I put online another update in which results for latest Packet are included.
Regards,
Stephan
I put online another update in which results for latest Packet are included.
Regards,
Stephan
Thanks Stephan Hi!
Thanks!Still some important compressors are missing...
![]()
I run out of time these days and months so please tell me which compressors I should test. Also the new bitmap and wave corpora
keep me busy, but who knows if anybody reads it since no reaction came up.
Regards,
Stephan
My latest software:
PIM v2.50
BALZ v1.15
LZSS v0.01
![]()
Hi Ilia,
please consider them as tested. I'll test them this sunday.
All new programs are already tested - except LZSS which needs longer than expected. It can take another two hours until the update is online.
I also included QPRESS and Bit 0.6.
Finally, with one day delay, here's another update of the main corpora.
Besides the programs mentioned above some results for bitmap and waveform
columns have been added and some timings of recent compressors such as BIT 0.2 and Flashzip 0.9.
Sorry for the delay, Ilia
Thanks a lot!![]()
Thanks Stephan! Did I want to ask you which format I/you/they are the waves sounds because I don't succeed in understanding because my filter Delta for the waves doesn't work with your files?
@Stephan Busch:
Thanks a lot for the updating! Seems I must work on a seperate analog model (bitmap, audio etc) for better compression![]()
Thanks Stephan!
http://www.squeezechart.com/
Hi Nania. The new waveform set consists of 16 stereo wav with 44.1 kHz at 16 bit. On compressors that do not support wildcards I use the tar variant
flashzip -m2 -s7 -b5 wav2.tar wav2.flashzip
This may prevent audio detection.
Since the interest in this testset is high, I have made it available to the public for one week from now. The archive contains OFR created with -maximumcompression switch and when unpacked may help you doing some tests.
www.squeezechart.com/TEST_Waveforms2.rar
For those interested: MMA 0.0a3 compresses them to 399.833.504 bytes
which is very close to NanoZip and only 10 MB are left for new highscore
which is currently held by WinRK (389 MB). And WinRK comes close to APE insane and TAK -p5m.
Regards,
Stephan
the wave files are united in an only file TAR or not ?
flashzip.exe -m2 -s1 -b5 C:\ofr\ABBA.wav t.zip 51614684->34095767
flashzip.exe -m2 -s1 -b5 C:\ofr\BenKing.wav t.zip 31620332->21828810
flashzip.exe -m2 -s1 -b5 C:\ofr\Desmon.wav t.zip 28306988->19384443
flashzip.exe -m2 -s1 -b5 C:\ofr\Enyas.wav t.zip 41477564->22626027
flashzip.exe -m2 -s1 -b5 C:\ofr\Hans.wav t.zip 39034412->30338441
flashzip.exe -m2 -s1 -b5 C:\ofr\Harajuku.wav t.zip 44165900->37766055
flashzip.exe -m2 -s1 -b5 C:\ofr\Jan.wav t.zip 37486124->23358070
flashzip.exe -m2 -s1 -b5 C:\ofr\JSD.wav t.zip 37686140->18383227
flashzip.exe -m2 -s1 -b5 C:\ofr\KennyG.wav t.zip 83303468->52733920
flashzip.exe -m2 -s1 -b5 C:\ofr\Laut.wav t.zip 45899324->27890167
flashzip.exe -m2 -s1 -b5 C:\ofr\Leonard.wav t.zip 62866988->49644821
flashzip.exe -m2 -s1 -b5 C:\ofr\Queen.wav t.zip 45640604->31596462
flashzip.exe -m2 -s1 -b5 C:\ofr\Royt.wav t.zip 40329260->31278621
flashzip.exe -m2 -s1 -b5 C:\ofr\Sash.wav t.zip 39819404->32092201
flashzip.exe -m2 -s1 -b5 C:\ofr\Shakir.wav t.zip 38246444->25510968
flashzip.exe -m2 -s1 -b5 C:\ofr\TravelinCare.wav t.zip 35154476->24223861
number of files=16 original=702652112 compressed=482751861 ENC=56.965000 sec. DEC=85.310997 sec.
THANKS STEPHAN ! FLASHZIP IS NOT GOOD IN TAR WAVE COMPRESSION!
Last edited by Nania Francesco; 19th August 2008 at 19:48.
Excuse Stephan if I am repetitive but the various groups of file, for instance BMP, with the single file compressor you always use a file TAR or no!
Thanks for all your effort, Stephan! And thanks for the test-set.
AFAIK, yes.Originally Posted by Nania Francesco
I use the TAR'ed sets when a compressor do not support wildcards
(p.ex. flashzip.exe -m2 -s1 -b5 C:\ofr\*.wav waveforms.zip)
flashzip -m2 -s7 -b5 compresses all .wav to 473.050.044 bytes
(each wav in a separate archive). This is a very good result (1 MB better
than current WinRar).
The TAR is used not to make the programmers angry but to save time.
Compressors without wildcard support would require to create each file in separate archive and adding timings.
if you want I make me to furnish you of the programs written in purebasic for every type of archiver with cotrol of the equality of the files and the time with choice of the folder to compress available [GUI] with creation of LOG file !
That sounds good, but wouldn't it be better to add support for wildcards?
When it comes to programming, I can declare myself a greenhorn so I don't know how much work is behind that wildcard thing.
Stephan Il problem I think, this is naturally one thought of mine and not a criticism to a great character of the compression that you are (Squeeze Chart forever!) , is that the programs in Console Version are too disadvantaged in comparison to the ARCHIVERs with the files TAR! l have been able to also see you FLASHZIP with the active filters improves of over 120 MB and they are not little!
a solution could be this:
you write a file .bat with the NOTEPAD or
1) phase
open Prompt
write c:\edit
write the commands [example]
flashzip - m2 - s7 - b5 test1.wav t1.fz
lashzip - m2 - s7 - b5 test2.wav t2.fz
..
..
flashzip - m2 - s7 - b5 testx.wav tx.fz
save you test encode.bat
2) phase
open Prompt
write c:\edit
write the commands [example]
flashzip - x test1.wav t1.fz
flashzip - x test2.wav t2.fz
..
..
flashzip - x testx.wav tx.fz
save you test decode.bat
3) phase
open Prompt
write c:\timer encode.bat
...
...
time xxx
4) phase
write c:\timer decode.bat
...
...
time xxx
5) for size compression result
c:\dir *. fz
Dear Nania,
where would we stay today without critics? And how would the Squeeze
Chart look today if no suggestions and critics would be made? I'm not a
great character but a man of visions. One vision is to make Squeeze Chart
better, and therefore I need feedback.
Of course, the TAR hurts compression unless predictors are integrated into
base algorithm, but also TAR simulates solid archiving which also manipulates
results. But I think I will not use TAR anymore on bitmap and waveform sets
at least on those compressors that do gain from getting pure (unTAR'ed)
data.
However, I do like the idea of FlashZip Archiver with GUI
FLASHZIP in the next release is new FLASH Archiver 1.0!
---
flashzip - m2 - s7 - b5 test1.wav t1.fz
lashzip - m2 - s7 - b5 test2.wav t2.fz
..
..
flashzip - m2 - s7 - b5 testx.wav tx.fz
---
not support for wildcards ...
---
i think it would be wonderful to have a implementation of
"compress all files within a directory inclusive all subdirectories"
---
but may be a compromise is possible?
what about to implement a standard for commandline like
compressor -specialswitch1 -specialswitch2 a archivefilename @filelist.txt
where the file "filelist.txt" contains a list of the full_filenames_inclusive_path
for example the file "filelist.txt" contains:
---
"c:\jpegger\AUTORUN.INF"
"c:\jpegger\JPegger.chm"
"c:\jpegger\jpegger.exe"
"c:\jpegger\readme.txt"
"c:\jpegger\cd_start\vstarter.exe"
"c:\jpegger\cd_start\VSTARTER.INI"
"c:\jpegger\cd_start\vstarter.jpg"
"c:\jpegger\images\sample1.jpg"
"c:\jpegger\images\sample2.jpg"
"c:\jpegger\images\first day on the beach.jpg"
"c:\jpegger\the beatles\the white album\title 1.jpg"
---
one line for one filename
the charset within the listfile can be 7bit-ASCII
or the default system charset within windows
(Windows-1252 / ISO-8859-15 / Latin9 in many cases)
or ...
may be the best would be UTF-8 ...
---
***
FOR EXAMPLE THE 7ZA.EXE SUPPORTS SUCH THINGS (and more):
***
listfile - a file containing a list of files
The filenames in such a listfile must be separated by new line symbol(s).
For list files, 7-Zip uses UTF-8 encoding by default.
commandline example:
7z a -tzip archive.zip @listfile.txt
It's strongly recommended to use only the real (long) file names.
***
i think
such a compromise will be easy to use for the benchmarkers of us
- only one time create such a list of filenames for each special testset
and i think
it would be easy to implement such a filelist-feature for the programmers
- more easier to implement then a full archive-solution
isnt it?
so it would be easy to do the benchmarks
and
the compressor-program can each file individually analyze and compress
@Nania and @Stephan:
could such a compromise be possible useful?
what do you think about such a compromise ?
@toffer
what do you think about implement such a listfile-feature?
it will be very complicate?
Last edited by joerg; 20th August 2008 at 04:03.
@ joerg
I sincerely hope not to meet too many problems with the new FLASH console archiver on which the version GUI will found him!
Hey! Such a file list support should be a piece of cake - but i'm usually too lazy, especially since i'm experimenting with other things atm. Maybe someone would be interested to write a standard archive container format with api? Ideally one should only supply (de)code routines for streams.
@toffer
thank you for your replay
*** such a @listfile support should be a piece of cake ***
i am glad to read you are lazy = lustig
and i am think you are not lazy = langsam
because your compressor is fast enough too
---
Maybe someone would be interested to write
a standard archive container format with api?
Ideally one should only supply (de)code routines for streams.
---
a wonderful idea
---
but i think such a simple listfile feature could be easy implemented
and would a good compromise to make benchmarking easier
no more no less
best regards