data compression technology

[dancingtiger]

I am not a tech head, but have become involved in helping a start up company find some money for an invention they have. The inventor has data compression up to 500 times lostless compression. Is this unheard of? I have heard that the industry standard for lostless compression is 2:1. This company has a patent and are needing more money to get a couple more patents before they let loose with the product. So, I want to get through the learning curve a bit on this. How important is data compression? Is it possible they do have something that can produce lostless compression 500 times? Does anyone want to see a business plan?

[Mntsnow]

Sure I would like to see a business plan so I can see what they are trying to do. But personally I'm very doubtful they will be able to give 500:1 compression technology at this point in time.

[OuTpaTienT]

500:1 compression is unheard of. And I don't see how it could even be possible...but stranger things have happened.

What type of data is the compression used on? If it's used for raw data files and/or text, then it better darn well be lossless, any data loss is not even an option.

If it's made to be used for things like graphics & images then something with that type of compression ratio would be a Godsend. But I just don't believe it...I want to believe it...just don't think it physically possible. Currently it seems that JPEG is the choice compression method for images. On a good day, with the right image, and with some loss, it can get 10:1 and better, but the loss quickly becomes visible at higher compression ratios.

And he's claiming 500:1 with NO LOSS?, sorry I just don't buy it. (and I even believe in cold-fusion!)

[Win_98]

certain thing cannot be compressed at all such as mp3, jpg, and gif as it already compressed. so why compressed them if they already are, the time it take to do it just a complete waste as it creating another file taking up more space.

[BBA]

Hmmm...well, it's not far off.

I have a friend who HAS applied for patent already on a compression technique...well, lets just say I was connected to his HOME computer (he has a 2-way cable connection) And I was on a SDSL at Fraglan (http://www.fraglan.com) at the time...

I was watching a DVD movie playing realtime from his PC...on my PC over that internet connection. You could not even tell it wasn't being played from my DVD drive.

When you see that kind of streaming from less than 300K/sec, you kinda believe in MAJOR compression actually works!

Besides, look at the little download for the gizmo screen saver...102K turns into 13M...you saw it yourself.

[OuTpaTienT]

Sure, you can get outrageous compression with moving pictures because so many pixels stay the same from one frame to the next. Therefore those pixels that don't change also don't need to be reprocessed until they do change.

Depending on the video content (of course) I can easily see 500:1 compression of flicks. But I would contend that you will never see that kind of compression with individual images.

[dancingtiger]

It's called the MINC (minimiziing information code), and in the business
breifing it claims lossless compression 100:1. All competition is at 2:1. My patient and friend
Dan Creed has seen it work at better than 500:1. It turned out to not be
rocket science either. There is only 2000 lines of code and it can fit
anywhere. The developers stepped out of the box.

The secret to the MINC data compression system is in the power of
sequencing. MINC proves that random, binary data can be restructured in a
way that creates the key elements necessary to compress that random data.
This is achieved through sequencing.

"With sequences, structure replaces randomness, order replaces chaos, and
mountainous data sets are reduced to small managable groups of data."
-Evan Al-Chokhachy, MINC Project Manager

In the 3 bit world there are eight permutations of random data. In the MINC
world there are only 2. As the bit length grows so grow the permutations of
random data creating more to handle and more to compress. But with MINC,
using the equation of N-1 the permutation remains very small and very
manageable and always less than the bit length.


In addition, while compression by traditional standards is achieved with a
single iteration, MINC can compress multiple iterations dictated by the
application. Thus compression ratios never thought possible can be achieved
on any file type.

A word document in the industry can be compressed 3.2:1. MINC does it 100:1
in 34 cycles. Excell spreadsheet is 2.7:1 in the industry and MINC does it
200:1 in 46 cycles. MPEG and ZIP are "impossible" by current industry
standards and MINC compresses them 8:1 in 24 cycles.

Concerning the guiding principles of the MINC, the underlying cornerstone of
its architecture is the concept of Sequences. Embodied within sequences is
a revolutionary strategy for dealing with binary data, one which reduces
what is an otherwise unmanageable world of random data to a greatly reduced
set of data which exhibits multiple ordering and repeating behaviors.
Reduction of raw data via this concept is dramatic and unprecedented.

In the random data world of 16 bit data streams, there are over 65,000
possible permutations. Within the confines of the MINC domain there are
only 15 such 16 bit combinations which need to be considered. Additionally,
the MINC architecture provides another distinct and important capability,
that of being cyclical in scope. Once fully implemented, compressed data
from any given compression step can be fed recursively as input to a
subsequent cycle, thereby achieving compression levels far exceeding that
which is obtainable with todays algorithms. Feeding the compressed data to
a subsequent cycle can be done as soon as it is derived, thereby enabling
"multiple-threaded" and/or "parallell-pipe" implementations of the
algorithm.
Pretty neet huh.

[reddog4629]

Holy sequences-batman!!!

[Szech]

That sounds really interesting. I don't doubt that there will be some revolutionary breakthrough in compression. Heck, look at MP3's. I remember I recorded one song in WAV format back in the day, and even at lowest quality monoraul, it took up 40 megs. Who knew that a program would be able to take out the stuff you don't hear anyway? Best of luck to this company, and I hope you knock PKZIP off its pedistal.

~~Edit: I once got about 3000:1 compression. Don't believe me? It was 2000 text files all 1 meg in size that contained the text "You're screwed now!" over and over. I zipped them up, and made the zip file executable (zip2exe), and renamed the file chat.exe. So if anyone downloaded it and ran it, 2000 megs of their hard drive would be gone like (snap) that! Hee hee. I'm such a jerk. Anyway, 2 gigs of text compressed to less than 1000 bytes.

[This message has been edited by Szech (edited 10-13-2000).]

[crockett]

what the heck type of compression did that gizmo screen saver use anyay.Did anyone ever figure that one out?

[superraton01]

remember Stacker? Id like to see a
business plan! =)