[Liquid!]

This is a new thread from the digital art forum at http://www.sijun.com/dhabih/ubb/Forum2/HTML/007012.html


Had a little bit of time, and couldn't resist to get the answer to the permutations question.

Anyhow, so I installed a c-compiler I found and wrote a small .exe file that I hope sheds some light on it. Keep in mind I haven't done this in a very long time, so I'm sure you could have written a more efficient and elgant program, however I hope this suffices.

What is the purpose of this program?

The purpose of this program is to calculate the possibilities of combinations that exist for a given target pixel percentage assuming that you attempt to double the size of the original picture. In other words, if a pixel was down rezed by 50%, thus a square of 4 pixels was averaged into one [the target pixel], how many 4 pixel combinations exist that would result in that one target pixels value?

What assumptions are being made?

There are a few very basic assumptions made regarding this.

1> Any inklimit greater than 375 is assumed to be indistinct from BLACK to the human eye and thus discarded.

2> Only integer values were used since, t=4 divided by 4 equals TARGET PIXEL 1%, however a t=6 divided by 4 equals 1.5% which in effect is either 2% or 1% depending on your rounding rules. For the sake of this query I'm only using multiples of 4 to arrive at even target pixel integers.

3>For all values of x[n] (the pixels that are being averaged x[1]-x[4]) must be no larger than 100, because you can't have more than 100% ink.

4> 100,50,0,0 is different than 100,0,50,0 and 0,100,50,0, and so forth. Each initial pixel is distinct since patterns must be followed.

Anyhow. The bottomline is that WITHOUT any intelligent analysis the highest amount of possibilities seems to exist at 46% ink in the target pixel at 543970 possibilities, which is very very low. Actually much lower than I anticipated.

Here's the table with all the other possibilities:

www.go-liquid.com/forum/perm.txt

and the c++ source code that generated this table:

www.go-liquid.com/forum/perm2.txt

I think both are pretty accurate, however if you find some errors, both in logic/syntax, please let me know.



-c

[This message has been edited by Liquid! (edited December 04, 2000).]

[This message has been edited by Liquid! (edited December 05, 2000).]

[A.Buttle]

Programs, shmograms. What I want are results!

------------------
So...which part of "get out of town freak" weren't you clear on, John?

Joe Dillingham

[email protected]
Three Times A Day

[[Shizo]]

Ye, me too!
I'm with A.Buttle on this one.

------------------
Nothing really matters to me

[Chapel]

You misspelled error.

MUAHAHAHAHAHAAAA!!

[Liquid!]

fixed... happy?!?

[Chapel]

I just thought it was ironic. Now that you have crossed over to the dark side of the forums.. you can get bombarded with useless replies and nobody can stop me... hahahaha!!

By the way.. I still say it isn't reasonable to use pixels for image enhancement. I've heard things about a different format that allows you to do it with the proper equipment though.

My company is building the next space shuttle by the way.

[[Shizo]]

MY company uses the first space shuttle (russian - Buran) to attract turists!! hehehe

[A.Buttle]

Who do you work for Chapel? JPL?

------------------
I'll just wait until you go to sleep tonight... then I'll cut you in half and count the rings.

Joe Dillingham
[email protected]
Three Times A Day

[Chapel]

LMCO

Who do you mean by JPL? Jet Propulsion Laboratories or that speaker company?

[This message has been edited by Chapel (edited December 06, 2000).]