Re: description of camera design I would like to have

phil-news-nospam@xxxxxxxx wrote:
On Thu, 2 Oct 2008 01:37:15 -0400 J. Clarke <jclarke.usenet@xxxxxxx>
wrote:

The specifics are that you have asserted with no proof that one can
easily produce a camera with a shorter distance between the rear
lens
element and the sensor than is current practice with DSLRs. The
specifics are that Leica has been producing such cameras for three
quarters of a century. The specifics are that when Leica decided
to
produce a digital version of that camera they found that existing
sensor designs were not satisfactory. The specifics are that Leica
had to have a special sensor designed that works only in their M8.
The specifics are that even so Leica could not obtain acceptable
performance over the full 35mm frame size.

You are the one that doesn't understand the issues involved. If you
did you could explain those issues and not need to have someone else
(who is not online here to answer pointed questions) do it for you.
Since you didn't just explain what the problems are, I interpret
that
action by you as something representing some combination of lack of
science understanding and you preference to just divert people you
don't like out of a discussion by trying to get them to deal with
something else in its place.

Interpret it any way you want to. I interpret your failure to
research the issues as being laziness.

Now, maybe you are smarter than Leica. If so I am sure that they
would be happy to hire you for whatever salary you ask.

Maybe. Maybe not. But they aren't a member of this group. You and
I
are. But you can't hold down a discussion of a
technical/science/theory topic on your own. Maybe I would discuss
this with someone from Leica
if they were here to discuss it. But you clearly won't or can't.

I'm not interested in typing a bunch of stuff that someone else has
already typed adequately.

In other words you want me to reinvent the wheel by explaining
things
to you that Leica has already explained adequately.

That's exactly how a _discussion_ works. It is NOT a discussion to
go
read some document ... unless the AUTHOR of the document is the one
who wrote it AND can answer questions about it (whether they be
questions about suspected errors, or just questions being asked to
better understand what was written about).

If we were sitting in a pub that would be correct, but we are not.

YOU are not discussing the science/theory/engineering. YOU are just
trying to pawn it off to somewhere that questions cannot be
answered.
Perhaps this is because you don't like this camera design idea and
want
to see the discussion of it come to a silent end. Or maybe you
really
do believe there is a technical flaw that would be impractical to
solve. But if it is the latter case, you certainly don't seem to
know
enough about it to discuss it.

Once again instead of actually looking something up and reading it you
blather about how evil others are who expect you to do your homework
before pontificating.

Clue 1: You can easily avoid being involved in a discussion by not
posting.

Clue 2: You can avoid being tempted to be involved in a discussion
by
not reading the thread at all.

So take your two clues and act upon them.

Prove it. Demonstrate how the geometry changes with aperture.

This is too basic. If this is something you cannot understand,
then
it
would be way too much trouble for me to bring you up to speed on
it.

In other words you can't.

It isn't the topic of the thread (yet). You are welcome to make it
the topic. But demanding a demonstration is not a discussion of it,
so it does not make it the topic.

You made the assertion, when asked to support it you first resorted to
insult and when that failed claimed it was off topic. The bottom line
is that you can't back up your assertion and you're too arrogant to
just admit it.

I think it's time for this thread to be done. It's clear I will
not
get anywhere with someone that does not even begin to understand
optics.

In other words you can't back up your assertions.

I won't back up assertions HERE in this thread, for topics that
really
need to be in their own thread.

Cop-out.

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 not advocating recessed lenses. I merely used recessed lenses
as
an example in the reverse when you said distance has no effect.

Please provide a quotation to support that statement.

While editing a post, I don't have the option to go back to scan all
your posts. If I get around to it later on, I might.

In other words you don't have any facts to support your assertion, as
usual.

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.

Are you talking about the rear elements being recessed?

I'm talking about Leicas, in which with some lenses the rear
element
was so close to the shutter that it would hit the metering sensor
on
the M5.

I've never had a Leica, or even handled one, much less pull the lens
off to see inside.

Pity.

So I don't know just HOW close they did that, or
even what kind of rear lens element is involved. So I cannot
comment
on their specific design. I may be able to add something if you can
provide certain parameters about it:

1. focal length in mm
2. largest aperture in f/stop
3. size of frame in mm (36 x 24 I presume)
4. distance from physical rear element to film in mm (not shutter)
5. optical distance from film to aperture opening in mm (as
modified
by all lens elements together behind the opening).

What I would want to measure, or with the complete design figure
out,
is just how much that aperture opening cross sectional area changes
with angle of view from the film, at various points from center of
the
frame to the farthest corner (21.6333 mm).

And no, I do NOT have enough interest to try to find engineering
diagrams of Leica lenses on the Leica site.

Yeah, that's the basic problem, you can't be assed to actually
research designs that present problems similar to those that you
advocate.

The particular lenses are the Super Angulon 21mm and the 28mm Elmarit,
both of which have rear elements that extend to within 8mm of the film
plane.

For the specific details you will need to contact Leica or in the case
of the Super Angulon Schneider. The Super Angulon is supposed to be
based on the Zeiss Biogon formula.

If you would bother to actually research a bit before spouting off
then we would not be having this discussion.

The research I did was sufficient to be relevant.

If you did not discover that a major and well respected manufacturer
had produced a design that had the reduced lens-to-sensor distance
that you advocate, and had problems with it indicates that the
research that you did was cursory at best.

So tell us what Leica did wrong, since you're so sure that they
screwed up.

You are the one who thinks what they did wrong is relevant. So you
need to be the one saying it.

You are the one asserting that they did something wrong.

Uh, what basic engineering principles are they not following? And
it's not "my precious Leica". If you aren't aware of Leica's
reputation and standing in the optics industry then you really have
no credibility at all.

I don't know what basic engineering principles are they not
following.

Then why are you asserting that they are not following such
principles? This seems to be simple empty-headed bashing.

I'm sure it has something to do with the digital aspect, as opposed
to
the lens aspect. But they may have preferred a lens solution rather
than a digital solution.

Uh, their sensor is required to work with every lens that they have
produced in the past 75 years except those that cannot be attached to
an M mounting.

Because Leica tried it and found that it was difficult, expensive,
and
the result left a good deal to be desired. You have not been able
to
point out how you would overcome the difficulties they encountered.

Maybe if they were working on an SLD design, it would be different.
Afterall, the RF design is not the same thing as an SLD design.

In what way is it different that is relevant to the issue of problems
inherent in a reduced lens-to-senstor distance?

There is no "diverging". You made an assertion that something was
easy. I gave an exmaple of a company with vast expertise and great
incentive to do it that found that it was not easy. You don't seem
to
want to accept that.

I'm not sure you even correctly identified "it". They had problems.
But they were not doing an SLD design. I'm still waiting to hear
from
you how you think their experiences apply to an SLD design.

You have asserted that the major benefit of the "SLD" design is
reduced lens-to-sensor distance when compared to an SLR. The Leica M
has that reduced distance.

I'll let Canon or Nikon do that. If Leica has solved some issue
and
has intellectual property on the solution, then Canon, Nikon, or
others
can license the IP.

Why would Leica license technology to a competitor?

For money? It happens. It happens a lot in the high end markets,
especially when the licensee is going into a lower end market. That
kind of thing lets them "get revenue a bite" out of a market they do
not want to dilute their trademark in.

When has Leica ever done 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.

I've done that geometry before.

Then scan it and put it up on flickr.

I have my own web servers.

Then put it up on them.

I doubt I have any of those papers,
anymore. Things do get tossed when moving. This was 25+ years ago
(before the internet came to the masses).

In other words all we have is your word for it.

Is that all that Leica encountered? You realize that when you
stop
down the aperature, you are approaching a pinhole.

Which means that all the light is coming from a point and none is
coming from directly above the pixel unless the pixel is beneath
the
exact center of the lens.

So. This is how lens imaging works.

To a first approximation.

And even if it is not a pinhole,
and is instead a wide aperature like f/2.0, pixels off-center get
less
of the light because more of the vectors are off at a greater angle.

Which is irrelevant to the problem at hand, which is more complex than
simple vignetting.

If the sensor is 22mm wide and the pinhole is 11mm from the angle
of
incidence is 45 degrees. That gives a reduction of 30% of light.
That's about 2/3 stop, and well within the range to compensate by
software.

And what is it after it has reflected off of the surface of the
microlens?

Too many variables here. What kind of microlens?

The kind that is on the sensor.

I'd look into going with a silicon bump design, not a lens overlay.

Still a microlens.

Not that this would necessarily be the best solution, but it could
be one alternative. Leica, being a lens company, might not have
even
considered it.

Uh, the sensor was designed by the same company that designed the
sensors that your beloved Hassleblad uses.

I did those experiments 25 years ago. I have no idea where the
negatives are. I'm not prepared to repeat the experiment right
now.

if you don't believe it, then I guess I'll just have to leave you
unconvinced. It is not practical ... and not worthwhile ...
trying
to convice you or teach you anything.

In other words you vaguels remember something that you did a long
time
ago.

I establish summaries and remember what I need to know. That's
easier
to do than remembering details. And back in those days, details
went
to paper, which turns out to be a lossy medium (in terms of losing
the
whole piece of paper).

Which doesn't alter the fact that you are reporting something that you
may or may not be remembering accurately.

Having done image processing by programmed software (software I
write, not
the use of some application program that may or may not really be
doing
what is needed), I can tell you that the artifacts that result
from
trying
to correct lens aberrations (chromatic and geometric) are greater
than what results from merely pixel by pixel exposure
compensation,

What makes you think that "pixel by pixel exposure compensation" is
sufficient?

What makes you think it isn't?

The fact that it wasn't.

At the lens distances I envision for
the SLD design, the light dilution at corners would not be less than
1/4. And that's within the range of compensation using a 14-bit
encoding (giving a 12-bit compensated range).

Again how much is it after you count losses from non-optimal
orientation of the microlenses, which losses are not calculable by
simple geometry?

especially with sensors like what Canon has in the 450D and 50D
that
gives 14 bits per
color plane per pixel. If it were only 8 bits, I might be more
inclinded
to say the geometry is easier to fix by software. But with what
we
have today, I'd rather have lenses with good geometry. And that
means lenses
with fewer compensating elements (which in turn helps reduce the
total
amount of reflections).

Again you are assuming that a digital sensor is film. It is not.

How is it that describing the differences makes you think that I am
not saying there are differences? How is it you think that means I
am saying there are no differences? Or is it that you didn't even
understand what I said?

I understood very clearly that you are bogged down in bits per pixel
and are not considering differences in the physical structure of a
digital sensor vs film.

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?

There are differences in how digital sensors behave, compared to
film. There are similarities, too.

And yet you ignore those differences in all of your discussion
above
in which you seem to think that vignetting is the issue.

The effect is LIKE vignetting. It's actually easier to calculate
than
real vignetting. It's smoother that common vignetting. The light
value drops gradually from center to corner.

You still have not addressed the issues that Leica encountered.

You never worked for Aircraft Armaments Incorporated, did you?

And this is relevant how?

You remind me of someone who nearly blew up an Air Force transport
once.

And my participation in this endless (and now determined to be
pointless) so called "discussion" is coming to a close.

Hope springs eternal.

I wonder why you even bother to post at all, given you never
intended
to discuss the science/theory or market/business aspects of the
camera design I proposed. You only intended to throw it over to
Leica.

Actually I intended to point out to you that your wonderful design
that is so simple to implement isn't quite as simple as you claim.
But you have chosen to argue ad-nauseum instead of taking the hint and
looking up the very, very well publicized problems that Leica had
implementing the M8.

I has assumed that you were actually interested in the practicality of
your proposed line.
--
--
--John
to email, dial "usenet" and validate
(was jclarke at eye bee em dot net)


.

Relevant Pages

  • Re: description of camera design I would like to have
    ... | easily produce a camera with a shorter distance between the rear lens ... The specifics are that when Leica decided to ... Perhaps this is because you don't like this camera design idea and want ...
    (rec.photo.digital)
  • Re: description of camera design I would like to have
    ... rear element will end up being placed further away from the ... What size of sensor are you talking about? ... Leica says it is, you say it isn't. ... design between breakfast and lunch than you are ever going to ...
    (rec.photo.digital)
  • Re: description of camera design I would like to have
    ... high end super-pro camera market business decisions ... had nothing to do with making a better lens design. ... plane can have more issues with a digital sensor than with film. ...
    (rec.photo.digital)
  • Re: description of camera design I would like to have
    ... fit the SWC because the SWC takes the same backs as their other ... of the many other discussions of the M8 sensor. ... I'm focused on the SLD design, not Leica's attempt at going digital. ... hence the lens can be placed closer to the sensor. ...
    (rec.photo.digital)
  • Re: Digital vs 35mm camera optics
    ... and optical complexity for either sensor. ... Since the lens is smaller, ... design more than apparent resolution. ... is that the digital camera lens could exhibit more relaxed tolerance ...
    (sci.optics)