I work for a Large Photo Lab and we are facing with the following
problem.

My Boss Tells Me That After We Compress/UnCompress (We Use WinZip)
Large (30MB+) Digital Pictures We See "ARTIFACTS"!!!
Basically, He Means That After We Do Some Work (We Use PhotoShop) On A
30MB Digital Picture and Compress It + UnCompress It The Picture slightly
introduces ARTIFACTS!

AnyOne Has Seen This Problem?

Thanks For All Your Help,

WinZIP compress/decompress should be perfectly lossless under all
circumstances.
And it should detect any archive media errors and grumble about them too.

The fastest test would be to Create a WinZIP archive of a known good file.
Save the original somewhere safe.
Restore the image from the ZIP archive file and then binary compare it
against the perfect original. I would be *very* surprised if they were
different. But hardware faults or malicious virus action cannot be ruled out.

If he still needs convincing subtract one image from the other and count the
number of colours in the difference.

If you are resaving the digital image in a lossy format like JPEG first then
all bets are off.

The compression algorithm used by WinZip is lossless, essentially the
same as native LZW compression used for TIFF.

A compressed file, when extracted (uncompressed) is bitwise identical to
the original (the same file that has never been compressed by WinZip).

Maybe your boss is compressing with JPEG?

--
Lucas Tam (REMOVEn...@rogers.com)
Please delete "REMOVE" from the e-mail address when replying.

If so, the use of photoshop to load, work on and then re-save could
introduce some quality loss but that's nothing to do with Winzip.  

Someone's got a wire crossed somehwere.

In article <6W2dnRTpyYHeU8GjXTW...@comcast.com>, ade...@yahoo.com
says...

1) Take an image. Make a copy of that file somewhere else.
2) ZIP original image, remove the source.
3) UnZIP the image back.
4) In Photoshop, open both unzipped image and the copy you created at the
step 1.
5) Image -> Calculations, set Source 1 and Source 2 to unzipped image and
copy of original, respectively; set Blending to Difference. Hit "Ok".
6) As a result, you'll get the image showing you the difference between the
two. Eventually, it will be completely black, meaning that the difference is
zero. Just to be sure, use Image -> Histogram and it will show you that the
only value across the image (and so the average and median values) is 0.

BTW a couple of good advices:

1) Use the comparison routine described above (calculating Difference) any
time you are unsure whether some routine\compression\something is lossless
or no. Simply checking difference image histrogram worth hundreds of
"opinions" you may hear, my own opinion included :-) If you get zero
difference, no loss occured. If you see something above zero, there is a
loss. No need to ask people for opinions - just check out yourself.
Experiment will give you the best proof.

2) Stop Using Leading Caps When They Aren't Really Necessary :-)))