Re: Anyone wanna help with a compression routine (new type)
- From: WM <mueckenh@xxxxxxxxxxxxxxxxx>
- Date: Thu, 6 Dec 2007 03:06:40 -0800 (PST)
On 28 Nov., 22:30, Ian Parker <ianpark...@xxxxxxxxx> wrote:
Taking another fairly trivial example. If we were to present movies in
the form of flash, that is to say individual objects moving, the data
will be very much more compressed than it is in conventional MPEG.
This idea could also be useful for compressing sets of numbers (which
is my main interest: numbers and their "addressability" or
"identifiability": A number which, in principle, cannot be addressed
by the bits available IMHO do not deserve the name "number").
To apply the flash method: If we have a natural number the bit
sequence of which consists of a noisy part A = a_1, a_2, ..., a_n with
n bits and a part of low entropy like B = 1,1,1,..., 1, 1, 1 with m
1's, then we can compress the noisy part A (as much as possible) and
add part B. All numbers of the form
B, a_1, a_2, a_3, ..., a_n
a_1, B, a_2, a_3, ..., a_n
a_1, a_2, B, a_3, ..., a_n
....
a_1, a_2, a_3, ..., a_n, B
can be compressed by (compressing A and) very few parameters.
I am curious to know whether such or related approaches exist already.
Regards, WM
.
- Prev by Date: Re: Anyone wanna help with a compression routine (new type)
- Next by Date: Re: Anyone wanna help with a compression routine (new type)
- Previous by thread: Re: Anyone wanna help with a compression routine (new type)
- Next by thread: Re: Anyone wanna help with a compression routine (new type)
- Index(es):
Relevant Pages
|
Loading
