Re: description of camera design I would like to have

phil-news-nospam@xxxxxxxx wrote:
On Tue, 30 Sep 2008 22:34:47 -0400 J. Clarke
<jclarke.usenet@xxxxxxx>
wrote:

phil-news-nospam@xxxxxxxx wrote:
On Tue, 30 Sep 2008 16:17:23 -0400 J. Clarke
<jclarke.usenet@xxxxxxx>
wrote:

So how well does your custom-made sensor with its
short-flange-distance compensators work with normal flange
distance
lenses?

So how well does your stock sensor with its normal flange distance
work with telephoto lenses that have a recessed rear element?

Why would I want such a thing?

Maybe you don't want telephoto lenses. That doesn't mean no one
does.

I have numerous telephoto lenses that work just fine without a
recessed rear element. In any case, you have been asserting that the
benefit of your proposed design would be that it made the production
of wide angle lenses easier.

Once the telephoto focal length reaches beyond a certain point, its
rear element will end up being placed further away from the sensor.
It will
have nothing to do with the mounting flange distance, other that
specify
how much "tube" length is needed to mount.

Uh, Canon doesn't seem to have any trouble producing a "telephoto"
with a 10mm focal length and no recessed rear element, so what does
that do to your argument?

The point is, the distances from the rear lens element, or it real
or
virtual center point, does not matter. What matters is the angle of
incidence. I does affect the ability of sensors to pick up light.
It
also affects the graininess of film. But these are things that can
be
compensated for.

But what does one gain to justify the cost of such compensation?

My point: the distance per se is not relevant.

Leica says it is, you say it isn't. Leica forgot more about camera
design between breakfast and lunch than you are ever going to know
so
I'll go with them.

You and I are the ones discussing this. I want to hear what YOU say
about a subject.

What I say is that I don't design cameras for a living, Leica does, if
they had problems then I'm satisfied that anybody will.

I am totally disinterested in what someone else says
about it UNLESS you have already said it (showing you understand
it),
and then using what someone else says to back up what you say if I
or
others disagree.

Well, I'm not going to play that game with you.

In other words, don't just point at somewhere else and leave it at
that.
Say it! Or don't bring it up.

You really don't want me to say what I think.

The only issue is
how well the sensor deals with greater angle of incidence. This
is
an issue that involves both the closeness of a lens rear element,
as
well as the opening size of the rear element.

How does "the opening size of the rear element" enter into it?

Angle of incidence.

Prove it. Demonstrate how the geometry changes with aperture.

Leica has problems with 35mm lenses. You're claiming that 10mm
won't
have problems. Leica has been designing cameras for more than 80
years. How long have you been designing cameras?

The way you describe this, it seems ONLY Leica has problems. If
they
are the only ones having problems, then THEY need to deal with it.

Leica is the only major manufacturer of interchangeable lens 35mm
cameras that produces something that is not an SLR. Nobody else has
the problem because nobody else has the very short flange distances
and deeply recessed lenses that you seem to be advocating.

I'm interested in the problems that are in basic theory and how they
can
be applied. Maybe Leica's problems are in manufacturing process.

No, their problem is that they have the shortest flange distance in
the industry and lenses with heavily recessed rear elements that were
never designed for use in an SLR or digital camera.
I
have
a hard time believing their _engineers_ would design something
faulty
in
the lab, and approve it as working when it does not, and then send
it
to manufacturing.

So who said it doesn't work? But it's expensive as Hell, like
everything Leica, and it's a crutch to deal with a design that was
never intended to be used with a digital sensor.

More likely, the manufacturing process could not
succeed
at some kind of quality control the engineers were able to do in the
lab.

Why don't you actually FIND OUT WHAT THE HELL YOU ARE TALKING ABOUT
before you talk trash?

If I did design a camera in a lab, and if it produced pictures with
a
purple tint and were vignetted, I would NOT be releasing it as a
final
design.

So what?

Just how much do you know about optics and lenses?

Enough to know that your proposed system does not confer the benefits
you claim.

If not much, then
it is pointless to continue between us. In that case, invite the
Leica engineers to join in. Maybe they can discuss application of
optical theory.

Since this matter is of such great concern to you, perhaps you should
contact them and ask them to do so.

I don't see where having the lens so close to the sensor that
you
need
to go to such lengths to get the sensor to function gains you
anything.

The angle of incidence is not as much of an issue with smaller
aperture lenses, which extreme wideangle tends to be. Would you
want
f/1.4 in a 10mm so much? A 10mm could be done in f/4.0 and
serve
most needs quite well. Even in f/5.6.

Huh? How is the angle of incidcence of light striking the sensor
different for an f/1.4 10mm from an f/4.0 10mm? And why would
one
need this special camera to use a 10mm? My 10mm is perfectly
satisfactory on a standard EOS mount.

The difference is in the wider aperture as seen in the rear
element.

Which would appear to reduce the required acceptance angle, not
increase it.

With a wider aperture, the light paths that impact a sensor cell or
a
film grain, will arrive at angles more away from the "straight on"
path that you would get with an f/256 pinhole "lens".

If you have a pinhole that is attempting to cover the full width of
the sensor, then it would seem to me that the angle of incidence at
the edges would be much shallower than if you have a lens whose
aperture is nearly that of the full width of the sensor doing so.

Try drawing yourself a picture and do some ray traces and see what you
come up with.

Although this
is 90 degrees from
the angle of the sensor or film, it is the 0 degree incident
reference.
Light arriving at other angles are then measured relative to that
"straight on" path. The wider the aperature is from the perspective
of that cell or grain, the greater these angles of incidents are
from
that reference path.

The word is _incidence_, not "incidents". And what makes you think
that a range of angles of incidence is a problem? The problem that
Leica encountered was that the angle of incidence at the edges of the
senser was extremely shallow.

There are two issues here.

One is that sensitivity varies with the angle of incidence. Film
dealt with this issue well enough because grains were mostly
randomly
oriented. So while some grains might be less sensitive to light
coming from some angle, other grains would be more sensitive. On
average it would work out well. But you can see this by creating an
exposure situation where there is exaggerated angles of incidence
with different colors from each angle. Photograph, on film, a two
color object with a sharp edge between the colors, but adjusted to
be
well out of focus, using a very grainy fast film (ISO 3200), shot at
one or two stops under, and push the development accordingly. Red
and blue would be good.

Why don't you show us an example?

A digital sensor is different. It doesn't have the random
orientations of film. But it can be addressed. Each cell can be
made to have a bump that can accept a wider range of angles of
incidence. And the exposure variation that differs between cells
can
be compensated for in software processing based on knowing the
aperture of the exposure.

And why would one want to go through all this rigmarole when simply
having an adequate distance between the rear element and the sensor
does the job adequately?

The other is that from "off angles" a flat aperature opening will be
more oval shaped and vary in size as "seen" from that angle (even if
the cells
and grains have no difference in sensitivity at different angles).
This is also possible to compensate for, and is more easily done for
digital.

Then why is it that cameras that for more than half a century have
been considered to be outstanding with film cannot easily be updated
with a digital sensor?

It's amazing how simple the engineer's job is to the layman who
doesn't actually have to _do_ it.

--
--
--John
to email, dial "usenet" and validate
(was jclarke at eye bee em dot net)


.

Relevant Pages

  • Re: Film slr lenses on DSLR question
    ... SLR wide angle lenses can cause problems on DSLR's because of the ... The fact there is a digital sensor at the end of the optical path ... Not that it varies from camera to camera, but more that it varies from lens ...
    (alt.photography)
  • Re: description of camera design I would like to have
    ... |> So how well does your stock sensor with its normal flange distance ... |> work with telephoto lenses that have a recessed rear element? ... |> how well the sensor deals with greater angle of incidence. ...
    (rec.photo.digital)
  • Re: Canon Response to 5D Dark Corners
    ... as extreme an angle of incidence as the Canon lens mount permits. ... There was essentially no difference in sensor sensitivity between these ... But you didn't simulate the light that falls on the sensor from a lens. ... Exactly, because the argument that you have made is that it isn't the lens that causes the problem, but that the *sensor* somehow has a different response to angle of incidence than film has. ...
    (rec.photo.digital.slr-systems)
  • Re: Confused about "Digital" Lens
    ... a perpedicular angle. ... corners of the sensor. ... telephoto" design of most wide angle lenses -- for film or digital. ... because this whole "digital lens" thing is just a load of crap? ...
    (rec.photo.digital.slr-systems)
  • Re: SLR lens on DSLRs
    ... >>> you should mention that what you say only applies to wide angle ... >>> lenses on FF digital cameras. ... >> rear element in a wide angle lens is from the sensor in a DSLR. ...
    (rec.photo.digital)