Hello users, please give me detailed explanation...
I am looking for the BEST algorithm for compression of continuous data stream.
To be exact, it´s not pure continuous stream since characters "Q" are randomly (yes, specific pattern can be seen there, but far away) mixed with letter "q". But I suppose that for compression algorithms capitalization will be irrelevant and therefore I will be able to compress down my file with high ratio.
My sample is attached below.
Only letter "q" can be replaced with something else (yes, I´ve also tried with spaces and special characters)... but not with "ABCDEF0123456789" nor with "Q", since the aforementioned letter acts as a separator and the big letter represents... I don´t want to say what.
It´s not complicated to calculate all characters in text document and interpret them in much simplier form like 1320(Q) where number is the repetition of a given character... actually, the problem lies how to properly distinguish uppercase characters between lowercase characters.
I don´think so that some further simplification and thus achieving better compression is even possible... I expect at least 98% space saving.
As far as I know, RLE itself is limited to 255 repetitions since characters are interpreted as a bits.
Also, I have knowledges about more sophisticated neural-network powered compressors like PAQ, CMIX, EMMA and its derivatives (PAQ8PXPRE etc).
But I am afraid that there isn´t a compression algorithm that don´t distinguish between "Q" and "q" - it is handled as a another character no matter if it´s capital letter or not.
Furthermore, I am unable to use characters from 0-9 and A-F and also Q;q at all for simplification.
After decompression, the output must be EXACTLY SAME as input, of course.
Please DO NOT OPEN FILE IN NOTEPAD! It takes 20 min or so. Use another editor instead.
So, is there any algorithm or, better, software (WINDOWS GUI preferably) that´s capable to achieve that I want?
Thanks a lot for your help.