Groups | Search | Server Info | Keyboard shortcuts | Login | Register [http] [https] [nntp] [nntps]


Groups > comp.compression > #1762

Re: Random Compression For Files Of Known Size

From Robert Wessel <robertwessel2@yahoo.com>
Newsgroups comp.compression
Subject Re: Random Compression For Files Of Known Size
Message-ID <qkr3i8tdohjpmjp7eckojcvcfddcssa3f2@4ax.com> (permalink)
References <d4625bb1-ff5d-4943-a007-9b8c0cc26745@googlegroups.com> <f5a0fa85-fdb2-4bf0-9c65-84d9a208d3b2@googlegroups.com> <3kqvh8d66j80t6ruf7pdk4r12drprdf91m@4ax.com> <e88b31bf-13b3-4211-b0a8-13c0aa79f9ac@googlegroups.com>
Organization Forte Inc. http://www.forteinc.com/apn/
Date 2013-02-18 03:08 -0600

Show all headers | View raw


On Sun, 17 Feb 2013 20:40:59 -0800 (PST), Fibonacci Code
<anglikai@gmail.com> wrote:

>On Sunday, 17 February 2013 04:27:09 UTC+8, robert...@yahoo.com  wrote:
>> On Sat, 16 Feb 2013 08:33:22 -0800 (PST), Fibonacci Code
>> 
>> <anglikai@gmail.com> wrote:
>> 
>> 
>> 
>> 
>> 
>> Please don't top post.  Fixed.
>> 
>> 
>> 
>> 
>> 
>> >On Wednesday, 23 January 2013 09:32:48 UTC+8, Geoffrey Sangston  wrote:
>> 
>> >> This method is clearly not feasible haha. It attempts to at least leave files the same size after compression. A drawback is both the compressor and decompressor must know the original file size.
>> 
>> >> 
>> 
>> >> 
>> 
>> >> 
>> 
>> >> Take the following example (where the LHS represents the binary string for an entire file):
>> 
>> >> 
>> 
>> >> 
>> 
>> >> 
>> 
>> >> 000 - 0
>> 
>> >> 
>> 
>> >> 001 - 1
>> 
>> >> 
>> 
>> >> 010 - 00
>> 
>> >> 
>> 
>> >> 011 - 01
>> 
>> >> 
>> 
>> >> 100 - 10
>> 
>> >> 
>> 
>> >> 101 - 11
>> 
>> >> 
>> 
>> >> 110 - 000
>> 
>> >> 
>> 
>> >> 111 - 001
>> 
>> >> 
>> 
>> >> 
>> 
>> >> 
>> 
>> >> Bit strings in the following set S(0,1,00,01,10,11,000,001,010,...) can be assigned to the decimal numbers D(0,1,2,3,4,5,6,...) by the following bijection:
>> 
>> >> 
>> 
>> >> 
>> 
>> >> 
>> 
>> >> Let n be the length of a word S and m be the decimal equivalent of that word. Then the paired decimal in D d = 2^n + m - 2.
>> 
>> >> 
>> 
>> >> 
>> 
>> >> 
>> 
>> >> To convert back from d:
>> 
>> >> 
>> 
>> >> 
>> 
>> >> 
>> 
>> >> n = leastInteger(log_2(d+2))
>> 
>> >> 
>> 
>> >> m = k - 2^n + 2
>> 
>> >> 
>> 
>> >> 
>> 
>> >> 
>> 
>> >> 
>> 
>> >> 
>> 
>> >> Now if we have a file of 100 bits then we can convert it to 99 bits with roughly  (but slightly less than) 50% probability, 99 bits with roughly (again less than) 25% probability, 98 bits with roughly 12.5% probability, etc...
>> 
>> >> 
>> 
>> >> 
>> 
>> >> 
>> 
>> >> This method only fails to compress the bit string in 2 elements of the total and compresses roughly 2 bits on the average. You also have to compute 2^1000 if you want to store a kilobyte haha
>> 
>> 
>> 
>> 
>> 
>> >Please continue work on this method,
>> 
>> >
>> 
>> >I have a complimentary algorithm that just require 50% of files to be reduced by 1 bit then we can achieve recursive compression.
>> 
>> 
>> 
>> 
>> 
>> And...  you can't.  You can get arbitrarily close to reducing 50% of
>> 
>> the files by one bit (IOW, 49.9999%, etc.), but you can't get to 50%.
>> 
>> And the closer you get to 50%, the closer to *doubling* the size of
>> 
>> the other 50% of the files you get.
>
>No, the method I found is a recursive compression technique by itself. It is bijective and have a properties of maximum increase 1 bit in any random file but compress the other. If this length removal method could reduce 1 bit at around 50% of the time, by layering the method alternatively, then recursive compression will have no problem.


If you have a maximum increase of one bit, you can make at most a
quarter of the files one bit smaller.

Back to comp.compression | Previous | NextPrevious in thread | Find similar | Unroll thread


Thread

Random Compression For Files Of Known Size Geoffrey Sangston <geoffrey.sangston@gmail.com> - 2013-01-22 17:32 -0800
  Re: Random Compression For Files Of Known Size Robert Wessel <robertwessel2@yahoo.com> - 2013-01-23 01:09 -0600
    Re: Random Compression For Files Of Known Size Geoffrey Sangston <geoffrey.sangston@gmail.com> - 2013-01-23 15:00 -0800
      Re: Random Compression For Files Of Known Size Robert Wessel <robertwessel2@yahoo.com> - 2013-01-23 17:27 -0600
        Re: Random Compression For Files Of Known Size Geoffrey Sangston <geoffrey.sangston@gmail.com> - 2013-01-23 15:56 -0800
          Re: Random Compression For Files Of Known Size Geoffrey Sangston <geoffrey.sangston@gmail.com> - 2013-01-23 16:21 -0800
            Re: Random Compression For Files Of Known Size Thomas Richter <thor@math.tu-berlin.de> - 2013-01-24 01:49 +0100
              Re: Random Compression For Files Of Known Size Geoffrey Sangston <geoffrey.sangston@gmail.com> - 2013-01-23 17:42 -0800
                Re: Random Compression For Files Of Known Size Robert Wessel <robertwessel2@yahoo.com> - 2013-01-23 20:05 -0600
                Re: Random Compression For Files Of Known Size Geoffrey Sangston <geoffrey.sangston@gmail.com> - 2013-01-23 18:12 -0800
          Re: Random Compression For Files Of Known Size Robert Wessel <robertwessel2@yahoo.com> - 2013-01-23 19:11 -0600
            Re: Random Compression For Files Of Known Size Geoffrey Sangston <geoffrey.sangston@gmail.com> - 2013-01-23 17:45 -0800
              Re: Random Compression For Files Of Known Size Geoffrey Sangston <geoffrey.sangston@gmail.com> - 2013-01-23 17:51 -0800
    Re: Random Compression For Files Of Known Size Geoffrey Sangston <geoffrey.sangston@gmail.com> - 2013-01-23 15:06 -0800
      Re: Random Compression For Files Of Known Size Thomas Richter <thor@math.tu-berlin.de> - 2013-01-24 00:25 +0100
  Re: Random Compression For Files Of Known Size Geoffrey Sangston <geoffrey.sangston@gmail.com> - 2013-01-23 18:46 -0800
    Re: Random Compression For Files Of Known Size Fibonacci Code <anglikai@gmail.com> - 2013-02-17 20:37 -0800
    Re: Random Compression For Files Of Known Size SG <s.gesemann@gmail.com> - 2013-02-18 05:38 -0800
  Re: Random Compression For Files Of Known Size Fibonacci Code <anglikai@gmail.com> - 2013-02-16 08:33 -0800
    Re: Random Compression For Files Of Known Size Robert Wessel <robertwessel2@yahoo.com> - 2013-02-16 14:27 -0600
      Re: Random Compression For Files Of Known Size "George Johnson" <matrix29@charter.net> - 2013-02-17 00:15 -0500
        Re: Random Compression For Files Of Known Size Robert Wessel <robertwessel2@yahoo.com> - 2013-02-17 01:58 -0600
      Re: Random Compression For Files Of Known Size Fibonacci Code <anglikai@gmail.com> - 2013-02-17 20:40 -0800
        Re: Random Compression For Files Of Known Size Robert Wessel <robertwessel2@yahoo.com> - 2013-02-18 03:08 -0600

csiph-web