Re: Computer 3-D

[P3D Larry Berlin <lberlin@xxxxxxxxx>]

>Lincoln Kamm  writes:
>
>The computer (assisted) image that I won a couple of contests with was a
>conversion from a 2-D painting by Hugo award winning painter "Kelly Freas".

*********************   A MOST EXCELLENT WORK!!!!!

>He gave me permission to do the conversion, and in that sense we
>collaberated on it.  I am more than willing to share the awards with Kelly.
>As far as there ever being a 3D button to make a 2D image 3D "with very
>little input/effort from the user" is just about impossible.  It took me
>over 30 hours to convert the Painting by Kelly, and I am sure that that is
>more time then was spent putting together most of the other slides in the
>competition even if you added up all of the other 400 slides to gether.  I
>am not saying that because I spent so much time on it that I should win, I
>just don't see how a computer could ever figure out depth information from a
>single 2-D picture.  The proccess I use is an artistic series of methods
>that I have spent almost 2 years developing using 4 different programs that
>are available to the public but were not designed for stereo, and is not a
>simle thing to do.  It definately needs LOTS of inormation from the user.
>
>With any computer program, the more you put into it, the more you get out.
>
>Lincoln Kamm
>Maker of 3-D STUFF
>http://www.concentric.net/~3dstereo
>

*******************  I completely agree with Lincoln. It's currently tedious
work with the biggest problem being that the software wasn't designed for
doing 3D conversions. However, by the same token, if someone were to design
some of this software with conversions in mind, it would make conversions
FAR easier to do than they are today. Other software tools PROVE that the
processes necessary to make conversions are NOT more difficult than other
computing problems, it just hasn't been done from the right perspective. The
little known secret here is that it is actually FAR easier than a whole lot
of other software accomplishments! This means that it could have been done a
long time ago with older software except no one (positioned to write tools)
thought it might be useful. Too many of today's tools lose their flexibility
with advancement. I consider most such advancements to be giant leaps backwards.

There is one other avenue of software that shows some promise. It allows you
to input two or more side by side shots of anything, then manually assign
common points in each of the images, then the software creates a 3D model of
the object based on the stereo (or multiple) photo input. This is most
useful for complex architectural renderings or other complex objects like
antique automobiles etc. It's also intended for creating 3D models rather
than stereo pair images for viewing. However, one could combine a number of
such models into a 3D rendering program and end up with stereo viewable
images that are quite different from the photography which originated the
objects.

I've read of other software that uses parallax in stereo pictures to
automatically assign depth levels to various parts of the original image.
Such routines apparently only exist in custom written software and I haven't
found it in any commercial stuff. However, if such routines could be placed
in a digital stereo camera by the use of microprocessors, the user could
adjust and tamper with the original images in a number of easy ways, all
while the images are still in the camera. It still requires at least two
images to determine the depth relationships. It seems we persist in our
*one-eyed* thought patterns by trying to make the computer see stereo with
only one eye. Give it two eyes and it too can see stereo for itself (once
given the programming for parallax).

Computers will never take the place of cameras, only make better or worse
use of what cameras can capture. It's the combination of technologies that
makes the advancement. It's the human mind that develops an idea of what we
want the combination of tools to be capable of. None of this techno stuff
takes us away from or substitutes for a human originated idea.

Larry Berlin

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


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