Noblex 150 resolution at infinity
RolandRB
Noblex 150 resolution at infinity. It is here
http://www.tawbaware.com/maxlyons/calc.htm
The way I am using it is to alter the circle of confusion size until
infinity is just lost out of the depth of field. I then take the
inverse of the circle of confusion diameter to give the resolution in
line pairs per unit at infinity.
Assuming the Noblex has a 50.75mm lens and it is fixed focussed to
10.4 meters then the following is what I got for the resolution at
infinity in line pairs per millimeter.
f4.5 CoC=0.0553mm (18 lp/mm)
f5.6 CoC=0.0444mm (22.5 lp/mm)
f8 CoC=0.0311mm (32 lp/mm)
f11 CoC=0.0227mm (44 lp/mm)
f16 CoC=0.0155mm (64.5 lp/mm)
My own conclusions are that if I were to use it for distant scenes, in
preference to my Horizon 202, then I would opt to use it at f11 or
smaller. At f16 it should theoretically give spectacular results.
jjs
"RolandRB" <rolan...@hotmail.com> wrote in message
news:52970945.04042...@posting.google.com...
> I found a useful online calculator so that I could calculate the
> Noblex 150 resolution at infinity. It is here
> http://www.tawbaware.com/maxlyons/calc.htm
Wow! With that link you don't have to worry about manufacturing variations,
lens quality, and all that stuff that requires getting off your ass to truly
determine.
What does a theoretical picture look like?
jjs
> Noblex 150 resolution at infinity. It is here
> f4.5 CoC=0.0553mm (18 lp/mm)
> f5.6 CoC=0.0444mm (22.5 lp/mm)
> f8 CoC=0.0311mm (32 lp/mm)
> f11 CoC=0.0227mm (44 lp/mm)
> f16 CoC=0.0155mm (64.5 lp/mm)
That's so strange, Roland. Why are you changing the CoC with each aperture?
You should get sharper results wide open. I mean, this IS theoretical,
right?
David J. Littleboy
"jjs" <nos...@please.xxx> wrote in message
news:108t02e...@news.supernews.com...
He's not changing the CoC, he's changing the aperture and _back calculating_
the CoC for a point at infinity from the aperture and the focus setting.
(That is, he's calculating the smallest CoC that will give a hyperfocal
distance of 10.5 feet for each f stop.)
That tells you how sharp the image at infinity will be for each f stop.
David J. Littleboy
Tokyo, Japan
Hemi4268
Yes, your calculations seem correct. A perfect in focus lens with give about
2000 l/mm center resolution at f-1 in noon summer sun. At f-2 1000l/mm and at
f-4 500 l/mm and so on.
So the table will look like this
f-1 =2000 l/mm
f-2 =1000
f-4 =500
f-8 = 250
f-16 =125
f-32 = 64
So your 64 l/mm at f-16 is a good figure since the image still is not in
perfect focus at infinity. Although, the true system resolution on the film
after exposure will be about half that or 32 l/mm.
Standard calculation will be lens reso times film reso over lens reso plus film
reso
(64x64)
(64+64) = 32
Larry
RolandRB
I can see you are getting interested in this side of photography. :o)
Stacey
>
> f4.5 CoC=0.0553mm (18 lp/mm)
> f5.6 CoC=0.0444mm (22.5 lp/mm)
> f8 CoC=0.0311mm (32 lp/mm)
> f11 CoC=0.0227mm (44 lp/mm)
> f16 CoC=0.0155mm (64.5 lp/mm)
>
> My own conclusions are that if I were to use it for distant scenes, in
> preference to my Horizon 202, then I would opt to use it at f11 or
> smaller. At f16 it should theoretically give spectacular results.
Of a theoretical subject using the theoretical lens on this theoretical
camera you say you'e owned for years but never have used... Me, I'd go
outside and take some pictures with it and see what they look like but I
guess that's being silly!
--
Stacey
Ralf R. Radermacher
> Me, I'd go outside and take some pictures with it and see what
> they look like but I guess that's being silly!
He owns it only theoretically.
Ralf
--
Ralf R. Radermacher - DL9KCG - Köln/Cologne, Germany
private homepage: http://www.fotoralf.de
manual cameras and photo galleries - updated March 30, 2004
Contarex - Kiev 60 - Horizon 202 - P6 mount lenses
RolandRB
The lens is (assumed to be) fixed focussed to 10.4 meters and will
have a depth of field that depends on what the acceptable circle of
confusion diameter is. So if I play around with the value of the
circle of confusion diameter then I can set it to a value where
infinity is only just squeezed out of the depth of field. Once I have
this value with infinity just out but very nearly in then the inverse
of it will give me the highest possible resolution in line pairs per
millimeter at infinity. I am working the formula backwards to find out
how sharp the image could be for infinity scenes.
What it means is no matter how sharp the lens is at 10.4 meters then
the highest achievable resolution at infinity for f4.5 is 18 lp/mm and
at f5.6 the highest resolution achievable is 22.5 lp/mm and so on as
listed.
RolandRB
I am always happy to read any of your posts because of your past
experience with lenses. Thanks for the list of resolutions for a
perfect lens. No lens is perfect, of course. I remember you once
posted a similar list for typical brand name lenses. Could you post
that here again please and let us know where the figures came from. I
think people easily confuse lens resolution with practical on-film
resolution. Since film resolution can vary then I think it is best to
keep lens resolution and film resolution as seperate topics and get
the effective on-film value by combining them using the formula you
quoted above.
Once I have the list of resolutions for typical brand-name lenses I
will do a comparison between the Horizon 202 fix-focussed at infinity
compared with a Noblex 150 fix-focussed at 10.4 meters for
photographing distant (effectively at infinity) scenes.
Hemi4268
>posted a similar list for typical brand name lenses.
Yes, let me post both tables so you can see the difference.
Again noon summer sun center resolution.
perfect lens typical name brand lens
f-1 2000l/mm 100
f-2 1000 300
f-4 500 500
f-8 250 250
f-16 125 125
You can see at about f-4 the perfect lens and the name brand lens kinda match.
Remember, we are still talking cente rand NOT edge resolution.
Larry
Hemi4268
Should read:
we are talking center resolution and NOT edge resolution.
This is not the first time my computer drops words or parts of words.
Larry
RolandRB
I have seen it mentioned that f8 is the "sweet spot" for most
practical lenses. That seems to contradict. Do you have an explanation
for this?
Hemi4268
>practical lenses. That seems to contradict. Do you have an explanation
>for this?
Actually f-8 is the sweet spot if you account for the depth of focus.
Lets try that list again using depth of focus for an additional output.
Again our chart for a perfect lens using noon summer sun.
F-stop Resolution FocusDepth
f-1 2000l/mm 1 micron
f-2 1000l/mm 4 microns
f-4 500 16
f-8 250 64
f-16 125 256
You can see the depth of focus is much larger at f-8 then at f-4. You can also
see that the fall off of resolution is really not all that great from f-4 to
f-8. If we go through the lens/film calculation using T-Max 100, a sharp
100+line image is very possible with the lens set at f-8.
Also the chances of getting the focus correct is 4 times better at f-8 then it
is at f-4.
Larry
David J. Littleboy
It's because that's more head-in-the-sand theory that blithely ignores
practical reality. The twit's talking about theoretically perfect lenses
imaging monochromatic blue light. Oh, yes. And those are aerial resolution
figures at 0.01% MTF.
Completely irrelevant to real photography.
Stacey
Because that is what people have found -actually using the lenses- instead
of calculating using a limited set of data points what should work best.
There are -HUNDREDS- of variables you are leaving out of your equations.
--
Stacey
RolandRB
You can use theory to give you an upper limit on what is achievable
and this is useful to know. For example, the circles of confusion for
a Noblex fix-focussed at 10.4 meters but used to photograph a scene
effectively at infinity, limit the resolution you can achieve and this
can be calculated. Also the loss of horizontal resolution when the
secondary principal point is moved off axis for a focussing model can
be calculated. There may be hundreds of variables to consider but none
of these will give you a BETTER result. Only a worse one.
RolandRB
There is a great deal of misconception about the resolving power of
lenses due to people doing tests on-film and publishing the results
because it is the film itself that is the major limiting factor for
resolution. You might be surprised at what is achievable for even a
simple lens. I am talking about the human eye. For example, how far do
two light sources have to be apart two miles away for you to see that
there are two light sources and not one? This is a practical example.
Make a best practical estimate for that and then consider that the
human eye focal length is 17mm. You will then be able to work out how
close those separate images are on the retina of the eye and so work
out the resolving power in lines per millimeter of the human eye. You
will find that the resolving power of the human eye is very high. I
invite you to do a calculation and post the results here.
Lassi Hippeläinen
>
> >I have seen it mentioned that f8 is the "sweet spot" for most
> >practical lenses. That seems to contradict. Do you have an explanation
> >for this?
>
> Actually f-8 is the sweet spot if you account for the depth of focus.
No, f:8 is the sweep spot if you include also the corners. Those numbers
were for the center of the image.
-- Lassi
Stacey
>
> You can use theory to give you an upper limit on what is achievable
> and this is useful to know. For example, the circles of confusion for
> a Noblex fix-focussed at 10.4 meters but used to photograph a scene
> effectively at infinity, limit the resolution you can achieve and this
> can be calculated. Also the loss of horizontal resolution when the
> secondary principal point is moved off axis for a focussing model can
> be calculated. There may be hundreds of variables to consider but none
> of these will give you a BETTER result. Only a worse one.
So? Doesn't mean the camera is useable at infinity or really anything in
actual use other than what it might be able to do in a perfect world, which
we don't live in!
--
Stacey
David J. Littleboy
Yep. And most of us still shoot on film. And those of us who don't, use
sensors with lower limiting resolution than film (although those sensors
have lower noise and higher MTFs at the frequencies one needs than film).
Hemi4268
>
Sure, as gas is irrelevant to driving.
Larry
Hemi4268
>will find that the resolving power of the human eye is very high.
Actually the resolving power of the eye is well documented. It's 8 lines per
millmeter.
Larry
Hemi4268
Actually your kinda right. Years ago I working on an "Image Quality" equation
for the government.
It was a typical government program with 1 PHd's in math as head of the project
with 5 workers and two years of work. After $1 million dollars spent we got
something.
Larry
Hemi4268
>The twit's talking about theoretically perfect lenses
>imaging monochromatic blue light. Oh, yes. And those are aerial resolution
>figures at 0.01% MTF.
>
>Completely irrelevant to real photography.
>
>David J. Littleboy
>Tokyo, Japan
>
could talk about heat content, octane, chemical mix. This all to get the most
power and distance from a tank of gas.
Yet to drive on the freeway all you need to do is gas up and go.
So sure, just like driving on a freeway all we need is to load a roll of film
and shot. Although, it's sometime nice to know how the system works and how
fast (sharp) we can make it.
Larry
David J. Littleboy
"Hemi4268" <hemi...@aol.com> wrote:
> >Completely irrelevant to real photography.
>
> Sure, as gas is irrelevant to driving.
The aerial resolution of lenses is about as relevant to real photography as
the energy available from running the hydrogen in gasoine through a fusion
reactor is to real driving.
RolandRB
> I found a useful online calculator so that I could calculate the
> Noblex 150 resolution at infinity. It is here
>
> The way I am using it is to alter the circle of confusion size until
> infinity is just lost out of the depth of field. I then take the
> inverse of the circle of confusion diameter to give the resolution in
> line pairs per unit at infinity.
>
> Assuming the Noblex has a 50.75mm lens and it is fixed focussed to
> 10.4 meters then the following is what I got for the resolution at
> infinity in line pairs per millimeter.
>
> f4.5 CoC=0.0553mm (18 lp/mm)
> f5.6 CoC=0.0444mm (22.5 lp/mm)
> f8 CoC=0.0311mm (32 lp/mm)
> f11 CoC=0.0227mm (44 lp/mm)
> f16 CoC=0.0155mm (64.5 lp/mm)
>
> preference to my Horizon 202, then I would opt to use it at f11 or
> smaller. At f16 it should theoretically give spectacular results.
And now to do the calculations comparing the Horizon 202 (which is
fixed-focussed to infinity) to the Noblex 150 (asumed fixed-focussed
to 10.4m) when both photographing a scene at effectively infinity. The
Noblex 150 has film area 50mm x 120mm and the Horizon 24mm x 58mm so
if the Noblex can put better than half the lines per millimeter on
film than the Horizon then it will win. The lens resolution on the
Noblex at infinity is limited by the circles of confusion as shown in
my post above. I'll assume a good film is being used which is capable
of 100 lines per millimeter and that the lens resolution of the
Horizon gives 110 lp/mm over most of the area of interest (not top and
bottom).
So for the Horizon the on-film resolution would be
110*100/(110+100) = 52 lp/mm
For the Noblex at f8
32*100/(32+100) = 24 lp/mm
so the Horizon just beats it theoretically at this f stop but the
Horizon would have to be tripod-mounted to have a chance of achieving
this in practise and the flaws on film would show up more, so
effectively the Noblex is at least equal.
For the Noblex at f11
44*100/(44+100) = 31 lp/mm
so the Noblex wins by a small margin
For the Noblex at f16
64.5*100/(64.5+100) = 39 lp/mm
so the Noblex wins by a comfortable margin
Hemi4268
You will find that with Tmax 400 film, your calculations and actual test
results with good reso targets will match almost exactly.
Larry
Ralf R. Radermacher
> You will find that with Tmax 400 film, your calculations and actual test
> results with good reso targets will match almost exactly.
Which says a lot about the practical relevance of this whole exercise.
Shooting a resolution target mounted at infinity. The mind boggles....
Ralf
--
Ralf R. Radermacher - DL9KCG - Köln/Cologne, Germany
private homepage: http://www.fotoralf.de
manual cameras and photo galleries - updated April 29, 2004
Hemi4268
>Shooting a resolution target mounted at infinity. The mind boggles....
If the lens is under 2 inches or so, infinity is really not all that far away.
About 100 yards or the distance of a football field will do it. Peopel walk
100 yard all the time. Some people get paid millions to run that far every
Monday evening in the Fall.
Larry
-
jjs
"Hemi4268" <hemi...@aol.com> wrote in message
news:20040430164029...@mb-m05.aol.com...
> >Which says a lot about the practical relevance of this whole exercise.
> >Shooting a resolution target mounted at infinity. The mind boggles....
>
> If the lens is under 2 inches or so, infinity is really not all that far
away.
What's the rule of thumb for determining infinity? Something like 400 *
focal length?
Hemi4268
You get close to infinity when the image is 1000 times smaller then the
original. This is about 200 ft with a 2 inch lens. Actual focus position will
be 50.050mm with a 50mm lens or 50 microns in from true infinity.
At f-8, which has 64 microns of focus depth, easly overlaps the above 50 micron
focus error and will actually overpass infinity.
Then again at 2000 ft, the actual focus position of a 50 mm lens is 50.005mm or
5 microns in from infinity. This is good enough for a f-2 lens which as 4
microns of focus depth.
Larry
Ralf R. Radermacher
> If the lens is under 2 inches or so, infinity is really not all that far away.
> About 100 yards or the distance of a football field will do it.
So, given this shooting distance, the focal length and the neg size of
the cameras in question, exactly how large a resolution target would you
need?
Bob Monaghan
Huh? If you have lens aerial resolutions, and the manufacturers' film
resolution values, you can predict on-film resolution rather closely using
the system resolution equation (see http://medfmt.8k.com/mf/lenslpm.html).
Film based testing then becomes an exercise in assessing how consistent
and close your overall technique is of reaching optimal results ;-) MTF
testing is another example of film-less (aerial) lens testing much
accepted by lens designers and the industry etc.
You can argue that other factors like shutter vibration and focusing
errors can negatively impact on-film results. But aerial lens testing will
show you the optimum potential of the lens - and how most of it is lost in
limiting film resolution ;-)
The flip side is that you can run system resolution estimates, using some
rather discouragingly modest film resolution limits (e.g., most color
print films), and discover why high aerial resolution lenses don't perform
noticeably better than lower resolution (and cheaper) lenses - the color
print film limiting resolution is by far the limiting factor. Again, a
valuable insight for 94% of consumers who shoot only color print films and
no slides or B&W ;-)
grins bobm
--
***********************************************************************
* Robert Monaghan POB 752182 Southern Methodist Univ. Dallas Tx 75275 *
********************Standard Disclaimers Apply*************************
David J. Littleboy
"Bob Monaghan" <rmon...@engr.smu.edu> wrote in message
news:c6usqt$dr$1...@blaze.seas.smu.edu...
>
> Huh? If you have lens aerial resolutions, and the manufacturers' film
> resolution values, you can predict on-film resolution rather closely using
> the system resolution equation (see http://medfmt.8k.com/mf/lenslpm.html).
Yes, but you can't predict how good the images will look, since you don't
have the MTF in the critical 20 to 30 lp/mm range. What you really want are
the resolutions at 50% and 30% MTF, since it's at those resolutions that
barely useful and barely recognizable imaging happens.
(Question: if you plug the 50% MTF lens resolution and the 50% MTF film
resolution into that equation, do you get the 50% system resolution or the
25% system resolution? Presumably the latter, right???)
Bob Monaghan
Why did you assume the horizon 202 lens _aerial resolution_ is only 110
lpmm? This is the performance level of a terrible fixed lens (in 35mm
format), much less than my and others experience with the Horizon 202.
Users of both Horizon and Noblex 135 (35mm) and 150 (MF) models have rated
it as only "slightly inferior to the Noblex regarding lens performance"
see http://www.pauck.de/marco/photo/panorama/horizon202/horizon202.html
Because of the swing lens design, only the center stripe of the lens image
circle is used to project an image, so edge or corner dropoffs are not an
issue either, right? Again, the center of the lens performance on the
short (vertical) 35mm film axis is all that counts, and is likely to be
very good (only 24mm high). The distortions seen in swing lens images are
generally the result of the swing lens design, not lens distortions, yes?
The larger negative size of the noblex 150 would make it the clear winner
in a limiting enlargement size contest against the 35mm film based horizon
202. The case is much less clear for the noblex 135 series of similar 35mm
format type - see http://www.a1.nl/phomepag/markerink/pan_35mm.htm
The other point, as another posted noted IIRC, is that there are various
models and variants of the Noblex 150 series, some of which are focused
for infinity (as is the horizon 202 series), others for closer distances.
I and others have shown how easy it is to use a fractional + diopter lens
to shift the horizon 202 focus to closer distances, so this is even less
of an issue with the horizon 202 (see http://medfmt.8k.com/mf/horizon.html
In short, the lens quality of the horizon 202 is very close or only
"slightly inferior" to that of the Noblex models costing 500% or so more $
(in same 35mm film format) or even more (in MF 120 formats).
I think it is important to challenge the case that the horizon 202 optics
are somehow seriously less capable or poor resolution (as implied by the
110 lpmm aerial lens resolution value cited in this thread and related
conclusions) because it might dissuade some buyers from considering these
Horizon 202 cameras on their merits - which are many, including low cost.
If anything, the horizon 202 users usually rate these cameras very highly
for their lens resolution and sharpness (as do the Noblex owners, with
similar good reasons ;-)
my $.02 ;-)
Bob Monaghan
the film resolution and lens aerial resolution values will only predict
maximum system resolution values, not contrast performance or MTF values.
I don't know of any way to predict lens contrast performance, other than
inputting the original parameters into a lens design software package
obviously and hoping the lens in your hand is a good match to the math?
Similarly, film contrast performance is so dependent on development
and proper exposure that we measure something easier like resolution ;p-)
Contrast is also more subject to issues like stray light and lens hood
efficiency, camera body flare (as with sundry K-mount camera bodies), and
even a thumbprint can reduce contrast by 25% (per tests in Modern Photo).
A rather pricey ISO standard optical target with various contrast bars is
now used to measure MTF for digital cameras, using a simple mouse maneuver
to highlight the target areas for analysis. If you can put your lens on a
digital camera body, then you can get an MTF chart for your system with
this setup (up to the sensor frequency limit, obviously ;-).
Most of us can't, so we will have to wait for someone to generate some
standard postscript test charts, scan in (B&W) film, and use similar
software to derive an MTF series. Given the software is already available
for downloading free (see links http://medfmt.8k.com/mf/mtftester.html at
top), it is a bit surprising that no one has adapted this approach for
film based MTF testing. Maybe we are all waiting for Godot to do it? ;-)
RolandRB
> Why did you assume the horizon 202 lens _aerial resolution_ is only 110
> lpmm? This is the performance level of a terrible fixed lens (in 35mm
> format), much less than my and others experience with the Horizon 202.
It was an assumption. Anybody can put in whatever values they like
into the calculations.
> Users of both Horizon and Noblex 135 (35mm) and 150 (MF) models have rated
> it as only "slightly inferior to the Noblex regarding lens performance"
> see http://www.pauck.de/marco/photo/panorama/horizon202/horizon202.html
Then they are comparing the Noblex at its focssing distance (assumed
10.4 m) with the Horizon at its focussing distance (infinity). I was
doing the comparison between both cameras used to photograph a scene
at effective infinity. After all, this is what most people buy
swing-lens cameras to do.
> Because of the swing lens design, only the center stripe of the lens image
> circle is used to project an image, so edge or corner dropoffs are not an
> issue either, right? Again, the center of the lens performance on the
> short (vertical) 35mm film axis is all that counts, and is likely to be
> very good (only 24mm high).
Put in whatver values you like into the calculation. What do you
suggest? But don't forget the diffraction that will limit the
resolution at small f stops.
> The distortions seen in swing lens images are
> generally the result of the swing lens design, not lens distortions, yes?
That is nothing to do with the resolving power of the lens. In fact, a
lens with improved correction of distortion might have a lower
resolving power than one that is not so well designed to correct this.
> The larger negative size of the noblex 150 would make it the clear winner
> in a limiting enlargement size contest against the 35mm film based horizon
> 202. The case is much less clear for the noblex 135 series of similar 35mm
> format type - see http://www.a1.nl/phomepag/markerink/pan_35mm.htm
That was stated in my calculations.
> The other point, as another posted noted IIRC, is that there are various
> models and variants of the Noblex 150 series, some of which are focused
> for infinity (as is the horizon 202 series), others for closer distances.
Which Noblex 150 model is focussed at infinity? Have you checked the
focussing distance in the specs? I think you will find that "infinity"
is not really infinity when you do. Another worry, for the focussing
models, is the movement of the secondary principal point off the exis.
If they are moving this to get infinity more in focus then if it is
done by moving the secondary principal point off the axis then this
will have a large negative effect on horizontal resolution because the
image of an object will not stay fixed at any point on the film
surface when the lens swings round.
> I and others have shown how easy it is to use a fractional + diopter lens
> to shift the horizon 202 focus to closer distances, so this is even less
> of an issue with the horizon 202 (see http://medfmt.8k.com/mf/horizon.html
That's irrelevent as this thread was to do with focssing at infinity.
> In short, the lens quality of the horizon 202 is very close or only
> "slightly inferior" to that of the Noblex models costing 500% or so more $
> (in same 35mm film format) or even more (in MF 120 formats).
At the distances they are focussed to, yes. But the comparison I made
was with a scene effectively at infinity. Then, for the Noblex 150
models, which are not fix-focussed at infinity, the circles of
confusion play a major role in limiting the sharpness.
> I think it is important to challenge the case that the horizon 202 optics
> are somehow seriously less capable or poor resolution (as implied by the
> 110 lpmm aerial lens resolution value cited in this thread and related
> conclusions) because it might dissuade some buyers from considering these
> Horizon 202 cameras on their merits - which are many, including low cost.
> If anything, the horizon 202 users usually rate these cameras very highly
> for their lens resolution and sharpness (as do the Noblex owners, with
> similar good reasons ;-)
Then do some research and come back with a realistic figure for lens
resolution over most of the film area. Don't forget that at f16,
diffraction will limit the lens resolution to 125 lp/mm and at f8 it
will be limited to 250 lp/mm. Also don't forget that you will be
limited by film resolution. Suppose you opted for 250 lp/m over most
of the film area (my guess is that this is too high). Let's do a
comparison with my suggestion of 110 lp/mm. Here are the calculations
for on-film resolution for film rated as being capable of 100 lp/m.
110*100/(110+100) = 52 lp/mm
250*100/(250+100) = 71 lp/m
It's an improvement, yes, but most people use the Horizon 202
hand-held and there is a lot of kick and vibration in that camera as
it starts its swing and during its swing so resolution will be limited
in any case, unless it is mounted on a sturdy tripod. Even then there
will be some vibration left in the camera affecting image quality. If
I could do a calculation to predict the negative effect then I would.
But I can't yet, so you are spared the misery of scanning over it and
it not sinking in.
I have the Horizon 202 and I am pleased with the results I get out of
it. It is not very well made, though. Nobody should expect them to
last for a large number of rolls through. And the banding of open blue
skies will be obvious if you use slide film. But while it is working,
it should give good results and if you know the right sources then
these cameras are cheap. The counter in mine jammed and broke while I
was at the top of a mountain in Germany, but the camera is still
working. I have to watch out for the aperture change so that it does
actually reflect the change in the lens. Sometimes I have to "loosen
it up" by putting it through a selection of settings. But don't let
that put anybody off buying one. I don't regret buying mine and when
it breaks I will probably get another one.
> my $.02 ;-)
>
> bobm
RolandRB
> Huh? If you have lens aerial resolutions, and the manufacturers' film
> resolution values, you can predict on-film resolution rather closely using
> the system resolution equation (see http://medfmt.8k.com/mf/lenslpm.html).
<rest of interesting stuff snipped>
I looked at that page in the section on the "sweet spots of lenses
explained". Could you show the calculations that gave the value for
system resolution in that table there?
RolandRB
<snip>
> The flip side is that you can run system resolution estimates, using some
> rather discouragingly modest film resolution limits (e.g., most color
> print films), and discover why high aerial resolution lenses don't perform
> noticeably better than lower resolution (and cheaper) lenses - the color
> print film limiting resolution is by far the limiting factor. Again, a
> valuable insight for 94% of consumers who shoot only color print films and
> no slides or B&W ;-)
But surely things have improved with colour print film.
http://creekin.net/films.htm
RolandRB
> >No lens is perfect, of course. I remember you once
> >posted a similar list for typical brand name lenses.
>
> Yes, let me post both tables so you can see the difference.
>
> Again noon summer sun center resolution.
>
> perfect lens typical name brand lens
>
> f-1 2000l/mm 100
> f-2 1000 300
> f-4 500 500
> f-8 250 250
> f-16 125 125
>
> You can see at about f-4 the perfect lens and the name brand lens kinda match.
> Remember, we are still talking cente rand NOT edge resolution.
>
> Larry
How big is this centre? 1mm diameter? How does it change as you move
away from the centre? What would it be at the edges, roughly? It would
be useful to have an idea of this.
Ralf R. Radermacher
> It's an improvement, yes, but most people use the Horizon 202
> hand-held and there is a lot of kick and vibration in that camera as
> it starts its swing and during its swing so resolution will be limited
> in any case, unless it is mounted on a sturdy tripod.
Most people are capable of keeping a camera remarkably steady although
for a limited period of time. A little training and a proper breathing
technique help a lot. This works nicely with the Horizon. Most of its
shake occurs at the very beginning and end of the drum rotation, i.e. at
times when the shutter is not yet open or has closed again.
The time my Noblex takes to even begin recording the frame wastes most
of my capability to keep it steady.
jjs
R. Radermacher) wrote:
> Most people are capable of keeping a camera remarkably steady although
> for a limited period of time. [...]
'Remarkably steady' has no number. If your heart beats, if you have the
normal human nervous system, you can't ever say you can handhold steadily.
RolandRB
> RolandRB <rolan...@hotmail.com> wrote:
>
> > It's an improvement, yes, but most people use the Horizon 202
> > hand-held and there is a lot of kick and vibration in that camera as
> > it starts its swing and during its swing so resolution will be limited
> > in any case, unless it is mounted on a sturdy tripod.
>
> Most people are capable of keeping a camera remarkably steady although
> for a limited period of time. A little training and a proper breathing
> technique help a lot. This works nicely with the Horizon. Most of its
> shake occurs at the very beginning and end of the drum rotation, i.e. at
> times when the shutter is not yet open or has closed again.
>
> The time my Noblex takes to even begin recording the frame wastes most
> of my capability to keep it steady.
>
> Ralf
For me, at least, the Horizon 202 is difficult to keep still. I have a
Leica M6 and assorted lenses and if I am taking a longish exposure
photograph and I do not have a tripod or monopod with me then I use
rapid breathing for a few seconds to give me a two second break to do
the shot. I can get down to quite long exposure times with that camera
with good results. I do this with the Horizon 202 handheld as well but
am well aware of the starting jolt. I keep looking through the
viewfinder to keep the bubble in the middle of the circle or at least
within the boundary of it. I always keep it within the boundary but I
am well aware that I have moved the camera body during this time. At
some stage I can bore everybody with an actual calculation of shake
and the effect on lens (not film) resolution but I haven't got round
to doing the maths yet. But I aim to do this in the future -- so
"sorry everybody".
Ralf R. Radermacher
> 'Remarkably steady' has no number. If your heart beats, if you have the
> normal human nervous system, you can't ever say you can handhold steadily.
It may not work in your theory. But it does in my practice.
Q.G. de Bakker
> It may not work in your theory. But it does in my practice.
Which show that whether things are good, or not, not just depend on the
quality of the thing itself, but in equal degree also on the standards we
set to measure them against.
Hemi4268
>and the effect on lens (not film) resolution but I haven't got round
>to doing the maths yet. But I aim to do this in the future -- so
>"sorry everybody".
These calculations have alread been done. The secret is to figure out shake in
milliradens per second. Then figure out image motion vs focal length and
resolution. Last is to calculate where in the sine curve the photographs are
most likely to be imaged.
The results can be read out in chances in 10. In other words with a 35mm lens
at 1/35 of a second you have 5 chances out of ten to at least get one image
with no motion. So several photographs will be enough.
Change the lens to 135mm with the same 1/35 of a second and the chances go down
to 1 in 10. In other words, you must take 10 images to get one that is motion
free.
Larry
jjs
hemi...@aol.com (Hemi4268) wrote:
> [...] In other words, you must take 10 images to get one that is motion
> free.
In other words, use a tripod.
David J. Littleboy
"Hemi4268" <hemi...@aol.com> wrote:
> > I can bore everybody with an actual calculation of shake
> >and the effect on lens (not film) resolution but I haven't got round
> >to doing the maths yet. But I aim to do this in the future -- so
> >"sorry everybody".
>
> These calculations have alread been done. The secret is to figure out
shake in
> milliradens per second. Then figure out image motion vs focal length and
> resolution. Last is to calculate where in the sine curve the photographs
are
> most likely to be imaged.
>
> The results can be read out in chances in 10. In other words with a 35mm
lens
> at 1/35 of a second you have 5 chances out of ten to at least get one
image
> with no motion. So several photographs will be enough.
Hey! The theory works! Surprise! (I was about to write a note reporting my
experiments that indicate that the probability of blur increases at slower
shutter speeds.)
RolandRB
Is this calculation for a leaf shutter or does it also work for a
focal plane shutter with a moving slit?
RolandRB
> >You
> >will find that the resolving power of the human eye is very high.
>
> Actually the resolving power of the eye is well documented. It's 8 lines per
> millmeter.
>
> Larry
8 lines per millimeter on the fovea? No chance. 8 lines per millimeter
at a standard distance, perhaps. Can you give me a URL where I can
check on this figure of 8 lines per millimeter? I have to wonder
whether they are measuring something else altogether. For vision on
the fovea I would estimate about 200 lines pairs per millimeter
on-cone resolution is possible for people with excellent vision. I am
looking at my computer screen at the moment and I think I can easily
resolve 5 line pairs per millimeter at 300mm away (1 feet). The focal
length of the eye is 17mm so I make that...
(300/17)*5 = 93 lp/mm
This is on the cones there. And that was a very rough check.
Hemi4268
>at a standard distance, perhaps. Can you give me a URL where I can
>check on this figure of 8 lines per millimeter?
Just about every good manual on printing, something that people have been doing
since the 1500's, will have notations that 8 l/mm is the standard. Lacking
that, try the SPSE manual.
Standard is 8 l/mm at 10 inches view distance.
Larry
RolandRB
As I thought. At a standard distance. Not very useful if quoted
without the distance. So, on the fovea with an eye lens with effective
focal length of 17mm the resolution in line pairs per millimeter on
the cones will be...
(254/17)*8 = 135 lp/mm
Is this for 20/20 vision? I suspect it is. For reading the bottom line
(the one above the line they cross out) of an eye chart? I have had my
eyes tested on a regular basis. Before I was 8 years old I used to
read out the makers name right at the bottom of the chart. I had lost
the ability to do that when I was 9 or 10.
Also which ethnic group is this? I know for Australian aboriginees it
is much higher. You could put another 4 lines below the one they
usually cross out and they can read them if their eyes are good.
The cones of the fovea have a diameter of 1.5 microns on average (this
varies a lot between ethnic groups). So you have a theoretical line
pair in 3 microns. So in line pairs per millimeter you have a
theoretical 333 lp/mm. It won't be quite as high as this since they
are colour cones. And I stress this varies between ethnic groups.
There was a famous Australian aboriginee who could draw a detailed map
of the moon and drew Venus as a crescent. I can not find a URL for
this. If true then that would put the maximum resolving power of the
eye much higher.
Bob Monaghan
see link to diffraction page http://medfmt.8k.com/mf/diffraction.html in
text; use rule of thumb lpmm~= 1600/f# to get end column of table showing
diffraction limits, e.g., 1600/4 (f/4) is 400 lpmm diffraction limit,
1600/8 (f/8) is 200 lpmm diffraction limit at f/8 etc. as shown etc.
the overall system resolution (middle chart) is there to illustrate how a
sample lens might vary in resolution, improving as you stop down and
reduce lens aberrations to some maximum value ("sweet spot"), then
dropping off as diffraction effects become controlling etc.
See postings by Larry at http://medfmt.8k.com/mf/diffraction.html for some
sample patterns for 35mm lenses (e.g., 300 lpmm at f/2, 500 lpmm at f/4,
250 lpmm at f/8, 125 lpmm at f/16 for a typical lens).
The mf/lenslpm.html shows how system resolution is critically dependent on
film resolution limits as the limiting factor. Using a higher resolution
film is the low cost way to improve system resolution ;-)
finally, most Med Fmt lenses are slower designs (f/3.5..) than most 35mm
lenses, and larger aperture physical size for given f/stop, so the sweet
spot of MF lenses tends to be a bit slower (e.g., f/8) than most 35mm
lenses (e.g., f/5.6). Hence the saying, f/8 and be there ;-)
hth
Bob Monaghan
improved? see http://creekin.net/films.htm for consumer speed print
films, most of which run somewhere around 125-130 lpmm (1000:1 contrast)
and 50 lpmm (1.6:1). So for real world scenes (6:1 ratios) you are stuck
in the 60-80 lpmm film resolution limit range, which means system
resolutions more like 50 lpmm on a good day ;-(
That's why tech saavy authors like Roger Hicks (the Lens Book) note that
if you are shooting color print film, you don't need to worry about lens
quality - the film is going to be the limiting factor. Quoting Kodak ;-)
From Lenses for 35mm (Kodak Workshop Series KW-18 1998, p. 33, Artur
Landt):
Photographers who use color negative films to make prints up to 8x10
inches can safely do without tests, since the differences in image quality
will not be noticeable at these small enlargements.
end-quote ;-)
actually, things have gotten worse, since you can't pick the slow color
print films like ektar 25 etc. they don't make it anymore ;-)
Bob Monaghan
it varys by tester, see http://www.hevanet.com/cperez/MF_testing.html for
Chris Perez's useful examples for center/middle/edge values on variations
most magazine reports only cite center and edge resolution...
for lots of sample lens resolution patterns (highlighting the f/8 and be
there rule in most cases ;-) see
http://medfmt.8k.com/third/variations.html
hth bobm
Bob Monaghan
but 1/125th second exposure is still 1/125th etc, even if it takes much
longer to expose the entire film length thru a narrow slit. So the total
vibration is only what you would see with an equivalent camera and lens at
1/125th second speed, not a camera and lens that is open during the
entire start-move-stop cycle. As Ralf noted, most of the motion/noise is
at the start and end of the cycle, when the film is not exposed.
Adding to that, the lens is still a 28mm lens, so you have the usual wide
angle lens benefits too, and on a 24x56mm sized film format IIRC - shot
many 28mm lenses on 6x6cm lately? ;-) Camera shake would be minimized etc
so it isn't any surprise to me that these cameras turn in surprisingly
sharp images.
my $.02 ;-) bobm
Bob Monaghan
the horizon 202 isn't focused at infinity either, right? The horizon 202
is focused at a closer midpoint than the noblex 135 series, so the
noblex 135 should be sharper at infinity, not less as argued, viz.:
noblex 135 f/5.6 2.9m to infinity focused at 7.5m
horizon 202 f/5.6 2.3m to infinity focused at less than 7.5m (6.5m+?)
The noblex 150 E2 is set for 10.4m, which is also closer to infinity than
the Horizon 202, right? ;-) This is the non-focusing model we argue about
;-) see manual specs listing links below etc.
The adjustable Noblex 150 UX lets you pick 17.2m, 6.5m, or 2.8m for
infinity, middle, and closest settings, yes? Again, their infinity setting
is only 17.2m vs. 10.4m in the original non-adjustable models, right?
The other key point is that for whatever COC they are using, Noblex seems
more conservative than horizon 202 (in same noblex 135 format), and this
seems to carry over in MF models too.
The same Noblex manual pages only give examples of using +diopter lenses
(as I did) because those are the only ones you are likely to need with the
camera, since it already always images infinity at its (more conservative
than the horizon 202s) hyperfocal settings.
So if the horizon 202 is good enough at infinity, the noblex should be at
least as good ;-)
grins bobm
==========
from p. 6 horizon 202 manual
http://www.pauck.de/marco/photo/panorama/horizon202/handbook/Horizon-202_Panorama.pdf
f/2.8 5.5m to infin
f/4 3.9m to inf
5.6 2.9m to inf
8 2 m to inf
from http://www.kamera-werk-dresden.de/english/noblex/cameras/frameset.htm
noblex 6/150 UX at inf.
f/4.5 6.9m to inf
f/5.6 6.1m to inf
f/8 4.7m to inf
noblex 6/150 E2
f/4.5 4.6 to inf
f/5.6 4.1 to inf
f/8 3.2m to inf
f/11 2.6m to inf
noblex 135U
f/4.5 2.6m to inf
f/5.6 2.3m to inf
f/8 1.8m to inf
f/11 1.4m to inf
f/16 1m to inf
Peter Irwin
> Let's look at the manuals and specifications again ;-)
>
> the horizon 202 isn't focused at infinity either, right?
I think it is.
I think the DOF for the Horizon is based on a circle of confusion of 0.05mm
and a 28mm lens focused at infinity. Plugging those figures into a DOF
calculator makes a pretty close match to the Horizon 202 specs below.
Peter.
Hemi4268
>text; use rule of thumb lpmm~= 1600/f# to get end column of table showing
>diffraction limits, e.g., 1600/4 (f/4) is 400 lpmm diffraction limit,
Just remember that when these rule or test were layed down, bees wax candles
were the medium of the day.
Noon summer sun is more representive for today.
Larry
Hemi4268
>focal length of 17mm the resolution in line pairs per millimeter on
>the cones will be...
>
>(254/17)*8 = 135 lp/mm
>
>Is this for 20/20 vision?
Sure 135 l/mm might be projected to the back of the eye but only 8 l/mm is
being picked up.
Same issue with a Nikon loaded with tri-x film. At f-8, 250 l/mm is being
projected on the film but the film really only picks up about 40 l/mm of the
250 l/mm.
Larry
RolandRB
At no point in that manual for the Horizon 202 does it say what
distance the lens is focussed to. If I am wrong then point it out to
me. I have taken night shots with that camera. 1/2sec exposure and
f2.8 aperture. The distant lights are sharp. If it isn't focussed at
infinity then it is very close.
Don't forget that depth of field is subjective. The acceptable circles
of confusion might be different in the cases quoted above. The Horizon
manufacturers might consider a larger circle of confusion is
acceptable. The Noblex makers might be assuming a smaller size. The
Noblex figures might be assuming the usual CoC for the 35mm format.
The Horizon manufacturers might be assuming a larger CoC because of
the wider film width than 24x36mm. They would be quite justified in
doing so. These can explain the differences in the figures above.
I'll do the calculatiions for you for the Horizon 202, if you like,
but no doubt I will have to stress them again. What the Horizon people
are doing is using a CoC of 0.05mm because the film is wider than the
24x36mm 35mm format size.
Here goes. Using the top calculator at
http://www.dudak.baka.com/dofcalc.html Use meters and a custom CoC of
0.05 and a focussing distance of 280m (1000 times focal length and
therefore good as infinity) and a focal length of 28mm and put in the
f-stop numbers. Here is what I get:
f-stop near focus far focus
2.8 5.5m inf
4 3.9m inf
5.6 2.8m inf
8 1.9m inf
Now compare that with your figures above as quoted in the manual which
I repeat here.
> f/2.8 5.5m to infin
> f/4 3.9m to inf
> 5.6 2.9m to inf
> 8 2 m to inf
They are as good as the same. So this lens is focussed at 280m or
thereabouts which is effectively infinity.
Now for the Noblex 135U. Now they say there are using an out-of-focus
field of 0.05 but they are not. They are using 0.04. So do the same
again with the lens focussed at 7m and not the 7.5 meters they say.
Here is what I get:
f-stop near focus far focus
4.5 2.7m inf
5.6 2.3m inf
8 1.8m inf
11 1.4m inf
16 1.0m inf
I will repeat your values here for the Noblex 135U
> noblex 135U
>
> f/4.5 2.6m to inf
> f/5.6 2.3m to inf
> f/8 1.8m to inf
> f/11 1.4m to inf
> f/16 1m to inf
Very nearly the same.
So my figures seem to prove that the Horizon 202 is effectively
fixed-focussed to infinity (1000 times focal length is effectively
infinity). And this is backed up with my taking night photos at f2.8
and having distant street lamps in focus. Their DoF values assume a
CoC of 0.05mm. The Noblex 135U are using a CoC of 0.04mm even though
they state 0.05mm and their lens is most likely fixed-focussed at 7m
and not 7.5 meters as they state at
http://www.kamera-werk-dresden.de/english/noblex/cameras/frameset.htm
I hope you have a chance to plug in these values into this calculator
so you are convinced.
But to state again -- Horizon 202 -- is fix-focussed to effective
infinity. The calculations above back it up and if you use it at f2.8
then infinity is in sharp focus.
Roland
RolandRB
This is on-cone resolution. You see it, right? If you are looking at 8
line pairs per mm ten inches away then you must have an image on the
fovea with an on-cone resolution of 135 lp/mm. That is where the image
is, is it not? On-cone, I stress, because this is what you are
actually seeing.
Bob Monaghan
the problem is that the lens is not used as a fixed 28mm on 24x36mm
format, but rather to create a 24x56mm image covering circa 110 degrees,
vastly more than a basic 28mm lens, so the DOF calculations done assuming
a fixed lens value don't seem to apply.
put another way, comparing the noblex 135 series with the horizon 202
(both are swing lens, 35mm film used, similar formats, slightly wider on
noblex (120 vs 110 deg), but here, the focusing point is circa 7.5 meters
per specs at noblex site, and horizon is nearly identical infinity to
close focus settings, so it must also be fixed focus for some point, not
infinity, but rather, close to 7.5 meters (but less than 7.5 meters).
Again, I posted these values and URLs in a related post in this thread...
so the horizon captures infinity at every setting, but it does so as one
end of its hyperfocal range, with the basic 28mm lens being fixed focused
at some point circa 7 meters from the camera, based on analogy to the
known case of the noblex 135 series which it closely matches in focusing
distances and format etc.
in short, it isn't infinity, or the close focusing points would be much
farther away than they are, in fact, it is focused closer than 7.5 meters
because the noblex 135 has farther away close focusing points by a bit, so
the horizon is focused closer.
I have heard and been told the horizon was focused at infinity too, but
looking at the noblex 135 specs makes it clear that isn't the case ;-)
regards bobm
Bob Monaghan
the problem with using that calculator is the underlying assumption of a
fixed lens and standard film format, although it is nice that you can vary
the CoC ;-)
the 28mm lens on the horizon 202 is not acting as a fixed 28mm lens on
35mm film format, which this calculator assumes. The swinging effect makes
it cover 110 degrees vs. much less for a fixed 28mm lens, and the film
format is not 24x36mm but 24x56mm, which is again hugely different etc.
in other words, I doubt the calculated values apply, unless you use an
equivalent format (perhaps 6x6cm) and equivalent lens FOV coverage for the
swing lens (and not the fixed 28mm of the basic lens).
that's why I used the tables provided by the mfgers of Noblex 135, and by
showing the horizon and noblex 135 of similar lens type and film format
and swing lens design, concluded the mfgers 7.5m focus distance for the
noblex 135 was nearly the same as that of the horizon 202 as I noted.
Your analysis requires us to assume the mfgers don't know the distance to
which their lens if fixed focused, which seems a rather odd conclusion ;-)
You also have to ignore the fact the calculator being used is designed for
fixed lenses and other formats, while the swing lens design is not
handled, nor do you account for the actual field of view of the swing lens
design. That's much wider than the basic 28mm lens, obviously, so your
calculations based on a fixed 28mm lens don't apply.
Again, I think Noblex 135 mfgers know where their lens is fixed focus in
meters (7.5m), and the match up with the similar design and format horizon
202 suggests the horizon is similarly focused at the same fixed focus
point, viz. 7 meters or so. Since your calculations yield 40 times farther
distances, I presume the calculator and assumptions used must be in error.
again, my $.02 ;-)
bobm
jjs
Does the Noblex lens revolve about its nodal point?
Q.G. de Bakker
> but 1/125th second exposure is still 1/125th etc, even if it takes much
> longer to expose the entire film length thru a narrow slit. So the total
> vibration is only what you would see with an equivalent camera and lens at
> 1/125th second speed, not a camera and lens that is open during the
> entire start-move-stop cycle. As Ralf noted, most of the motion/noise is
> at the start and end of the cycle, when the film is not exposed.
Two things come to mind.
First, people will disagree about the detrimental effects of handholding
(which are present at any (!) speed). Seems part of what being human is. I
think we have arrived at the "you would not believe, though i'm asking you
to, the olympic proportions of the rock-solidness of my handholding" stage.
;-)
Second, while any bit of exposed film may only be exposed for a very short
time, reducing the effects on shake a bit, handholding a scanning camera
will lead to something else than unsharpness, to wit distortion.
But that's not what was being discussed, so please ignore this. ;-)
RolandRB
> the problem with using that calculator is the underlying assumption of a
> fixed lens and standard film format, although it is nice that you can vary
> the CoC ;-)
>
> the 28mm lens on the horizon 202 is not acting as a fixed 28mm lens on
> 35mm film format, which this calculator assumes. The swinging effect makes
> it cover 110 degrees vs. much less for a fixed 28mm lens, and the film
> format is not 24x36mm but 24x56mm, which is again hugely different etc.
This makes no difference to what it is calculating which is the depth
of field given a CoC which you choose.
> in other words, I doubt the calculated values apply, unless you use an
> equivalent format (perhaps 6x6cm) and equivalent lens FOV coverage for the
> swing lens (and not the fixed 28mm of the basic lens).
>
> that's why I used the tables provided by the mfgers of Noblex 135, and by
> showing the horizon and noblex 135 of similar lens type and film format
> and swing lens design, concluded the mfgers 7.5m focus distance for the
> noblex 135 was nearly the same as that of the horizon 202 as I noted.
Well you got it wrong. It is clear from my calculations that the
Horizon 202 depth of field was assuming a CoC of 0.05mm and the Noblex
135U, although stating a CoC of 0.05mm, was in fact using a CoC of
0.04mm. The figures prove it.
> Your analysis requires us to assume the mfgers don't know the distance to
> which their lens if fixed focused, which seems a rather odd conclusion ;-)
The most important adjustment to make if any, on a swing lens camera,
is putting the secondary principal point bang on the axis. If this is
not done then the image of an object on the film through the slit will
shift as the lens rotates and ruin the horizontal resolution. If this
needs adjusting then the distance the lens is focussed to will change.
> You also have to ignore the fact the calculator being used is designed for
> fixed lenses and other formats, while the swing lens design is not
> handled, nor do you account for the actual field of view of the swing lens
> design. That's much wider than the basic 28mm lens, obviously, so your
> calculations based on a fixed 28mm lens don't apply.
That has nothing to do with it.
> Again, I think Noblex 135 mfgers know where their lens is fixed focus in
> meters (7.5m), and the match up with the similar design and format horizon
> 202 suggests the horizon is similarly focused at the same fixed focus
> point, viz. 7 meters or so. Since your calculations yield 40 times farther
> distances, I presume the calculator and assumptions used must be in error.
All you have to do is carefully follow through my argument, as anybody
doing this is free to do, plug in the values and look at the results.
I hope you get time to do that one day.