On Fri, 26 Apr 2002 05:02:25 -0500, Samuel W. Heywood wrote:

> On Fri, 26 Apr 2002 03:29:35 -0400, Clarence Verge wrote:

> <snip>

>> Tell him to zip the .BMPs and rename them to .ZBM.

> I have heard some stories about a phenomena called
> "lossy compression"  According to these stories it is OK to zip
> images in some formats, but not others.  Among those formats that
> are supposedly OK to zip are the BMPs.  If you zip a JPG and then
> unzip it, the resolution will not be as keen as seen in the original
> before it was zipped.  I don't know if these stories are true and
> I have never taken the time to conduct any experiments to see for
> myself.  If it is true that the phenomena called "lossy compression"
> does in fact exist, is the loss of resolution noticeable to the
> naked eye?  Also, if such a phenomena does indeed exist, how can it
> be explained?  Two other questions I have are these:

Not true.
An unzipped file is exactly the same as it was before being Zipped.
Every single byte of it.
(pun intended)<g>

> 1.  If there is such a thing as "lossy compression", could it be
> mimimized by using ARJ, ARC, DWC, LHA, TAR, or some other alternative
> to PKZIP?

None of those formats is "lossy".

However... the Jpeg image format itself IS "lossy".

BTW,
It took Clarence quite some time to convince me of that fact.
(but I finally DID "see the light")<VBG>

> 2.  Of the many image formats that we know of, and if it is true that
> so called "lossy compression" is a problem in the case of some image
> formats, which types should not be zipped?

All of them CAN be zipped and then unzipped with no problem
what-so-ever.
However.... the only format which will end-up-with a significant
reduction in size is .BMP

All of the rest of them are already compressed just about as small in
size as they can get.

See what I mean?<g>

Searching ZIP: C:/1SEND/A-SCREEN.ZIP

 Length Method  Size  Ratio   Date  Time   CRC-32  Attr Name
 ------ ------  ----- ----   ----   ----  -------- ---- ----
1737426 DeflatX 20085 99%  04-27-02 16:56 81656dcf --w- A-SCREEN.BMP
  19178 DeflatX 18513  4%  04-27-02 16:58 2c45fbd3 --w- A-SCREEN.GIF
 140734 DeflatX 84235 41%  04-27-02 16:57 6c01d9d8 --w- A-SCREEN.JPG
 ------         ----- ---                                 -------
1897338        122833 94%                                    3

(made the screen-shot in 800x600 Hicolor BMP
 and converted it to GIF and JPG)

As you can see. The GIF is smaller than even the zipped BMP
But.... it is only 256 colors instead of 16million colors.
And.... it really looks crappy.

That 41% compression ratio on the JPG is very uncommon.
Most times, you're lucky to get a 10% reduction in size.

Grab the zip file from:
http://www.angelfire.com/id/glenndoom/images.htm

> It could be that all this talk I have heard about "lossy compression"
> is just a bunch of bull, but I tell you that a lot of people believe
> it.  I won't believe it unless I can find an authoritative source on
> the subject which confirms the existence of the reported phenomena.
> The truth is out there.  Does anyone know the URL?

http://www.jpeg.org/ <VBG>

-- 
 Glenn
 http://arachne.cz/
 http://www.delorie.com/listserv/mime/
 http://www.angelfire.com/id/glenndoom/download.htm
 http://www.thispagecannotbedisplayed.com/

Reply via email to