Or rather, admit defeat and just give up on the idea that you''ve come up with a new way of compressing data so that MP3''s quarter in size and zip files do the same.

Pythius

quote: Original post by LackOfKnack

I hope to someday study this, but not at the present.

That is a much better attitude towards the situation.

quote:
Kinda. So using 5 different compression programs, to you, is considered one algorithm?

Yes, your algorithm is to use 5 sub-algorithms. I consider that one algorithm.

quote:
What was I talking about? Window size? What I meant was if you have some bits from a random file:

1011 0100

But you break them into 3-bit bytes:

101 101 00...

Then now you have a pattern, but before you did not. That would depend on the particular file or sequence.

Yes, that's waht I was talking about.

quote:
I was trying to show that things are able to be compressed repeatedly. People keep saying stuff like "I could store an entire movie on my Zip drive! That, of course, is impossible." But they don't say why exactly.

Sorry, but a good compression algorithm like Huffman will not see any increases after one compression.

quote:
Seems to work on paper.

If I had a doller for every time that was said, and then the real-world implementation failed....

quote:
And so you're saying not every file is immediately compressible. That's where the data massage comes in.

Yes, that's what I'm saying. You know what, forget a compression algorithm. I'd like to see a data massaging and unmassaging algorithm that makes data more compressable.

Great, all this BS and all we get is a Huffman implementation. If your implementation cuts a mp3 by 2% I would be impressed.

All of you that still believe that data can be compressed so well should follow Lack's example and try your ideas out for yourself. Good luck,

Mike

Edited by - Vetinari on 4/19/00 6:25:55 PM

Hey kieren_j, about making the file more compressable, I was thinking, if text files usually only use the first 7 bits, you could take about 100 bytes, and order it like all the first bit, then all the second bits, then all the third bit, and so on and so on. Then you do Bit Run Length Encodeing, Is that what your doing, Id like to know how your doing Bit Run length encoding, if you are. ARE YOU?

How are people still posting to this thread while its closed?

-Dan Smith
dans@3dgamedev.com

From all this, I infer logically that Mr./Ms./Mrs. LackOfKanck has no life, and is basically a hired amateur so-called "lawyer" for kieren_j.

LackOfKanck--go get a life for once!! Eat some snickers, do the freedom Dance!! Get off that fat ass and go outside for the first time!!

kieren_j--you sound too good to be true. I''m NOT saying what you''re talking about is impossible, but I am still skeptical...

More peremptory arguments from...
Captain FREEDOM!!

people who have been in orbit (say, on a space station) are technically "time travelers". They are a few milliseconds younger than if they would have stayed on earth.

They''ve also took 2 very accurate clocks, flown one around the world very fast, and left one on earth. The one on earth was a (very small) fraction of a second faster, its true.

Im not going to say anything about compression, sorry. its just not my area to talk (heh like time travel is). none the less I''ve found this thread interesting....I hope it doesnt get killed.....

first i believe to kerian_j!
for what he need to lie? what he will get form it?

2. i believe him because he is not the only one who know ways to do awesome compression like kerian_j discover,
i know 2 guys who discover a way to compress something like 3GB to 200-300MB!!

and trust me , im not lie too!

Guys (and girls) I hate to tell you guys this but there realy is a way to compress ANY file (yes ANY!) into a very small hand-full of bytes (like 40-100 depending on size, randomness, and length of time to compress). This is known as fractal compression, I will show you how to do this in a sec but after I do you will see the reason it is not used.
First I must define what type of fractal is good for compression. The fractal type that I use is called an I.F.S. (iterated function system) to create this you use a set of variables that are inputed into a function repetedly, the output of the function is put back into the equation. now most of you are thinking WHAT!! how can a equation just POP out the result? thats the problem. you can get a very ordered and repetable result but how do you get the equation from the data? No one has ever figured out a way to generate the variables from the data. Also the functions that can generate the data may take MONTHS on a fast computer (I am talking supercomputer here). now a simple example.

here is the most common IFS equation to date.

Xn+1 = A1+A2*Xn+A3*Xn^2+A4*Xn*Yn+A5*Yn+A6*Yn^2
Yn+1 = A7+A8*Xn+A9*Xn^2+A10*Xn*Yn+A11*Yn+A12*Yn^2

Useing this we can use multiple sets of the variables to generate extreme flexiblity. note using JUST one equation and only counting a range for each variable of -1.2 to 1.2 in incroments of .1 generates (wait for it) 25^12. and of these about 10^15 have chaotic boundrys(good) and that is with only one equation with a small range. now you also must save the seed to the random number generator if you are using the monte carlo method (I didn''t make that name up), and also the number of times to run the equation.

good sources:

Fractal Horizans by Clifford A. Pickover
and my fave:
The Beauty Of Nature by Benoit Mandelbroit





I have always been lost!

please right to me if you wish to talk about this or ICQ me at ad_ice45@yahoo.com and 70233548

To quote a previous post on this thread,

"MadHack, I think you''ve been hitting the crack pipe a little too often."

Pythius