Compressions

jlliagre · 11-25-2004, 02:20 AM

Code:

I've heard 7zip is open source, so, if true, not a problem as far as platform.

Right, 7zip was actually ported to linux last week, I just found it out in http://sourceforge.net/projects/p7zip/

How are your images encoded ?

JanusPaul · 11-25-2004, 10:36 AM

When working with the actual image data, they'll be used in rawest form (BitMap). 16/24/32/-bit all included, then compressed when saved.

jlliagre · 11-25-2004, 11:11 AM

Are these images computer generated or real photos ?

JanusPaul · 11-25-2004, 07:47 PM

All the above o.O?

jlliagre · 11-26-2004, 02:40 AM

There's an lossless algorithm comparison at http://www.hpl.hp.com/loco/HPL-98-193R1.pdf

From the jpeg faq about 24 bit images:

Quote:

The best known lossless compression methods can compress such
data about 2:1 on average.

Why do you want lossless compression for images, which precisely are the kind of data that have very efficient lossy compression methods available ?

gamehack · 11-26-2004, 07:05 AM

Have you tried UHARC? It squezed 2 GB into a 300mb file... the compression/decompression indeed is very slow, but hey, you save a lot of space

For more info take a look at this site http://www.maximumcompression.com/programs.php

Cheers,
gamehack

Marius2 · 11-26-2004, 08:00 AM

Quote:

Originally posted by JanusPaul
The compression in a dailup modem DOES EXIST! And it's not going to be found in a 'driver' file. It's HARD BUILT into the hardware itself. When one modem sends data over the wire to another modem, one divides the data by the polynomial expression, sends the remainder, the receiving end applies the remainder back into the expression. There ya go, a perfect 4:1 lossless compression for any case! Awesome/Brilliant as it may seem, it's possible!

If you're talking 4:1 lossless reduction, the trick is probably within the modulation of the signal on the line
(that is, take 4 or 8 or 16 states of the signal instead of only 2 - number of states limited by line quality).
I don't think that a software algorithm can losslessly compress 4:1 (except for special cases), since usually
there isn't that much redundancy in typical data.

JanusPaul · 11-28-2004, 11:32 PM

Quote:

Originally posted by Marius2
If you're talking 4:1 lossless reduction, the trick is probably within the modulation of the signal on the line
(that is, take 4 or 8 or 16 states of the signal instead of only 2 - number of states limited by line quality).
I don't think that a software algorithm can losslessly compress 4:1 (except for special cases), since usually
there isn't that much redundancy in typical data.

4-bits = 16 states, a way to level out the data.

jlliagre · 11-29-2004, 01:24 AM

Quote:

4-bits = 16 states, a way to level out the data.

and 1 bit = 2 states, that's the basics, nothing is leveling out or off here ...

These hardware algorithms are not compressing the data, but just improving their use of an analogical transport medium.

It's just like claiming a text written with a smaller font is a compressed version of the same one with a larger font ...
Nothing is related with data compression here.