Science > Physics > data compression and expansion: SPONGE method by dedanoe
| Topic: |
Science > Physics |
| User: |
"dedanoe" |
| Date: |
17 Feb 2006 06:33:51 AM |
| Object: |
data compression and expansion: SPONGE method by dedanoe |
(X1,X2,1,1)x(1,1,y1,y2)=(n1,1,1,1)x(1,m1,m2,m3)
The idea is the vector x to be constant of the method, the vector y to
be the input, the vector n to be the compression and the vector m to be
the expansion. The method follows certain physical sense of compression
and expansion: In optimally relaxed situation the sponge has force x
and distance y while if we compress the force to n we must expand the
distance to m. The method works with four unknowns both left and right
connected with four equations. That insures certain linear projection.
We obtain those equations this way:
....
read it here: http://dedanoe.tripod.com/SPONGE.doc
http://dedanoe.tripod.com
dedanoe wants heavy medal
.
|
|
| User: "Puppet_Sock" |
|
| Title: Re: data compression and expansion: SPONGE method by dedanoe |
17 Feb 2006 11:40:32 AM |
|
|
dedanoe wrote:
(X1,X2,1,1)x(1,1,y1,y2)=(n1,1,1,1)x(1,m1,m2,m3)
The idea is the vector x to be constant of the method, the vector y to
be the input, the vector n to be the compression and the vector m to be
the expansion.
[snip]
Can you explain your notation a bit?
Also, can you explain how two inputs (y1 and y2) becoming
four outputs (n1, m1, m2, and m3) is in any sense compression?
Socks
.
|
|
|
| User: "dedanoe" |
|
| Title: Re: data compression and expansion: SPONGE method by dedanoe |
18 Feb 2006 05:02:44 AM |
|
|
you have in the .doc a system of four equations with four unknowns. the
unknowns are (n1, m1, m2, m3) and the knowns are (x1, x2, y1, y2). i
have been working on it lately and i get that the mothod will work with
all the accuracy. say we start with a1, a2, a3, a4 to get b1 to b6.
then we go further to get c1, c2, c3. if we set the equations ci=ai for
i=1,2,3 then we get a4, y1, y2 as functions of a1, a2, a3 in exactly
three equations.
http://dedanoe.tripod.com
.
|
|
|
| User: "Puppet_Sock" |
|
| Title: Re: data compression and expansion: SPONGE method by dedanoe |
20 Feb 2006 08:51:38 AM |
|
|
dedanoe wrote:
you have in the .doc a system of four equations with four unknowns. the
unknowns are (n1, m1, m2, m3) and the knowns are (x1, x2, y1, y2). i
have been working on it lately and i get that the mothod will work with
all the accuracy. say we start with a1, a2, a3, a4 to get b1 to b6.
then we go further to get c1, c2, c3. if we set the equations ci=ai for
i=1,2,3 then we get a4, y1, y2 as functions of a1, a2, a3 in exactly
three equations.
You may think you answered the questions. You simply mumbled.
dedanoe wrote:
(X1,X2,1,1)x(1,1,y1,y2)=(n1,1,1,1)x(1,m1,m2,m3)
Can you explain your notation a bit? You seem to have
a vector multiplied by a vector. What do you mean by this?
Also, can you explain how two inputs, y1 and y2, becoming
four outputs, n1, m1, m2, and m3, is in any sense compression?
Socks
.
|
|
|
| User: "dedanoe" |
|
| Title: Re: data compression and expansion: SPONGE method by dedanoe |
22 Feb 2006 08:27:26 AM |
|
|
the answer is out here:
http://dedanoe.tripod.com/sponge
http://dedanoe.tripod.com/sponge/spn.zip
and the usual: http://dedanoe.tripod.com
.
|
|
|
| User: "Puppet_Sock" |
|
| Title: Re: data compression and expansion: SPONGE method by dedanoe |
22 Feb 2006 09:56:51 AM |
|
|
dedanoe wrote:
the answer is out here:
http://dedanoe.tripod.com/sponge
No it's not. You never explained your notation. Nor did you
explain how two inputs becoming four outputs is in any
sense compression. Do you plan to explain this?
http://dedanoe.tripod.com/sponge/spn.zip
I don't download zips from untrusted sources. That's you.
Socks
.
|
|
|
| User: "dedanoe" |
|
| Title: Re: data compression and expansion: SPONGE method by dedanoe |
23 Feb 2006 05:52:11 AM |
|
|
ok, so you have:
A x Y = N x M
A has two variables and that's the original (a1,a2,1,1)
Y has two variables and it's free of choice (1,1,y1,y2)
N has one variable and that's the compression for A (n,1,1,1)
M has three variables and that's the expansion for A (1,m1,m2,m3)
here on google I can't express math well.
You'll just have to see what I have on
http://dedanoe.tripod.com/sponge
the file http://dedanoe.tripod.com/sponge/spn.zip
is program made in dellphi 7 and this is what it does:
1) compress the original then expand the compression
then compress the result then expand the last compression
in most of the case the calculus converges to some particular values.
2) works in reverse direction.
i almost forgot: http://dedanoe.tripod.com
.
|
|
|
| User: "Puppet_Sock" |
|
| Title: Re: data compression and expansion: SPONGE method by dedanoe |
23 Feb 2006 09:45:22 AM |
|
|
dedanoe wrote:
ok, so you have:
A x Y = N x M
A has two variables and that's the original (a1,a2,1,1)
Y has two variables and it's free of choice (1,1,y1,y2)
N has one variable and that's the compression for A (n,1,1,1)
M has three variables and that's the expansion for A (1,m1,m2,m3)
[snip]
Sigh. So, you are dedicated to pretending to be a rat-hole
down which I am supposed to pour water. Fine.
When you say
(a1,a2,1,1) x (1,1,y1,y2) = (n,1,1,1) x (1,m1,m2,m3)
what do you mean? There are many ways to combine two
vectors, several of which commonly use x as the symbol.
For example, there's the cross product. (Though it's not
likely a cross product for 4 element vectors.) What do you
mean by the x here? You could say *something* about this
notation. For example, one thing it *could* mean (though
there's no evidence I've seen that it does) is that your
really want a matrix instead of a vector. That is, by
(a1,a2,1,1) you really mean a two-by-two matrix
[ a1 a2 ]
[ 1 1 ]
and by x you mean matrix multiplication. There are various
other things you could mean by this x. But you don't say.
For example, you could mean
(a1,a2,1,1) x (1,1,y1,y2) = (a1 x 1, a2 x 1, 1 x y1, 1 x y2)
but that would not make a lot of sense regarding the rest
of your posts and web page. So, just what does that
notation mean? You seem to mean that a1, a2, etc., are
real numbers. What do you mean by this expression?
(a1,a2,1,1) x (1,1,y1,y2)
That is, show explicitly how you combine these two
vectors, and *ONLY* these two vectors, and what the
result is. For example, is the result a vector? Or is
it some other thing, say a single value?
The usual meaning of "compression" is that you manage
to send information in fewer bytes than were originally
used to express the message. As for example, "zip"
compression uses Huffman coding (see the wikipedia article
http://en.wikipedia.org/wiki/Huffman_coding for more).
The result is, you can transmit a smaller set of data, but
at the destination you can expand the data again and get
the original data back. Your scheme, what little you have
managed to present of it, seems to turn two numbers into
four. That is, you seem to get four numbers, all of which
are required to recover the original two numbers. So, this
in no sense seems to be compression. What is it you
think is gained through this transform? If you were to
transmit your "n" value, would it be possible, without
the m1, m2, m3 values, to recover a1 and a2? Even
supposing one knew the y1 and y2 values?
Socks
.
|
|
|
| User: "dedanoe" |
|
| Title: Re: data compression and expansion: SPONGE method by dedanoe |
23 Feb 2006 03:41:01 PM |
|
|
i see, by x i mean cross vector product of what is left and what is
right regarding that operation. say we have (a1,a2,1,1) x (1,1,y1,y2)
then as a result we get (k1,k2,k3,k4) where k1=a1-a2, k2=y1a2-1,
k3=y2-y1 and k4=1-a1y2. out of the particular original A combined with
the free of choice Y i get simultaneously the compression N and the
expansion M. i pick only whatever i need from N and M. in reverse order
i have two ways of recovery: 1) A x N / sqr(norm(N)) = Y x M /
sqr(norm(Y)) here i know M and Y to get A and N but Y has to meet
certain condition for success and 2) M x A / sqr(norm(M)) = Y x N /
sqr(norm(Y)) here i know N and Y to get A and M but Y must meet a
condition to return unique solution. once again
http://dedanoe.tripod.com/sponge
thanks for that Huffman site.
.
|
|
|
|
|
|
|
|
|
|
|
|
Related Articles |
|
|
