simple compression algorithm could compress that to "A 1000 times". If it was an assortment of AAAABBBABABBBBBBAAAABBBAAAABBBBBB it could compress that in various ways such as reducing it to a string of numbers like 4311164344. Basically, if something can be reduced down to a pattern or formula that yields the same result every time, it will be.>Random data can't be compressedSeeds.
>>110180386access speeds
>>110183510Sneeds.
>>110180386soul
>>110180455nice try autist
>>110180514tzd
>>110180575But that won't reduce its size because RLE requires some extra data to determine whether a particular number is repeated or not. You can only compress things with RLE if there's more repetitions than non-repetitions, in which case it's not random
>>110180455>Now an algorithm replaces nigger with 1awhat if there's a document where there's a graph named "Figure 1a", what happens when you unzip the file?
>>110190727Actually you're right. I kneel
>>110190997It’s replaced with Figure \1a and put back to Figure 1a on decompression.
>>110190997Go figure nigger.
>>110180386nothing, it is as such:>the man walked down the road>th ma wal do th ro>tmwdtrbut its done in software language, and is reverse decoded (additive) when extracted. then they make it work and then you can infinitely scale it
>>110180386size
