Hi folks,
another update:
bitmap:
+ Libra8 1.0.0.1
+ PWC
main:
+ LPAQ1
+ WinTurtle 1.0
+ TarsaLZP (4.8.07)
+ LZPM 0.08
special treatment (JPEG):
+ Paq8fThis
+ PackJPG 2.2
Cheers,
Stephan
Hi folks,
another update:
bitmap:
+ Libra8 1.0.0.1
+ PWC
main:
+ LPAQ1
+ WinTurtle 1.0
+ TarsaLZP (4.8.07)
+ LZPM 0.08
special treatment (JPEG):
+ Paq8fThis
+ PackJPG 2.2
Cheers,
Stephan
you need to update upx v3.01
and try this command
upx --ultra-brute *.exe
it will try 72 trial..
for images add the program deflopt for png as last process
http://www.walbeehm.com/download/
for mp3 add .. mp3packer that convert the cbr files vice versa to vbr
first you need to use it as
mp3packer -z input.mp3 output.mp3
add it as a compressor then as a preprocessor to soundslimmer
http://omion.dyndns.org/mp3packer/mp3packer.html
great site..you brought me memories
TrueColor [Artificial, Scanned] (ABC Newspaper 5100x6980x24 Bit)
.PLR Libra8 1.0.0.1 (Andrew Polar) (21.07.07) -3
It takes another's place![]()
Thanks Stephan!![]()
A new version of ANDpakk2 has been released - 0.18 (download link)![]()
not bad .. but works on sinlge EXE file with link to DLL..
Well, although it has been improved to work on more executables, I still haven't found any common exe, that can be successfully ran after compression.![]()
You can check for valid file with great FS utility:
ftp://ftp.sac.sk/pub/sac/pack/fs.zip
Use this command: fs -s %file_name%. "RVA" of the first section must be "00001000". For example WinTURTLE is compressible. But proper RVA doesn't garantee correct compressionBut i hardly find "bad" executable.
P.S. ANDpakk2 itself shows RVA before compression.
Thansk nimdamsk! WinTURTLE belongs into compressible group, but upon successful decompression instead of running the executable itself it crashes, that's what I meant...
//edit: Yes, you both are right, on second PC with XP SP2 it works fine.![]()
May be there is something with yours pc as i have no problems with decompression on both WinXP Pro SP2 i tested.
Also works OK on my Sempron 2400+, WinXP SP2 machine.
Compressed size = 53.3 KB (54,656 bytes)
The new Update of Squeeze Chart 2007 ? Hi Stephan!
Squeeze Chart updated!
new:
LPAQ1
LPrePAQ 1.1
LZPM 0.11pre
PAQ 8o SSE2 v2
Precomp 0.3.6
WinTurtle 1.3.0
enhanced 'Special Treatment' page
Tried to contact SoundSlimmer Team several times in order to ask
them for support of more audio formats (.AAC, AC3, .MP4,
.OGG, .WMA) but they do not answer at all (dead? or allied to Stuffit?)
some new records on 'Records by filetype' page
To Ilia: The most of the increased time LZPM 0.11pre was spent on
highly redundant data such as CD Image subchannel data, airy encyclopedia file and on Nokia Testset.. On those data type LZPM
seems to be very very slow.
To Nania: WinTurtle 1.3.0 has lost some of its compression power
compared to WinTurtle 1.0 - Is that intentional?
Greetings,
Stephan
Thanks Stephan!![]()
Its due to an optimized parsing with no search limits. The "9" is about the max compression possible. As you know the decoder is the same for all compression levels and its very fast.Originally Posted by Stephan Busch
The official LZPM 0.11 might be even slower, but with higher compression, especially on binary data. The original v0.11 is still under development anyway.
![]()
Cool! Waiting for release!Originally Posted by encode
![]()
afaik, mp3 compression algorithm was developed here, in Russia. its quite substantial work and taking into account that SS was probably not commercially successful, i dont think that there are any chances that the same work will be repeated for aac or any other format. moreover, i bet that these formats has less redundancy (not saying that they are still less popular)Originally Posted by Stephan Busch
you should rather write recompression algorithm as mp3 was developed in germanyOriginally Posted by Bulat Ziganshin
an compressor for divx (ie. mpeg4 asp + mp3) would be good![]()
Thanks Stephan!![]()
Some new results are ready..
TarsaLZP 04.08. 138th place (2.276.263.447) in 2036 sec
TarsaLZP Interim 109th place (2.178.453.621) in 1836 sec
WinTurtle 1.3.0 126th place (2.224.841.376) in 1812 sec
WinTurtle 1.0 102th place (2.146.564.413) in 2207 sec
HOOK 1.0 136th place (2.273.818.524) in 5639 sec
CMM 17.09. 150th place (2.302.658.240) in 7333 sec
Concluding that TarsaLZP Interim is the best TarsaLZP ever built;
WinTurtle 1.3.0 is faster but compresses worse compared to 1.0,
Hook 1.0 comes close to best Hook 0.8c and CMM in his first run
reaches WinACE class (which is remarkable).
To Bulat:
I think that compressing already compressed data such as PDF, SWF, MP3, GIF, PNG, MP4, MPG, OGG and AC3 has future since
those formats are used very often and do exist on all computers.
Maybe WMA and AAC are seldom, but lossless recompression is
both wish of millions of users and a nice selling strategy.
To Donkey7:
DIVX is a junction of MPEG4 and MP3 and it is very likely that one
day there'll be a lossless recompressor for it.
To Ilia:
The info about LZPM timings was just statistical, not meant to be
critic. I have listened to your song over at myspace.com and it reminds me of my past time when I made songs..![]()
the PDF,SWF,GIF,PNG, is already preprocessed with Precomp
JPG you packjpg and precomp has it to support JPG
MP3 has lossless reduce (mp3packer http://omion.dyndns.org/mp3packer/) and there is a packer to reduce it about 20% its called soundslimmer www.soundslimmer.com currently it has MP3 plugin and it will add other formats soon (AAC)..nothing new for over two years..
MP4 is a container not a type format like MP3 it mainly uses AAC
MPG has similar structure as JPG so it would be easily linked (i hope)
OGG is similar to MP3 but more advanced it has lossless phases that can be used to repackit in smaller form...
the main problem is that all these are time consuming for developer and for end user .. the advance in computing speed would remove this problem but not now maybe within five years when we have a processor with 10-cores with the cost of single core-processor..
Hey maadjordan,
I know all this - and a little more - you know, as benchmarker you hav access to non-public pre-releases.
And I can tell you that it won't take 5 years until we have preprocessors for all that filetypes.
Of course, development can take years. But all those filetypes
can be recompressed losslessly with the nearly same technics as
they are used in Precomp and PackJPG, so both authors could
modify their programs so that those can process the files mentioned.
Regards,
Stephan
BTW: SoundSlimmwe team will not release anything new.
They gave up development. So its up to other programmers.
strange.. the sound slimmer has been updated on 2007-augest.. but with minor changes.. and if they give up with it.. why don't they open-source the engine.. so every body take benefit of it..
BTW it would become popular if the provided plugin to play the resulted file (mpz) in winamp or foobar or windows media player..
i remembered one thing regarding OGG there was a tool called "rehuff" it would repack the huffman coded bits into smaller size and it would reduce 1-2% of file size..
Can you give me a link to that OGG preprocessor, please?
I am dying to know how it performs on my test file @ special treatment page![]()
i can't find it any more .. it was issued for testing on 2002 and never get released.. i can provide some links on the matter but none can provide direct optimizing tool.. but here is some topics around it and there is some experiments done to use arithmetic entropy..
http://osdir.com/ml/multimedia.ogg.vorbis.devel/20 06-08/msg00012.html
http://osdir.com/ml/multimedia.ogg.vorbis.general/ 2003-01/msg00082.html
http://web.interware.hu/rudas/