The split prism is an interesting optical device and what you say makes
sense. The split prism works in the following fashion: Think of the light
coming to focus on the screen. The light focussed by the lens arrives at a
point of focus shaped like a cone. Broad at the front surface of the lens
and the tip of the cone at the screen (called a pencil of light). This is
called convergence of light after being refracted by the lens system.
The light after coming to a focus at the point diverges outwards so that it
blurs once again. So you can picture it like two cones with point to point
as thus but in 3D--> X. The split prism shows a split image when the image
is in any of the areas that are off the point and the reason why at the
focus point the split image suddenly comes together. I don't quite know how
the focal length affects the pencil of light coming through the lens. It
probably depends on the exit pupil size of the lens more than any other
factor but this is just my speculation.
This optical mechanism of alignment in the split prism is very very accurate
because humans have a higher Vernier acuity than what there visual acuity
would predict due to cortical processing of retinal images. Vernier acuity
is the ability to align lines or detect any misalignments. The visual acuity
of humans is limited at about 1 minute of arc, whereas the Vernier acuity is
down to the order of 5 seconds of arc (there are 60 seconds of arc per
minute).
The problem therefore does not depend on any inherent problems with the
split prism. It lies in the assumption that the screen is at the same
distance as the film and CH Ling's URL shows in a nice way how this can be
corrected although I wouldn't be as game to put tape in under my mirror pin.
It is also true that light comes to focus at different points depending on
wavelength composition (chromatic abberration) and therefore could affect
focus also. The human eye is maximally sensitive to different wavelengths
depending on its state of dark adaptation. I could list so many variables
that what becomes amazing is that we get any decent shots at all given all
the confounding factors.
Oben
----- Original Message -----
From: "Winsor Crosby" <wincros@xxxxxxxxxxxxx>
To: <olympus@xxxxxxxxxxxxxxx>
Sent: Wednesday, May 09, 2001 2:06 PM
Subject: Re: [OM] Gary's lens test...
> >Oben Candemir wrote:
> >>
> >> The problem here I believe can be two things:
> >>
> >> 1. Refractive error of the photographer. However the use of
microprism/split
> >> prism to focus eliminates this in most cases.
> >>
> >
> >Split image is technically proved to be not accurate enough if you
> >need to make use of the max resolution from your lens. According to
> >Popular Photography the AF cameras use this technique for focusing,
> >the basic accuracy is only around 45lp/mm, but there are some software
> >tracking method that can increase the focusing accuracy a bit but not
> >always work.
> >
> >Also, don't expect DOF work for serious lens test, for 35mm system, it
> >only mean you will get not less than 30lp/mm.
> >
> >C.H.Ling
>
> I believe that both the split image and the microprism depend, like a
> rangefinder, on a baseline. The baseline on a rangefinder is fixed.
> The baseline on the split image/microprism is related to the diameter
> of the lens and varies with the magnification of the focal length.
> That is why rangefinder camera focus with a fixed, long baseline, is
> more accurate up to about a 100mm focal length. After that the
> magnification of the lens on the SLR increases the baseline of the
> microprism/split image beyond what it would be with a fixed
> rangefinder. So, accuracy is dependent on focal length, at least for
> manual focus. Ground glass focus with WA lenses on an SLR is
> probably the way to go for the most critical focus. I am not sure how
> AF is related to that.
>
> Winsor
>
> Winsor
> --
> Winsor Crosby
> Long Beach, California
> mailto:wincros@xxxxxxxxxxx
>
> < This message was delivered via the Olympus Mailing List >
> < For questions, mailto:owner-olympus@xxxxxxxxxxxxxxx >
> < Web Page: http://Zuiko.sls.bc.ca/swright/olympuslist.html >
>
>
< This message was delivered via the Olympus Mailing List >
< For questions, mailto:owner-olympus@xxxxxxxxxxxxxxx >
< Web Page: http://Zuiko.sls.bc.ca/swright/olympuslist.html >
|