Re: WHOA!! Backup the JPG truck...

[P3D Larry Berlin <lberlin@xxxxxxxxx>]

>Date: Sat, 06 Sep 1997
>From: P3D Jonathan Gross cmments:
>
>Dan Shelley wrote:
>> It's an anaglyph saved at 100% quality (0% compression). THe images were
>> both originally from a Photo-CD, used at thier highest resolutions. I
>> think it's pretty clear myself...
>> 
>> JPG's can be used quite effectively for anaglpyhs IMHO.
>
>I agree with you, but what are the advantages left from JPEG for image
>compression?
>
>For loss-less image compression, seperate out the RGB into three 8-bit
>image channels, and compress each channel using a loss-less LZW (PKZIP)
>method. You'll get between 3:1 and 5:1 compression with no loss of
>important boundary data.

******  I'm not at all sure why anyone would want to store separate 8 bit
channels other than scientific curiosity. What's the point?

>
>And, for high quality stereo pairs you might want to use 1024X768 field
>pairs.  If people published their JPEG and JPS images with 0%
>compression and 2X1024X768 resolution, then it might not be so bad.  But
>is that what people are doing?

*****  Why field pairs? Again, what is the point here? Using the JPEG scale
set at zero or rather an maximum quality results in a file that isn't as
small as it could be with more compression. It's seldom as small as is
optimum for web page use. The point isn't to use maximum quality all the
time, it's to use it when there is justification for a larger file option.
Otherwise, moderate amounts of compression are the intent and the common
practice. Why in the world should stereo images be excluded from this common
usage? No one has said that high quality no-loss images should NOT be used.
There is plenty of room for both ranges of quality. That's the whole point.

>
>How about alignment information?  Without it, it's like handing someone
>two loose film chips and asking them to look at the stereograph.  In
>that respect, two JPEG images are better than one JPS pair.  At least
>you can input them into an alignment utility without having to seperate
>them first.

*****  The intent of a JPS file is to align the images BEFORE saving in that
type of format. What would be the point in creating a JPS stereo pair that
isn't aligned? The very idea is absurd. No one has to separate a JPS image
in order to use it, at least not manually. The software splits it right down
the middle. When the file is created, the alignment is done BEFORE putting
them together as a pair to save. JPS is nothing at all like handing someone
loose film chips. That's ridiculous. We aren't talking raw images here, but
that seems to be your reference. Two separate JPG files can be handled
separately, everyone know that. It's useful at times even, but so what. The
JPS file type eliminates the alignment problem you are suggesting by doing
the alignment before saving. A JPS file is like a properly mounted and
aligned stereo slide, not even close to *loose chips* in the wildest stretch
of imagination.

>
>All of a sudden, a special format for stereo pairs doesn't look so bad; 
>High quality loss-less compression, non-destructive cropping and
>alignment information, and indelable copyright notification are
>important features for commercial-quality stereo images IMHO.

*****  why must you scrabble in the dirt trying to find ways, any weak way
at all, to disparage something as simple as a JPS stereo image? The purpose
of a JPS includes use of JPG compression, not because anyone likes
compression for it's loss factors, but because most everyone using the
internet for images likes to download them in a reasonable time. Certainly
tradeoffs exist. What's your real point?  Why argue about the color of the
sky? There are formats available which provide NO-LOSS compression if and
when it is needed. If you DO happen to want to use compression, then a JPS
answers that very prevalent need and does so for anyone wanting to use it,
even you. Pretty convenient. No one has said that JPS is the only stereo
format in existence, nor is it necessarily the best, depending on any
particular set of definitions for *best*. We all know that each person might
define that in a different way. That's the point. Let each user use what
they see fit to use for their own purposes. If you want NO LOSS, then don't
use JPS. Pretty simple.

To make sure I'm clear here, let's take your list one item at a time

1. *a special format ... doesn't look so bad*  - This statement seems to
contradict your earlier observations. But yes there is a need for good
stereo formats.

2. *High quality loss-less compression* - Stereo files can be saved in
several No-Loss formats (TIF, BMP, TGA, to name a few) and compressed with
PKZIP for a lossless compression. Or PNG provides a No-Loss compression
method and can be made into the stereo file type of PNS. It would be nice if
other compression schemes come along that address special needs in stereo
images. Are you suggesting one?

3.  *non-destructive cropping* - This sounds like a contradiction in terms.
What exactly do you mean here? When I align and crop a stereo pair of
images, I intend that the cropped areas be discarded and the resulting
images be fully matched to the stereo information that is intended in that
particular image. I wouldn't want someone else to undo that basic primary
work and end up with image areas that aren't properly a part of the stereo
pair that I intended. If someone wants to align my image pair differently
than I have established, fine, use the information I've provided and further
crop if that's your wish. How does that affect any kind of file type? All
anyone has to do is open the images in some editing progam and re-edit. No
problem, but not what the originator of an image intended.

4.  *alignment information* - This is inherent in setting the vertical
alignment and setting the stereo window. Once those factors are set to
encompass the optimum area of stereo image, they don't need further setting
unless someone wants to play around. The use of any file type from TIF, BMP,
TGA, JPG, PCX, or stereo files of BMS, PNS, JPS, will still allow a viewer
to load the image into an editing program and set the window or alignment
wherever they want. Again, that is extraneous to what the originator of the
image intended. If that originator does a good job of optimizing these
factors, part of any basic stereo image process, any further adjustments are
likely to be less ideal, but not impossible to do.

5.  *indelable copyright notification* - This is something that is an issue
for anyone using any type of image file. Why should it be considered any
different for stereo images than for 2D images? Are artists working in 2D
not interested in this issue? What's wrong with various means used for
copyright information which ranges from adding a signature or mark of some
kind to the image itself, or providing copyright information on adjacent web
pages and other materials. There are even text programs that add such
information to various image file types in text blocks. Maybe the issue
hasn't been fully resolved, but that's true for any image type. This
deserves to get attention and be subjected to discussion totally aside from
any stereo image file type. If and when some image type comes along that
allows such information to be recoreded, fine. Until it is available, it's
an open issue not directly relating to stereoscopy. The presence of such an
issue does not negate any immediate need for formats to share stereo images
with. We have such file formats available. Copyright issues still exist. The
sun still shines and it rains occasionally too. So what?

6.  *commercial-quality stereo images* - What is meant by this? Who defines
what is a commercial quality image? Should everyone wishing to share images
have to subject their image to an evaluation board to determine first
whether or not they are of commercial quality? I've seen lots of commercial
images, not 3D images, that are far lower than any image I would ever hope
to use or show, yet they are part of the *commercial quality environment.* I
can go out today and buy a set of these poor images (one brand that I know
of that fits this category is *Expert Software*) off the same shelf that
contains high quality images. Obviously different standards exist even
within that category of *commercial*. Typically commercialism dictates that
what the market will bear is commercial. As long as people pay real money
for garbage images, then those images must also be a part of commercial
quality. I often wonder how a company using poor quality images can stay in
business, yet some of them do. What does this have to do with stereo images
or formats for the conveying of stereo images?

>
>JPEG can be used effectively for anaglyphs, but be careful about
>extending that to stereo pairs.

*****  What a twisted conclusion that doesn't have any logic to back it up,
just to attack something that's as workable and useful for you as for anyone
else.

Larry Berlin

Email: lberlin@xxxxxxxxx
http://www.sonic.net/~lberlin/
http://3dzine.simplenet.com/


------------------------------

End of PHOTO-3D Digest 2271
***************************
***************************
 Trouble? Send e-mail to 
 wier@xxxxxxxxxxxxxxxxxxxxxxxx 
 To unsubscribe select one of the following,
 place it in the BODY of a message and send it to:
 listserv@xxxxxxxxxxxxxxxxxxxxxxxx 
   unsubscribe photo-3d
   unsubscribe sell-3d
   unsubscribe overland-trails
   unsubscribe icom
 ***************************