Re: questions, questions...

[P3D <PK6811S@xxxxxxxxxxxxxx>]

Peter Homer asks:

>... if it was
>3D what was the seperation would it be limited by the diameter of the
>single lens used to take the original or would it depend on the degree to
>which it was out of focuss as with some other SL3D systems such as the one
>using red green filters like  the Q-dos . This would seem to be a way to

SL3D lenses take two separate pictures and overlay them on the same
film, they are reconstructable because they do not share wavelengths.
Depth of field is controlled by the size of the aperture and is not
inherently less in SL3D than in other photography.  

Peter Davis writes:

>Well, I'm beyond the limits of my understanding here, but seems to be
>that by putting the clear and black stripes in the center of their
>filter, Vivitar has both reduced the area of the filters (increasing
>depth of field?) and increased the separation of the two apertures
>(increasing the 3D-depth?).  I'm guessing you're using "aperture" here
>to refer to the area of each filter, as opposed to the opening of the
>diaphragm in the lens.  So, what's wrong with the Q-dos
>implementation?  Or have a totally misunderstood your explanation?

I can't comment on the quality of the anaglyphs produced by Q-dos, 'cause
I haven't seen any.  Each filter has its own aperture which is the
area not covered by the diaphragm.  In another implementation they
might have separate fixed or independent variable apertures.

The 'clear stripe' bothers me a bit.  Is it really transparent, so
that there are actually three images captured?  One red from the left,
one blue from the right, and one of all colors from the middle?  Seems
like that would cause considerable ghosting, but the mind might
be able to overlook it.  Or maybe that is how they overcome the
red/blue shading at the sides of the picture.  I'll have to make up
some new masks and try it!  Since it is the separation of the weighted
centers of the two filters that matters, maybe it is ok if they
overlap.

>I haven't noticed any red/blue shading at the edges.  I do see
>red/blue "halos" around objects in the picture, which is what I
>thought was providing the depth information.

Good and yes, respectively.

>By the way, another artifact I've noticed is "cardboarding," the
>illusion that the picture is composed of different planes of objects
>at different distances, rather than continuously varying depth.  Is
>this also a Q-dos-specific problem?  What would cause that?

You can see two red and blue 'pupils' by looking through the front
of the lens at a white light source.  Put a ruler across the front
and measure the distance between their middles.  That should be the
effective separation of your two 'lenses' which will surely be less
than the average human eye separation of 65mm.  Greater separation
improves depth discernment, lesser separation reduces it.  Since
multitudinous other depth clues are lost in all forms of photography,
including SL3D, the mind which can't discern that an object has depth
may be able to see that it is closer/farther than other objects,
thus 'cardboarding'.

John B. writes:

>Now wait a minute!  8-)  Didn't someone earlier say that if there was black 
>in the scene, it would appear black in both images and so mess up the depth 
>because you couldn't control these black images independently?  I thought 
>the theory expressed earlier forbade this.  But then again, what about Q-DOS?  
>Well I'm confused.  Anyone want to explain it again, bringing these two 
>examples into the theory?

A large black object with no clear edges will look flat in any kind
of stereo photography.  But if there are distinguishable objects in front or
behind it, then their mutual alignments will look different to your
left and right lenses.  Same in Q-dos.  Again, the principle is that
an anaglyph created in a single-lens system consists of two images taken
from two different perspectives, which are recorded on the same piece
of film, and reconstructable through color selection.

Try thinking of it as two-aperture 3D, maybe that will help.

Paul Kline
pk6811s@xxxxxxxxxxxxxx


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