Masked Compression (Ultimate compression in one go)

[Razor12911]

The only thing you must figure out is how the header is structured.

Normally the the precomp header gives the version. It must tell you where the processed streams are located then after finding then, it makes it very easy to make the unpacker.
Another tip, if using handles, never write anything on cmd window that isn't part of the data. Meaning all that Precomp v0.4.4 - UNSTABLE bla bla has to go else you'll end up with errors unnecessarily.

[panker1992]

thanks, dont worry i know where to find you in case i find trouble :P

i hope you will be determined to help then as well

[Razor12911]

Anytime mate

[Amsal]

Panker, one question, what is the required ram for decompressing this data which is compressed using Masked Compression?

[panker1992]

a lot of i expect you know that you can pack with x64 program and unpack with 32 bit program

if you use srep64 for decomp you can see 4 gigs of ram easily

and if you use nanozip it uses the same amount of ram as for compression

if you use my xz method which is ultra lz77 it uses total 500~800mb ram :P

[felice2011]

Hi Panker, using the methods with the final mask configured in this way

"msc+srep64+delta+exe2+xz/$jpgsolid=jpg+zstd/$void=srep64+zstd/$text=nz/$precomp=precomp+msc+srep64+delta+lz77"

the compression is successful only with the "Precomp 0.38" with all other versions I get this error:



Is there any solution, or for the time being, the insertion of the algorithm Precomp in final mask is recommended only with the use of precom0.38.

[panker1992]

well i have studied a lot of precomp source code lately,

i can tell some facts that has come to my attention

precomp 0.38 was a masterpiece of software, all other precomps are pretty unstable the way i see it, precomp 0.40 and above when fail to see zlib header come with this error in 90% of the cases

second fact , i havent uploaded nor updated my Masked Compression to use precomp 0.44 or even precomp 0.45 which is unstable :P not even previous precomps

my private messages are flooded by people begging to implement precomp 0.44 or even other precomps

i will not yet update the masked compression to support, because it spawn errors like these!!!

today i emailed Christian Shneider, critical error was fixed with memory leak, and i sent him library with tak, next precomp will most likely support wav precompression
and it already takes support for mp3 precompression

in the meantime precomp 0.44 offers nothing, nothing new is inside precomp 0.44 that will make precomp 0.38 obsolete

i literally beg people to not put precomp 0.44 to official public type tools for massive usage, and i havent released precomp 0.45 for this reason as well because i have the exe file but i keep it under heavy testing

therefore Felice i conclude that precomp 0.44 MUST be withdrawn anbd not yet be massively used under any circumstance especially on critical Data

Please Bare with me, for any changes on precomp i will post myself
it's not like i keep it for myslelf or something

sorry for this long post

[felice2011]

OK masks works very well, with the configuration in the last mask "/$precomp=" with Precomp added, should only be used with precomp 0.38, other versions Precomp can be used safely without entering Precomp in the last mask.

Last Mask : "/$precomp=precomp+msc+srep64+...etc" (Only precomp0.38)

Last Mask : "/$precomp=msc+srep64+delta+exe2...etc" (All Precomp Version)

Panker you're doing a great job.

[panker1992]

i will tell you a secret,

i have implemented it in my last v2.1 update

$precomp= ...etc etc is a mask to filter data that lz77, xz, fazip and generally any lz77 variant can compress only 5~8%, this is why you put precomp here because using precomp before lz77 can increase ratio by about 45~60% max

you can use precomp filter mask to select out uncompresable data and store using

$precomp=msc+srep+zstd this mask will allow you to select uncompressable data

for store purposes .... i have used it in my sonido mask (sonido = speed of sound )
my ultra fast mode in my tools

and yes

use can use any precomp if you put it in the beginning
for example
ultra = precomp+msc+srep64+delta+exe2+lz77/$jpgsolid=jpg+zstd/$void=srep64+zstd/$text=nz

but that will also increase time so a dual shaped sword is more like it

[Ichiraku2001]

Hey @panker1992! First of all you're doing a great work!

I'm getting this error while decompressing!

Game's MAD MAX
used -mrefl8 to compress!
Input Size = 1.29 GB (one of the splitted folder)
Output Size = 355 MB

But when trying to decompress with WPI for masked compression! it shows the above error! I've checked the setup file, its not corrupted!
Please Help

[panker1992]

The reflate razor made cannot be used in inno
It errors every time you need to have manual reflate for mad max
This is known issue I am sorry I cannot help

Reflate is beta tool for now

[panker1992]

To show you my Gratitude for your continued support, i decided to roll Masked Compression v2.2

i already posted everything in the main post
so to change climate i will post some photos of my tools under usage

if you happen to test them give me your honest opinion

A) Would you like for me to add a new feature ??

B) Do you think i missed something important ??

anything you want i will test and add if i see potential

3 pics below: bik encoding using my script, success on decompressing steam folder after using on it all my tools simultaneously , and my new mask on usage offering 18% ratio up from 11%


Enjoy

[felice2011]

Do not worry I do not update beta versions tester on my utility..
You run too much with your masks..., you have added the mask for the recompression data in MP3, not hear any difference in the optimization of compression threshold.

Tests executed on a data folder mp3 approximately 234MB

Masked method used for Test: "precomp+msc+srep64+delta+exe2+lz77/$jpgsolid=jpg+zstd/$mp3solid=mp3+zstd/$void=srep64+zstd/$text=nz"

Your update: v 2.2..



M_S_Kompressor with optimized profile based multi mask By Panker..





Personally, I may be wrong, I not have so much experience in data compression, but exceed on a number of algorithms containing in the method multi-maschera, involves a loss of a compression stream, increasing the compression time, and decreasing the compression threshold.

[panker1992]

Hmmm use the mask directly .... test this maskless

and remove msc :P

both msr and msc can compress mp3 files ...either -mp3=0 or remove msc entirely from the test

to see if my mask really works :P else i need to find what it goes wrong :/

[felice2011]

I had already tried...
The compression level of an audio file (*. mp3) and higher using a profile with MSC.
The same level of audio compression the lowers considerably by using MSR.
I'm sorry but I prefer to MSC