Sign in with
Sign up | Sign in
Your question

Poll on *Really* Wide Angle Lenses

Last response: in Digital Camera
Share
August 5, 2005 4:44:09 PM

Archived from groups: rec.photo.digital.slr-systems,rec.photo.digital (More info?)

If a rectilinear 8-28mm zoom or perhaps rectilinear primes in the
6-10mm range were available, would you be seriously interested? Focal
lengths listed are in full-frame 35mm equivalents, so for a DX format
the above-mentioned zoom would have an acutal focal length range of
5.33 - 18.67mm). Distortion would be nearly zero, as would
illumination falloff.

Brian

More about : poll wide angle lenses

Anonymous
August 5, 2005 4:52:32 PM

Archived from groups: rec.photo.digital.slr-systems,rec.photo.digital (More info?)

BC wrote:

> If a rectilinear 8-28mm zoom or perhaps rectilinear primes in the
> 6-10mm range were available, would you be seriously interested? Focal
> lengths listed are in full-frame 35mm equivalents, so for a DX format
> the above-mentioned zoom would have an acutal focal length range of
> 5.33 - 18.67mm). Distortion would be nearly zero, as would
> illumination falloff.

I would only consider this optic if, and only if, in addition to no
illumination falloff, no distortion, it also reduced diffraction
effects to zero and possessed absolutely no aberrations of any kind.
Additionaly, the focus motors would also have to be perpetual motion
machines -- let them be powered by the vacuum energy.
Anonymous
August 5, 2005 7:15:19 PM

Archived from groups: rec.photo.digital.slr-systems,rec.photo.digital (More info?)

Anyone can find a use for any lens, even distorted ones like a lensbaby
or fisheye. Don't see why someone couldn't use something like this.
Related resources
Anonymous
August 5, 2005 7:52:33 PM

Archived from groups: rec.photo.digital.slr-systems,rec.photo.digital (More info?)

I never wanted anything for my 35mm SLR wider than 20mm, and I only
used that once in a very long while. So having a 10-22mm for my 20D is
more than enough for me, and I doubt that I would use wider than 12mm
most of the time, unless I was after the 'dramatic' or 'unusual'
perspective of the 10-12 range. Equivalent to 16mm in the 35mm FF
format, 10mm probably won't get that much use from me, but I never had
it so easy to get at before (in a zoom, rather than having to always
switch lenses).

--Wilt
August 6, 2005 12:50:41 AM

Archived from groups: rec.photo.digital.slr-systems,rec.photo.digital (More info?)

On 5 Aug 2005 12:52:32 -0700, "eawckyegcy@yahoo.com"
<eawckyegcy@yahoo.com> wrote:

>BC wrote:
>
>> If a rectilinear 8-28mm zoom or perhaps rectilinear primes in the
>> 6-10mm range were available, would you be seriously interested? Focal
>> lengths listed are in full-frame 35mm equivalents, so for a DX format
>> the above-mentioned zoom would have an acutal focal length range of
>> 5.33 - 18.67mm). Distortion would be nearly zero, as would
>> illumination falloff.
>
>I would only consider this optic if, and only if, in addition to no
>illumination falloff, no distortion, it also reduced diffraction
>effects to zero and possessed absolutely no aberrations of any kind.
>Additionaly, the focus motors would also have to be perpetual motion
>machines -- let them be powered by the vacuum energy.


as long as it didn't cost too much.
Anonymous
August 6, 2005 2:41:53 AM

Archived from groups: rec.photo.digital.slr-systems,rec.photo.digital (More info?)

In article <1123271049.621406.294920@g14g2000cwa.googlegroups.com>, BC
<brianc1959@aol.com> writes
>If a rectilinear 8-28mm zoom or perhaps rectilinear primes in the
>6-10mm range were available, would you be seriously interested? Focal
>lengths listed are in full-frame 35mm equivalents, so for a DX format
>the above-mentioned zoom would have an acutal focal length range of
>5.33 - 18.67mm). Distortion would be nearly zero, as would
>illumination falloff.
>
>Brian
>
Brian, at first sight it sounds very attractive. However, from using a
14mm ffl lens on a 35mm film body, and getting to know the serious
geometric effects inherent in rectilinear drawing (especially in the
corners) I would be quite wary of what the images looked like. I felt
that 14mm was about as far as I would wish to go in that format -
indeed, it's one for occasional use only. For a smaller sensor DSLR,
then make that 9 or 10mm, of course.

Not saying I would rule it out, simply that I would view it with some
curiosity and a genuine open mind, but would have some expectations of
not liking the images too much. Even accepting you could cure the
illumination and linearity (and I'm sure it could be done) this drawing
problem is, as you know, completely inherent in the geometry of the
situation; you can get the lines straight or the area right, but not
both.*

And, of course (as someone already said) it would depend to some extent
on the price.

What I really would like would be a 10-12mm ffl lens that worked on my
10D and cost about £400. Dream on ....

*If I'm wrong, and there is a way of squaring this particular circle,
then all bets are off, of course!

David
--
David Littlewood
Anonymous
August 6, 2005 2:41:54 AM

Archived from groups: rec.photo.digital.slr-systems,rec.photo.digital (More info?)

With, DSLR, why not just take multiple shots and stitch them together?
Panorama software is much cheaper than an expensive lense. Is there
something I am missing? (I am a newbie to all of this cool stuff!)

David Littlewood wrote:
> In article <1123271049.621406.294920@g14g2000cwa.googlegroups.com>, BC
> <brianc1959@aol.com> writes
>
>> If a rectilinear 8-28mm zoom or perhaps rectilinear primes in the
>> 6-10mm range were available, would you be seriously interested? Focal
>> lengths listed are in full-frame 35mm equivalents, so for a DX format
>> the above-mentioned zoom would have an acutal focal length range of
>> 5.33 - 18.67mm). Distortion would be nearly zero, as would
>> illumination falloff.
>>
>> Brian
>>
> Brian, at first sight it sounds very attractive. However, from using a
> 14mm ffl lens on a 35mm film body, and getting to know the serious
> geometric effects inherent in rectilinear drawing (especially in the
> corners) I would be quite wary of what the images looked like. I felt
> that 14mm was about as far as I would wish to go in that format -
> indeed, it's one for occasional use only. For a smaller sensor DSLR,
> then make that 9 or 10mm, of course.
>
> Not saying I would rule it out, simply that I would view it with some
> curiosity and a genuine open mind, but would have some expectations of
> not liking the images too much. Even accepting you could cure the
> illumination and linearity (and I'm sure it could be done) this drawing
> problem is, as you know, completely inherent in the geometry of the
> situation; you can get the lines straight or the area right, but not both.*
>
> And, of course (as someone already said) it would depend to some extent
> on the price.
>
> What I really would like would be a 10-12mm ffl lens that worked on my
> 10D and cost about £400. Dream on ....
>
> *If I'm wrong, and there is a way of squaring this particular circle,
> then all bets are off, of course!
>
> David
Anonymous
August 6, 2005 4:07:38 AM

Archived from groups: rec.photo.digital.slr-systems,rec.photo.digital (More info?)

I like wide angle, but damned if I'd go past 15-16mm equivalent for
rectilinear.
--
http://www.pbase.com/bcbaird
Anonymous
August 6, 2005 4:43:48 AM

Archived from groups: rec.photo.digital.slr-systems,rec.photo.digital (More info?)

In article <xIqdnX3HMrpdm2nfRVn-rg@rogers.com>,
nicholas.wittebol@rogers.com says...
> With, DSLR, why not just take multiple shots and stitch them together?
> Panorama software is much cheaper than an expensive lense. Is there
> something I am missing? (I am a newbie to all of this cool stuff!)

You can get some neat shots from the perspective effects present when
using a wide angle lens... and there's times where it isn't practical or
easy to take multiple shots to stitch things together.
--
http://www.pbase.com/bcbaird
Anonymous
August 6, 2005 4:43:49 AM

Archived from groups: rec.photo.digital.slr-systems,rec.photo.digital (More info?)

Ahhhh...I see. I'm learning...slowly! ;-)

Brian Baird wrote:
> In article <xIqdnX3HMrpdm2nfRVn-rg@rogers.com>,
> nicholas.wittebol@rogers.com says...
>
>>With, DSLR, why not just take multiple shots and stitch them together?
>>Panorama software is much cheaper than an expensive lense. Is there
>>something I am missing? (I am a newbie to all of this cool stuff!)
>
>
> You can get some neat shots from the perspective effects present when
> using a wide angle lens... and there's times where it isn't practical or
> easy to take multiple shots to stitch things together.
Anonymous
August 6, 2005 4:59:18 AM

Archived from groups: rec.photo.digital.slr-systems,rec.photo.digital (More info?)

"Brian Baird" <no@no.thank.u> wrote in message
news:MPG.1d5dd1d06e980769989799@news.verizon.net...
> In article <xIqdnX3HMrpdm2nfRVn-rg@rogers.com>,
> nicholas.wittebol@rogers.com says...
>> With, DSLR, why not just take multiple shots and stitch them together?
>> Panorama software is much cheaper than an expensive lense. Is there
>> something I am missing? (I am a newbie to all of this cool stuff!)
>
> You can get some neat shots from the perspective effects present when
> using a wide angle lens... and there's times where it isn't practical or
> easy to take multiple shots to stitch things together.
> --
> http://www.pbase.com/bcbaird

I concur.........*High quality* panorama's (of the stitched variety) aren't
exactly a cake walk - key word is high quality.
David
Anonymous
August 6, 2005 1:08:26 PM

Archived from groups: rec.photo.digital.slr-systems,rec.photo.digital (More info?)

Charles wrote:
> On 5 Aug 2005 12:52:32 -0700, "eawckyegcy@yahoo.com"
> <eawckyegcy@yahoo.com> wrote:
>
>>BC wrote:
>>
>>> If a rectilinear 8-28mm zoom or perhaps rectilinear primes in the
>>> 6-10mm range were available, would you be seriously interested? Focal
>>> lengths listed are in full-frame 35mm equivalents, so for a DX format
>>> the above-mentioned zoom would have an acutal focal length range of
>>> 5.33 - 18.67mm). Distortion would be nearly zero, as would
>>> illumination falloff.
>>
>>I would only consider this optic if, and only if, in addition to no
>>illumination falloff, no distortion, it also reduced diffraction
>>effects to zero and possessed absolutely no aberrations of any kind.
>>Additionaly, the focus motors would also have to be perpetual motion
>>machines -- let them be powered by the vacuum energy.
>
>
> as long as it didn't cost too much.

It should also have a lifetime warranty and a universal mount that adapts to
any camera mount with full functionality and no degradation in performance.
Anonymous
August 6, 2005 2:59:00 PM

Archived from groups: rec.photo.digital.slr-systems,rec.photo.digital (More info?)

"BC" <brianc1959@aol.com> wrote in message
news:1123271049.621406.294920@g14g2000cwa.googlegroups.com...
> If a rectilinear 8-28mm zoom or perhaps rectilinear primes in the

Zooms *are* primes. A prime lens (popular misusage notwithstanding) is the
camera lens as opposed to some auxiliary lens used with it. Makes no
difference whether it's a zoom or fixed focal length (FFL), a prime lens is
still a prime lens.

N.
Anonymous
August 6, 2005 3:00:39 PM

Archived from groups: rec.photo.digital (More info?)

"Nicholas Wittebol" <nicholas.wittebol@rogers.com> wrote:
> With, DSLR, why not just take multiple shots and stitch them together?
> Panorama software is much cheaper than an expensive lense. Is there
> something I am missing? (I am a newbie to all of this cool stuff!)

Doing that is (a) a pain*, and (b) results in a cylindrical projection,
which can result in serious distortion**. If you get a dSLR that takes a
standard mount, then the Stigma 12-24 will mount on a cheap film body, and
you can use that for your extreme wide angle work. I didn't do that because
I was afraid I'd spend too much time shooting 12mm shots on film and not do
any digital<g>.

*: Although it's not bad. I find that if I shoot on a tripod, Panorama
Factory will stitch nicely with no intervention, as it did in the image
below.

*: http://www.pbase.com/davidjl/image/47227687/large

David J. Littleboy
Tokyo, Japan
Anonymous
August 6, 2005 3:00:40 PM

Archived from groups: rec.photo.digital (More info?)

On Sat, 6 Aug 2005 11:00:39 +0900, "David J. Littleboy"
<davidjl@gol.com> wrote:

<snips>

>If you get a dSLR that takes a
>standard mount, then the Stigma 12-24 will mount on a cheap film body, and
>you can use that for your extreme wide angle work.

That's exactly what I did. Except the idiots at Sigma neglected to fit
an aperture ring on the lens, so there's no way to adjust the
diaphragm on my old manual film bodies. Indeed it remains stopped all
the way down, unless you put a 'kludge' against the pin before
mounting the lens. Unsatisfactory, but the only way to get a 12-24mm
zoom on my old Canon F-1!

Mike
--
http://www.corestore.org
'As I walk along these shores
I am the history within'
Anonymous
August 6, 2005 3:00:40 PM

Archived from groups: rec.photo.digital (More info?)

In article <dd15uu$d0s$1@nnrp.gol.com>, davidjl@gol.com says...
>
> "Nicholas Wittebol" <nicholas.wittebol@rogers.com> wrote:
> > With, DSLR, why not just take multiple shots and stitch them together?
> > Panorama software is much cheaper than an expensive lense. Is there
> > something I am missing? (I am a newbie to all of this cool stuff!)
>
> Doing that is (a) a pain*, and (b) results in a cylindrical projection,
> which can result in serious distortion**. If you get a dSLR that takes a
> standard mount, then the Stigma 12-24 will mount on a cheap film body, and
> you can use that for your extreme wide angle work. I didn't do that because
> I was afraid I'd spend too much time shooting 12mm shots on film and not do
> any digital<g>.
>
> *: Although it's not bad. I find that if I shoot on a tripod, Panorama
> Factory will stitch nicely with no intervention, as it did in the image
> below.
>
> *: http://www.pbase.com/davidjl/image/47227687/large
>
> David J. Littleboy
> Tokyo, Japan

Have you tried Panorama Tools (ptools)? That gives you the option of
rectilinear projection for your final output.
--
http://www.pbase.com/bcbaird
Anonymous
August 6, 2005 3:00:40 PM

Archived from groups: rec.photo.digital (More info?)

David J. Littleboy wrote:
> "Nicholas Wittebol" <nicholas.wittebol@rogers.com> wrote:
>> With, DSLR, why not just take multiple shots and stitch them
>> together? Panorama software is much cheaper than an expensive lense.
>> Is there something I am missing? (I am a newbie to all of this cool
>> stuff!)
>
> Doing that is (a) a pain*, and (b) results in a cylindrical
> projection, which can result in serious distortion**.

Only if you ask for cylindrical projection, surely? I've seen programs
which allow spherical and rectilinear projection as well.

David
August 6, 2005 3:01:26 PM

Archived from groups: rec.photo.digital.slr-systems,rec.photo.digital (More info?)

Chris Brown wrote:
>
> Illumination falloff is a direct consequence of the lens being rectilinear.
> It happens because, by maintaining the correct angles near the egde of the
> frame, you stretch the light out (that's why circles turn into ellipses).
> Same light over a larger area = illumination falloff. The only way to
> mitigate it is to use a circular ND grad filter, that is denser in the
> middle, and then expose for longer.
>
> The only way to avoid it without resorting to graduated filters is to make
> the lens a fisheye.
>
> Furthermore, you get diminishing returns in terms of nagle of view with
> shorter rectilinear lenses. You can't ever get to 180 degrees - it's
> physically impossible (would require an infinitely large sensor). This too,
> is not a problem with fisheyes.
>
> Might I suggest that what Sir wants is a circular fisheye, with a 180
> degrees field of view, and a copy of Panotools to produce cropped
> rectilinear projections from it? This has the advantage that there really
> will be no light falloff, but the edge resolution will be lower than the
> middle, because you'll be interpolating.

Hi Chris:
I've already designed examples similar to the specs I listed, so I
wasn't just making up stuff that is physically impossible. One of the
lenses I presented at a recent conference covers 140 degrees at f/1.0
with less than 0.1% distortion. Nevertheless, the illumination in the
corners of the image is slightly *higher* than in the center. In this
case you would have to introduce some vignetting in order to achieve
uniform illumination. I've designed other rectilinear examples
covering up to 150 degrees that have almost perfectly uniform
illumination.

Two things determine illumination falloff. First is the angle of the
chief ray in image space (not object space), and second is pupil
aberration. An unvignetted telecentric lens with no pupil aberration
will have perfectly uniform illumination regardless of distortion or
field of view. With favorable pupil aberration you can actually have
an illumination curve that rises as you approach the corners, even when
distortion is zero. Numerous telecentric projection lenses have been
built that exhibit this behavior.

Agreed that you can't actually get to a full 180 degrees with a
rectilinear lens. But you can get alot wider than conventional wisdom
might lead you to believe. The old rectilinear limit was set by the
Goerz Hypergon at around 135 - 140 degrees, but this lens is only
useable at a tiny aperture and it has tons of illumination falloff. I
suspect the real practical limit is somewhere in the 160 to 170 degree
range. Such lenses are very different from fisheyes, and I'm
interested in finding out more about what market possibilities might
exist.

I'm aware of the Panotools option, since I've used it extensively for
about 4 years now. I've always felt the image degradation due to
interpolation which you mention is too severe unless working with
ultra-high resolution stitched images. I feel the same way about
correcting keystone distortion.

Also, if you shoot a flat target with an ultrawide angle rectilinear
lens, then circles in the object will imaged as circles in the image
plane. Its only when you consider three-dimensional objects that
things start to get strange.

Brian
August 6, 2005 3:18:13 PM

Archived from groups: rec.photo.digital.slr-systems,rec.photo.digital (More info?)

Whether we like it or not, language is one of the most democratic
things around. "Popular misusage" ineveitably becomes "correct". In
the case of zoom vs. prime, I just don't care any more.

Brian
August 6, 2005 3:22:50 PM

Archived from groups: rec.photo.digital.slr-systems,rec.photo.digital (More info?)

I suspect most people would agree with you, although I still have alot
of enthusiasm for a super-wide zoom, especially if it has some shifting
capability.

Brian
Anonymous
August 6, 2005 3:28:47 PM

Archived from groups: rec.photo.digital.slr-systems,rec.photo.digital (More info?)

In article <1123271049.621406.294920@g14g2000cwa.googlegroups.com>,
BC <brianc1959@aol.com> wrote:
>If a rectilinear 8-28mm zoom or perhaps rectilinear primes in the
>6-10mm range were available, would you be seriously interested? Focal
>lengths listed are in full-frame 35mm equivalents, so for a DX format
>the above-mentioned zoom would have an acutal focal length range of
>5.33 - 18.67mm). Distortion would be nearly zero, as would
>illumination falloff.

Illumination falloff is a direct consequence of the lens being rectilinear.
It happens because, by maintaining the correct angles near the egde of the
frame, you stretch the light out (that's why circles turn into ellipses).
Same light over a larger area = illumination falloff. The only way to
mitigate it is to use a circular ND grad filter, that is denser in the
middle, and then expose for longer.

The only way to avoid it without resorting to graduated filters is to make
the lens a fisheye.

Furthermore, you get diminishing returns in terms of nagle of view with
shorter rectilinear lenses. You can't ever get to 180 degrees - it's
physically impossible (would require an infinitely large sensor). This too,
is not a problem with fisheyes.

Might I suggest that what Sir wants is a circular fisheye, with a 180
degrees field of view, and a copy of Panotools to produce cropped
rectilinear projections from it? This has the advantage that there really
will be no light falloff, but the edge resolution will be lower than the
middle, because you'll be interpolating.
Anonymous
August 6, 2005 3:50:57 PM

Archived from groups: rec.photo.digital.slr-systems,rec.photo.digital (More info?)

In article <GXTIe.2993$sv.603@trnddc08>, David A says...

> I concur.........*High quality* panorama's (of the stitched variety) aren't
> exactly a cake walk - key word is high quality.

No problem obtaining high quality seamless panoramas, when using the
Panorama Tools with the layered output. People are even doing seamless
360° panoramas.
Also, the final resolution will be substantially higher, because you
stitch together many images. And the geometrical lens aberrations will
have been corrected.
--

Alfred Molon
------------------------------
Olympus 4040, 5050, 5060, 7070, 8080, E300 forum at
http://groups.yahoo.com/group/MyOlympus/
Olympus E300 resource - http://myolympus.org/E300/
August 6, 2005 3:57:16 PM

Archived from groups: rec.photo.digital.slr-systems,rec.photo.digital (More info?)

Jeremy Nixon wrote:

"Lenses are more than wide enough already. Make 'em faster instead.
Give me a good 50mm f/0.7, and we can talk about market possibilities.
:)  "

--

Unfortunately, the sensors used in all consumer cameras (including
DSLR's) cannot adequately handle the marginal ray angle from an f/1
beam (30 degrees), let alone from f/0.7 (45.6 degrees). By the time
you get to f/0.7 you may start getting substantial crosstalk between
pixels. I suspect this is the main reason that Canon no longer offers
a 50mm f/1 lens: opening up from f/1.4 to f/1 does little good when
using electronic sensors.

>From a pure lens design perspective, its easy to design an ultrawide
lens having an aperture much faster than f/0.7, but the size can
quickly get well out of the practical range for manufacturing.

Brian
Anonymous
August 6, 2005 4:05:43 PM

Archived from groups: rec.photo.digital.slr-systems,rec.photo.digital (More info?)

>>"Lenses are more than wide enough already. Make 'em faster instead.
>>Give me a good 50mm f/0.7, and we can talk about market possibilities. :)  "

Agree! And 'affordable', too!
I have a 35mm film system with 20mm f/2.8 lens and 24mm f/2 lens. I
wish I had something that large for max aperture to work with my 20D!
A 12.5mm f/2.8 and a 15mm f/2 lens equivalent to what I already own for
35mm film.
August 6, 2005 4:26:10 PM

Archived from groups: rec.photo.digital.slr-systems,rec.photo.digital (More info?)

Jeremy Nixon wrote:
> BC <brianc1959@aol.com> wrote:
>
> > Unfortunately, the sensors used in all consumer cameras (including
> > DSLR's) cannot adequately handle the marginal ray angle from an f/1
> > beam (30 degrees), let alone from f/0.7 (45.6 degrees). By the time
> > you get to f/0.7 you may start getting substantial crosstalk between
> > pixels.
>
> Is this inherent to digital sensors, or just the current implementations?
>

The sensors used in certain scientific and space applications don't
have the limitation. I think the terminology is "back-illuminated" vs.
"front illuminated", and has to do with whether the sensor circuitry is
in front of the active areas (thereby causing shadowing effects) or
behind them. I don't really understand the subject very well, but I
suspect that there may be technical or cost reasons why the sensors
that work well with large ray angles (whether chief ray or marginal
ray) aren't used in consumer products.

Personally, I agree that f/0.7 or faster lenses would be a cool thing
to have, but only if my camera could deal with them. Unfortunately you
would also need an entirely new lens mount to deal with them. For
example, the EOS mount can only deal with an f/1 cone of light, and the
Nikon F mount can only handle f/1.2. The Olympus 4/3 mount might have
better capability, but I haven't studied the specs.

Brian

> --
> Jeremy | jeremy@exit109.com
Anonymous
August 6, 2005 6:20:04 PM

Archived from groups: rec.photo.digital (More info?)

"Brian Baird" <no@no.thank.u> wrote:
> In article <dd15uu$d0s$1@nnrp.gol.com>, davidjl@gol.com says...
>>
>> Doing that is (a) a pain*, and (b) results in a cylindrical projection,
>> which can result in serious distortion**. If you get a dSLR that takes a
>> standard mount, then the Stigma 12-24 will mount on a cheap film body,
>> and
>> you can use that for your extreme wide angle work. I didn't do that
>> because
>> I was afraid I'd spend too much time shooting 12mm shots on film and not
>> do
>> any digital<g>.
>>
>> *: Although it's not bad. I find that if I shoot on a tripod, Panorama
>> Factory will stitch nicely with no intervention, as it did in the image
>> below.
>>
>> *: http://www.pbase.com/davidjl/image/47227687/large
>
> Have you tried Panorama Tools (ptools)? That gives you the option of
> rectilinear projection for your final output.

It's on my list of things to do. But note that the image above is way beyond
what can be reasonably done with rectilinear projection. I should figure out
how the angles work out, but it seems to me that with anything wider than a
normal lens, as soon as you rotate the camera, you are beyond what
rectilinear projection can handle.

David J. Littleboy
Tokyo, Japan
Anonymous
August 6, 2005 6:20:05 PM

Archived from groups: rec.photo.digital (More info?)

In article <dd1hai$g3b$1@nnrp.gol.com>, davidjl@gol.com says...
> It's on my list of things to do. But note that the image above is way beyond
> what can be reasonably done with rectilinear projection. I should figure out
> how the angles work out, but it seems to me that with anything wider than a
> normal lens, as soon as you rotate the camera, you are beyond what
> rectilinear projection can handle.

It's my understanding Panorama tools will stretch/compress the image to
make it fit.

You'll have to trim for the jaggies, but it does seem to work very well
from what I've seen.
--
http://www.pbase.com/bcbaird
Anonymous
August 6, 2005 6:58:35 PM

Archived from groups: rec.photo.digital (More info?)

"Brian Baird" <no@no.thank.u> wrote in message
news:MPG.1d5e1958d97ee7279897a2@news.verizon.net...
> In article <dd1hai$g3b$1@nnrp.gol.com>, davidjl@gol.com says...
>> It's on my list of things to do. But note that the image above is way
>> beyond
>> what can be reasonably done with rectilinear projection. I should figure
>> out
>> how the angles work out, but it seems to me that with anything wider than
>> a
>> normal lens, as soon as you rotate the camera, you are beyond what
>> rectilinear projection can handle.
>
> It's my understanding Panorama tools will stretch/compress the image to
> make it fit.

Yes. That I understand. My point was that rectilinear projection itself is
limited. Beyond a certain point, you need to go to a cylindrical (or
spherical) projection.

David J. Littleboy
Tokyo, Japan
Anonymous
August 6, 2005 6:58:36 PM

Archived from groups: rec.photo.digital (More info?)

In article <dd1jvu$gms$1@nnrp.gol.com>,
David J. Littleboy <davidjl@gol.com> wrote:
>
>Yes. That I understand. My point was that rectilinear projection itself is
>limited. Beyond a certain point, you need to go to a cylindrical (or
>spherical) projection.

Indeed. The limit is 180 degrees FOV. You can't ever reach it with a
rectilinear projection - the projection would be infinitely wide.

There's an easy mental model you can use to prove this to yourself - imagine
you are looking straight down at a pair of railway lines, stretching off
into the distance, and you want a rectilinear projection of 180 degrees,
showing both horizons. The nature of a rectilinear lens would mean that the
spacing between the lines, which we know to be parallel, would have to
remain constant - they would never meet. The factor by which you need to
scale the distance between the rails in a rectilinear image will increase
with distance from the observer.

But as you tend to the horizon, the observed gap between the lines tends to
zero, so your scale factor tends to infinity. Can't ever work.
Anonymous
August 6, 2005 7:05:33 PM

Archived from groups: rec.photo.digital.slr-systems,rec.photo.digital (More info?)

"BC" <brianc1959@aol.com> wrote in message
news:1123352293.580216.253240@g44g2000cwa.googlegroups.com...
> Whether we like it or not, language is one of the most democratic
> things around. "Popular misusage" ineveitably becomes "correct".

In some cases it does, yes. One example is "battery," which originally meant
two or more cells, typically connected in series to produce a higher
voltage. E.g., a long flashlight might use five cells, but that was still
just one battery. Now people use "battery" when they mean "cell." What was
originally misusage in this case has become standard usage, and now even the
manufacturers use "batteries" to mean "cells," as in a package of eight
"batteries."

This I think is the tipping point: When *manufacturers* of the item adopt
the misusage it becomes standard and correct.

As far as I know, neither Nikon nor Canon nor Pentax nor any other major
prime lens manufacturer uses "prime" to mean FFL. I know that Minolta does
not. I have Minolta lens literature going back about three decades and they
have never misused "prime" in this way.

On the contrary, lens manufacturers such as Zeiss and Schneider catalog
lenses which they identify as "variable prime" lenses, i.e. prime lenses of
variable focal length. (Not zooms, because they are not parfocal which
strictly speaking a zoom is supposed to be.) Since they have been in the
business of making photographic lenses for probably 80 or 90 years before
this misusage of "prime" ever occurred, it seems to me they should be
regarded as authoritative where the terminology is concerned.

N.
Anonymous
August 6, 2005 7:44:29 PM

Archived from groups: rec.photo.digital.slr-systems (More info?)

BC wrote:

> If a rectilinear 8-28mm zoom or perhaps rectilinear primes in the
> 6-10mm range were available, would you be seriously interested? Focal
> lengths listed are in full-frame 35mm equivalents, so for a DX format
> the above-mentioned zoom would have an acutal focal length range of
> 5.33 - 18.67mm). Distortion would be nearly zero, as would
> illumination falloff.
>
> Brian
>

a 12 or 13mm rectilinear prime would be nice for the 7D.

Or perhaps if Minolta were to adopt the same 10 Mpix 1.37x cropped
sensor that Leica used, then a 15mm

In both cases, f/2.8 without vignetting / falloff would be required.

Not cheap in any case.

Cheers,
Alan.

--
-- r.p.e.35mm user resource: http://www.aliasimages.com/rpe35mmur.htm
-- r.p.d.slr-systems: http://www.aliasimages.com/rpdslrsysur.htm
-- [SI] gallery & rulz: http://www.pbase.com/shootin
-- e-meil: Remove FreeLunch.
Anonymous
August 6, 2005 9:03:46 PM

Archived from groups: rec.photo.digital (More info?)

In article <dd1jvu$gms$1@nnrp.gol.com>, davidjl@gol.com says...
> > It's my understanding Panorama tools will stretch/compress the image to
> > make it fit.
>
> Yes. That I understand. My point was that rectilinear projection itself is
> limited. Beyond a certain point, you need to go to a cylindrical (or
> spherical) projection.

Very true. I don't know what the FOV limit is, but I'd imagine you'd
max out prior to 180 degrees.
--
http://www.pbase.com/bcbaird
August 6, 2005 9:37:58 PM

Archived from groups: rec.photo.digital.slr-systems,rec.photo.digital (More info?)

Jeremy Nixon wrote:
> BC <brianc1959@aol.com> wrote:
>
> > Personally, I agree that f/0.7 or faster lenses would be a cool thing
> > to have, but only if my camera could deal with them. Unfortunately you
> > would also need an entirely new lens mount to deal with them. For
> > example, the EOS mount can only deal with an f/1 cone of light, and the
> > Nikon F mount can only handle f/1.2.
>
> There have been Nikon F mount lenses faster than f/1.2; Rodenstock has
> made a couple (one was f/0.75), and there was a Nikon 85mm f/1.0 at one
> point. Does the fact that they were all macro/close-up lenses make the
> difference there?
>
> (sorry, I know you were asking about super-wide lenses, but you clearly
> know more about this than most around here and super-fast lenses interest
> me greatly...)
>
> --
> Jeremy | jeremy@exit109.com

f/#'s are almost always marked for infinity focus. The effective f/#
is greater than the infinity f/# when you focus up close. I know that
Bjorn Rorslett has mounted one of the ultra-fast Rodenstock lenses to a
Nikon, but this was only for closeup photography. Ultimately, the
effective f/# is limited by the registration distance together with the
inside diameter of the lens mount.

Brian
Anonymous
August 6, 2005 10:36:34 PM

Archived from groups: rec.photo.digital.slr-systems,rec.photo.digital (More info?)

BC <brianc1959@aol.com> wrote:

> Such lenses are very different from fisheyes, and I'm interested in
> finding out more about what market possibilities might exist.

Lenses are more than wide enough already. Make 'em faster instead.
Give me a good 50mm f/0.7, and we can talk about market possibilities. :) 

--
Jeremy | jeremy@exit109.com
August 6, 2005 10:55:37 PM

Archived from groups: rec.photo.digital.slr-systems,rec.photo.digital (More info?)

David Littlewood wrote:

"Sounds exciting, Brian. Don't let my slightly doubtful response put
you
off; if it wasn't too expensive, I'm enough of a technology geek to be
tempted anyway. The parameters you describe sound very good - is there
anything which is not so good? Like the thought of it needing a
centre-clear filter though.

It would be interesting to see more of how the geometry of 3D objects
would look using such a lens - as that was my main concern. Are there
any web pictures we could look at?


Best of luck with it anyway, "

Hi David:
I would expect a hyper-wide rectilinear lens to have a pretty small
following due to the geometrical effects. That's why a zoom would be
more useful: you could always zoom out of the hyperwide range into a
more normal wide-angle range. Hyper-wide coverage would also be useful
if you wanted to design a really wide-angle shift lens. The best way
to get some trial hyper-wide images is to use stitching. I admit that
when stitching I often get images that are just *too* wide. But
sometimes the ability to go hyper-wide is just magical. The problem
here is that you really can't judge the composition accurately until
stitching is complete.

I suspect my real market will be DMD or LCD projection, aerial recon.,
surveillance, and possibly a handful of oddball photographers. The
problem is, my heart is really into photography, and not the other
applications.

Brian
Anonymous
August 6, 2005 11:10:09 PM

Archived from groups: rec.photo.digital.slr-systems,rec.photo.digital (More info?)

BC <brianc1959@aol.com> wrote:

> Unfortunately, the sensors used in all consumer cameras (including
> DSLR's) cannot adequately handle the marginal ray angle from an f/1
> beam (30 degrees), let alone from f/0.7 (45.6 degrees). By the time
> you get to f/0.7 you may start getting substantial crosstalk between
> pixels.

Is this inherent to digital sensors, or just the current implementations?

--
Jeremy | jeremy@exit109.com
Anonymous
August 6, 2005 11:17:00 PM

Archived from groups: rec.photo.digital.slr-systems,rec.photo.digital (More info?)

In article <1123351286.207516.226250@f14g2000cwb.googlegroups.com>, BC
<brianc1959@aol.com> writes
>
>Hi Chris:
>I've already designed examples similar to the specs I listed, so I
>wasn't just making up stuff that is physically impossible. One of the
>lenses I presented at a recent conference covers 140 degrees at f/1.0
>with less than 0.1% distortion. Nevertheless, the illumination in the
>corners of the image is slightly *higher* than in the center. In this
>case you would have to introduce some vignetting in order to achieve
>uniform illumination. I've designed other rectilinear examples
>covering up to 150 degrees that have almost perfectly uniform
>illumination.
>
>Two things determine illumination falloff. First is the angle of the
>chief ray in image space (not object space), and second is pupil
>aberration. An unvignetted telecentric lens with no pupil aberration
>will have perfectly uniform illumination regardless of distortion or
>field of view. With favorable pupil aberration you can actually have
>an illumination curve that rises as you approach the corners, even when
>distortion is zero. Numerous telecentric projection lenses have been
>built that exhibit this behavior.
>
>Agreed that you can't actually get to a full 180 degrees with a
>rectilinear lens. But you can get alot wider than conventional wisdom
>might lead you to believe. The old rectilinear limit was set by the
>Goerz Hypergon at around 135 - 140 degrees, but this lens is only
>useable at a tiny aperture and it has tons of illumination falloff. I
>suspect the real practical limit is somewhere in the 160 to 170 degree
>range. Such lenses are very different from fisheyes, and I'm
>interested in finding out more about what market possibilities might
>exist.
>
>I'm aware of the Panotools option, since I've used it extensively for
>about 4 years now. I've always felt the image degradation due to
>interpolation which you mention is too severe unless working with
>ultra-high resolution stitched images. I feel the same way about
>correcting keystone distortion.
>
>Also, if you shoot a flat target with an ultrawide angle rectilinear
>lens, then circles in the object will imaged as circles in the image
>plane. Its only when you consider three-dimensional objects that
>things start to get strange.
>
>Brian
>
Sounds exciting, Brian. Don't let my slightly doubtful response put you
off; if it wasn't too expensive, I'm enough of a technology geek to be
tempted anyway. The parameters you describe sound very good - is there
anything which is not so good? Like the thought of it needing a
centre-clear filter though.

It would be interesting to see more of how the geometry of 3D objects
would look using such a lens - as that was my main concern. Are there
any web pictures we could look at?

Best of luck with it anyway,

David
--
David Littlewood
August 6, 2005 11:17:13 PM

Archived from groups: rec.photo.digital.slr-systems,rec.photo.digital (More info?)

Nostrobino wrote:

"This I think is the tipping point: When *manufacturers* of the item
adopt
the misusage it becomes standard and correct.

As far as I know, neither Nikon nor Canon nor Pentax nor any other
major
prime lens manufacturer uses "prime" to mean FFL. I know that Minolta
does
not. I have Minolta lens literature going back about three decades and
they
have never misused "prime" in this way.

On the contrary, lens manufacturers such as Zeiss and Schneider catalog

lenses which they identify as "variable prime" lenses, i.e. prime
lenses of
variable focal length. (Not zooms, because they are not parfocal which
strictly speaking a zoom is supposed to be.) Since they have been in
the
business of making photographic lenses for probably 80 or 90 years
before
this misusage of "prime" ever occurred, it seems to me they should be
regarded as authoritative where the terminology is concerned. "

One good example of a manufacturer that uses "prime" and "zoom" in the
way most people do is Panavision. This page on their website clearly
identifies "prime" lenses as "fixed focal length" lenses:
http://www.panavision.com/lenses.php

However, I would reject the idea that manufacturers should define the
language. After all, its companies like Schneider and Rodenstock that
have completely trashed the meaning of perfectly good optical terms
like "apochromatic" in the interest of hawking their wares. Sounds
more like fraud than authoritativeness to me.

Brian
Anonymous
August 6, 2005 11:47:47 PM

Archived from groups: rec.photo.digital.slr-systems,rec.photo.digital (More info?)

BC <brianc1959@aol.com> wrote:

> Personally, I agree that f/0.7 or faster lenses would be a cool thing
> to have, but only if my camera could deal with them. Unfortunately you
> would also need an entirely new lens mount to deal with them. For
> example, the EOS mount can only deal with an f/1 cone of light, and the
> Nikon F mount can only handle f/1.2.

There have been Nikon F mount lenses faster than f/1.2; Rodenstock has
made a couple (one was f/0.75), and there was a Nikon 85mm f/1.0 at one
point. Does the fact that they were all macro/close-up lenses make the
difference there?

(sorry, I know you were asking about super-wide lenses, but you clearly
know more about this than most around here and super-fast lenses interest
me greatly...)

--
Jeremy | jeremy@exit109.com
Anonymous
August 6, 2005 11:59:10 PM

Archived from groups: rec.photo.digital (More info?)

"David J Taylor" wrote:
> David J. Littleboy wrote:
>> "Nicholas Wittebol" <nicholas.wittebol@rogers.com> wrote:
>>> With, DSLR, why not just take multiple shots and stitch them
>>> together? Panorama software is much cheaper than an expensive lense.
>>> Is there something I am missing? (I am a newbie to all of this cool
>>> stuff!)
>>
>> Doing that is (a) a pain*, and (b) results in a cylindrical
>> projection, which can result in serious distortion**.
>
> Only if you ask for cylindrical projection, surely? I've seen programs
> which allow spherical and rectilinear projection as well.

Yes. I was talking about what is, for me, the most common case: stitching
together a few wide angle images in a single row to create a generic
panorama. Rectilinear will be common for people using stitching to get
higher resolution, spherical for folks doing VR stuff.

Another thing is that the distortion is only really a problem in "panoramas"
with a wide angle yet a more conservative (less "panoramic") aspect ratio.

Another problem with panoramas: the local pro shop will only print panoramas
as crops on a larger print, and you get charged for the whole larger print.
Sigh. Time to figure out how to use roll paper.

David J. Littleboy
Tokyo, Japan
Anonymous
August 6, 2005 11:59:11 PM

Archived from groups: rec.photo.digital (More info?)

David J. Littleboy wrote:
[]
> Yes. I was talking about what is, for me, the most common case:
> stitching together a few wide angle images in a single row to create
> a generic panorama. Rectilinear will be common for people using
> stitching to get higher resolution, spherical for folks doing VR
> stuff.
> Another thing is that the distortion is only really a problem in
> "panoramas" with a wide angle yet a more conservative (less
> "panoramic") aspect ratio.

Agreed on both, David.

> Another problem with panoramas: the local pro shop will only print
> panoramas as crops on a larger print, and you get charged for the
> whole larger print. Sigh. Time to figure out how to use roll paper.

I think one of the printing shops in Edinburgh (leaflets and posters etc.)
has a roll-printing facility (they have panoramas on display), so it might
be worth checking round. Paying for the whole print seems unfair.

Cheers,
David
Anonymous
August 7, 2005 2:02:18 AM

Archived from groups: rec.photo.digital.slr-systems,rec.photo.digital (More info?)

In article <1123271049.621406.294920@g14g2000cwa.googlegroups.com>,
BC <brianc1959@aol.com> wrote:
>If a rectilinear 8-28mm zoom or perhaps rectilinear primes in the
>6-10mm range were available, would you be seriously interested?

Define "seriously". :-)

I would *like* to have such a lens to play with and see what
effects it has on various possible images. But at the same time, I
don't see enough of a use for *my* photography to go much above about
$300.00 -- which would be dirt cheap for such a lens. This is an
important factor because I am retired, and can't spend as much as I
would like to. (There are lenses which I want to get right now which I
can't afford.)

> Focal
>lengths listed are in full-frame 35mm equivalents, so for a DX format
>the above-mentioned zoom would have an acutal focal length range of
>5.33 - 18.67mm).

Hmm ... would it cover a full 35mm frame? Or just a crop factor
of 1.5 or so? For my current DSLRs, it would not make any difference,
but I do have old film cameras (Nikon F's, and with the right backs,
the Nikon N90s.)

> Distortion would be nearly zero, as would
>illumination falloff.

It sounds nice. What would the physical size and weight be
like? Would the front element have to bulge out beyond the mounting
hardware, thus being more exposed to possible damage? Would a filter be
possible in front of it for protection or light modification, or would
filters (for light modification) have to be placed between elements or
behind the rear-most element and live with no protective cover for the
front element?

Enjoy,
DoN.
--
Email: <dnichols@d-and-d.com> | Voice (all times): (703) 938-4564
(too) near Washington D.C. | http://www.d-and-d.com/dnichols/DoN.html
--- Black Holes are where God is dividing by zero ---
Anonymous
August 7, 2005 3:11:55 AM

Archived from groups: rec.photo.digital.slr-systems,rec.photo.digital (More info?)

In article <1123354636.809674.212570@g14g2000cwa.googlegroups.com>, BC
says...

> Unfortunately, the sensors used in all consumer cameras (including
> DSLR's) cannot adequately handle the marginal ray angle from an f/1
> beam (30 degrees), let alone from f/0.7 (45.6 degrees). By the time
> you get to f/0.7 you may start getting substantial crosstalk between
> pixels.

Does a lens design such as the 4/3 system, which is supposed to give you
more perpendicular rays, help here ?
--

Alfred Molon
------------------------------
Olympus 4040, 5050, 5060, 7070, 8080, E300 forum at
http://groups.yahoo.com/group/MyOlympus/
Olympus E300 resource - http://myolympus.org/E300/
Anonymous
August 7, 2005 3:14:06 AM

Archived from groups: rec.photo.digital.slr-systems,rec.photo.digital (More info?)

In article <11fa0pib8do000a@corp.supernews.com>, Jeremy Nixon says...

> Lenses are more than wide enough already. Make 'em faster instead.
> Give me a good 50mm f/0.7, and we can talk about market possibilities. :) 

Well, such a lens would be fine for portraits, but the low DOF (when
used with an (D)SLR) would make it unsuitable for other applications,
such as landscape photography for instance. You'd have to stop down the
lens.
--

Alfred Molon
------------------------------
Olympus 4040, 5050, 5060, 7070, 8080, E300 forum at
http://groups.yahoo.com/group/MyOlympus/
Olympus E300 resource - http://myolympus.org/E300/
Anonymous
August 7, 2005 3:40:31 AM

Archived from groups: rec.photo.digital.slr-systems,rec.photo.digital (More info?)

On Sat, 06 Aug 2005 19:47:47 -0000, Jeremy Nixon <jeremy@exit109.com>
wrote:

>BC <brianc1959@aol.com> wrote:
>
>> Personally, I agree that f/0.7 or faster lenses would be a cool thing
>> to have, but only if my camera could deal with them. Unfortunately you
>> would also need an entirely new lens mount to deal with them. For
>> example, the EOS mount can only deal with an f/1 cone of light, and the
>> Nikon F mount can only handle f/1.2.
>
>There have been Nikon F mount lenses faster than f/1.2; Rodenstock has
>made a couple (one was f/0.75), and there was a Nikon 85mm f/1.0 at one
>point. Does the fact that they were all macro/close-up lenses make the
>difference there?

Pretty brave of Rodenstock to make such a fast lens when they couldn't
even make decent enlarger lenses.
-Rich
August 7, 2005 4:53:12 AM

Archived from groups: rec.photo.digital.slr-systems,rec.photo.digital (More info?)

On Sat, 6 Aug 2005 09:08:26 -0500, "AustinMN" <Austin260@comcast.net>
wrote:

>Charles wrote:
>> On 5 Aug 2005 12:52:32 -0700, "eawckyegcy@yahoo.com"
>> <eawckyegcy@yahoo.com> wrote:
>>
>>>BC wrote:
>>>
>>>> If a rectilinear 8-28mm zoom or perhaps rectilinear primes in the
>>>> 6-10mm range were available, would you be seriously interested? Focal
>>>> lengths listed are in full-frame 35mm equivalents, so for a DX format
>>>> the above-mentioned zoom would have an acutal focal length range of
>>>> 5.33 - 18.67mm). Distortion would be nearly zero, as would
>>>> illumination falloff.
>>>
>>>I would only consider this optic if, and only if, in addition to no
>>>illumination falloff, no distortion, it also reduced diffraction
>>>effects to zero and possessed absolutely no aberrations of any kind.
>>>Additionaly, the focus motors would also have to be perpetual motion
>>>machines -- let them be powered by the vacuum energy.
>>
>>
>> as long as it didn't cost too much.
>
>It should also have a lifetime warranty and a universal mount that adapts to
>any camera mount with full functionality and no degradation in performance.


and auto composition correction. It's somewhat like image
stabilization, it shifts internal elements to eliminate distracting
things from the picture.
Anonymous
August 7, 2005 4:53:13 AM

Archived from groups: rec.photo.digital.slr-systems,rec.photo.digital (More info?)

Charles wrote:
> On Sat, 6 Aug 2005 09:08:26 -0500, "AustinMN"
> <Austin260@comcast.net>
> wrote:
>
>> Charles wrote:
>>> On 5 Aug 2005 12:52:32 -0700, "eawckyegcy@yahoo.com"
>>> <eawckyegcy@yahoo.com> wrote:
>>>
>>>> BC wrote:
>>>>
>>>>> If a rectilinear 8-28mm zoom or perhaps rectilinear primes in
>>>>> the
>>>>> 6-10mm range were available, would you be seriously interested?
>>>>> Focal lengths listed are in full-frame 35mm equivalents, so for
>>>>> a
>>>>> DX format the above-mentioned zoom would have an acutal focal
>>>>> length range of
>>>>> 5.33 - 18.67mm). Distortion would be nearly zero, as would
>>>>> illumination falloff.
>>>>
>>>> I would only consider this optic if, and only if, in addition to
>>>> no
>>>> illumination falloff, no distortion, it also reduced diffraction
>>>> effects to zero and possessed absolutely no aberrations of any
>>>> kind. Additionaly, the focus motors would also have to be
>>>> perpetual motion machines -- let them be powered by the vacuum
>>>> energy.
>>>
>>>
>>> as long as it didn't cost too much.
>>
>> It should also have a lifetime warranty and a universal mount that
>> adapts to any camera mount with full functionality and no
>> degradation in performance.
>
>
> and auto composition correction. It's somewhat like image
> stabilization, it shifts internal elements to eliminate distracting
> things from the picture.

I think that's an extra-cost plug-in.

--
Frank ess
Anonymous
August 7, 2005 6:23:23 AM

Archived from groups: rec.photo.digital.slr-systems,rec.photo.digital (More info?)

Martin Brown wrote:
> Nicholas Wittebol wrote:
>
> > With, DSLR, why not just take multiple shots and stitch them together?
> > Panorama software is much cheaper than an expensive lense. Is there
> > something I am missing? (I am a newbie to all of this cool stuff!)
>
> It is difficult to do a whole sky camera for meteor watch that way. You
> want the wide angle view taken at the same time. But for landscapes
> where nothing much is changing then a good tripod and stitching software
> can do a pretty good job much cheaper too!
>
> There is a fixed focal length 8mm Russian Pelang (sp?) fisheye lens for
> 35mm SLRs offered from time to time in photomags that looks interesting
> for this (and cheaper than Pentax/Canon/Nikon etc). I have been told
> that it isn't too bad but not seen any results from one myself.

Try "peleng" on ebay. Right now, it will cost you around 210 bucks US
delivered.
Anonymous
August 7, 2005 11:02:25 AM

Archived from groups: rec.photo.digital.slr-systems,rec.photo.digital (More info?)

BC <brianc1959@aol.com> wrote:

> f/#'s are almost always marked for infinity focus. The effective f/#
> is greater than the infinity f/# when you focus up close. I know that
> Bjorn Rorslett has mounted one of the ultra-fast Rodenstock lenses to a
> Nikon, but this was only for closeup photography.

Yes; I've seen one of the f/1.1 Rodenstocks with a Nikon mount (not the
f/0.75 one) and it did not have infinity focus. I figured that was just
a cost decision, given that the poor quality of the lens would make it
one you'd not want to use for general photography anyhow.

> Ultimately, the effective f/# is limited by the registration distance
> together with the inside diameter of the lens mount.

Ah, well, I guess I'll just have to live with f/1.2 and a dream. :) 

--
Jeremy | jeremy@exit109.com
August 7, 2005 12:26:54 PM

Archived from groups: rec.photo.digital.slr-systems,rec.photo.digital (More info?)

Chris Brown wrote:

"The only way you can make the corners equally bright is to discard
some of
the light hitting the centre."

This sentence seems to sum up most of your position, so I've snipped
out the rest. The short answer is, fortunately, no. You really can
avoid corner falloff without throwing away on-axis light. When I set
out to design one of these lenses at, say, f/2.8, it stays f/2.8
regardless of how good the off-axis illumination is. In other words,
if you start restricting the field of view by introducing smaller and
smaller field stops the on-axis image brightness remains constant.

If you look at the incoming ray bundles at the front of some of these
lenses you will see that the off-axis bundles are huge compared to the
on-axis bundle. Its true that the front end of the lens must cope with
these large off-axis ray bundles, but in no way does this constitute
"discarding some of the light hitting the centre". Physics has not
been violated, and you can still use a normal iris diaphragm. If you'd
like, I can email you a paper that I recently presented that addresses
on this subject.

Bear in mind that I do have actual rectilinear designs covering 140
degrees with with *negative* falloff (edges brighter than the center),
and up to 150 degrees with virtually no falloff.. Admittedly, I
haven't built prototypes - but you really can use software to
accurately calculate things like illumination falloff.

I think its worth re-stating that in the cos^4(theta) rule the angle
theta is not the field angle (i.e. the angle of the chief ray in object
space), but is rather the angle of the chief ray in image space. The
chief ray is the one that passes through the center of the aperture
stop and the extreme corner of the image. For pinholes and certain
simple types of lenses the two angles are identical, or nearly so.
However, there are large classes of lenses in which the chief ray angle
in image space is much less than in object space. The best example is
a telecentric lens, where the chief ray angle in image space is zero
regardless of field angle. As I pointed out in an earlier post, a
telecentric lens with no vignetting or pupil aberration will have
exactly zero illumination falloff regardless of field angle. And it is
not necessary to discard light from the center to achieve this.

Brian
!