I try to deflate quite a number of tomographical image datasets. A dataset is a folder of 20-30 GB full of TIFF images like this:
Can you suggest the best compression algorithm from 7z cohort for such data please?
1) If all TIF files are 16-bit images, you can use delta:2 filter for them:
in command line:
If all BMP files are 8-bit images, you can use delta:1 filter for them:
2) PPMD method can be slightly better than lzma or lzma2, but decompression will be much slower.
3) If all files are different (there are no duplicates), you can switch off solid mode. So decompression of one file from set will be fast.
Note also that if you compress 16-bit data with lossless method (like 7-Zip), compression ratio is not impressive. 16-bit data contain too much random data that can not be compressed.
That is why compression ratio for 8-bit files is much better.
3) Unfortunately most of files in my datasets are 16-bit images.
2) PPMD increased compression ratio!
1) For some reason
D:\Temp\7z.exe a -mf=Delta:2 "D:\Temp\Raw.7z" "D:\Temp\Raw"
The parameter is incorrect.
What could be the reason for this?
it must work in 7-Zip 9.30 or 9.32.
-mf=delta:2 -m0=ppmd -mx
and for old version 9.20:
Is 9.30 or 9.32 stable?
I use 9.20.
I tested with version 7-zip 9.20 and WinRAR 4.11 and got the following rating (1 for the best):
1) WinRAR with Best option
2) 7-zip with -m0=delta2 -m1=lzma -mx=9 -ms=off -mtc=on
3) 7-zip with -m0=PPMd -mx=9 -ms=off -mtc=on
Have you something to add?
Can be slightly better for some files (not all).
WinRAR 4.* supports special multimedia algorithm. So it's good for such 16-bit data.
How can I use
with version 9.20?
Or these switches are for alpha-versions only?
works in 9.20 and later.
-mf works only after 9.21
But there is no difference. It's just simpler to write -mf in some cases.
-m0=delta:2 -m1=lzma for 9.20
results in the same compression ratio as
-m0=delta:2 -m1=ppmd or -mf=delta:2 (9.30 or 9.32)
Is this correct?
Default method is lzma2 in 9.30 and 9.32.
But lzma2 and lzma provide same compression ratio for your data. So it's better to use lzma in your case.
But you can use lzma2 with -mx3 or -mx1 switch, if you need
fast compression with multithreading.
and it will work in any version.
Log in to post a comment.
Sign up for the SourceForge newsletter:
You seem to have CSS turned off.
Please don't fill out this field.