hyperfocal distance

Gisle Hannemyr wrote:
M Barnes writes:
Gisle Hannemyr wrote:
M Barnes writes:

Anyway, isn't the circle of confusion a characteristic
of the lens, not the sensor?

No - it is proportional to sensor size.

Hm. Okay, I think I get it. How is the CoC computed
for lenses then?

It doesn't make sense to compute a CoC for a lens.

Ah, OK. I see where I went wrong.
This site:
http://www.nikonlinks.com/unklbil/dof.htm#method
talks about using the USAF 1951 lens test chart to
determine a customized CoC value. I read this
long ago, and somewhere in the back of my mind
believed that CoC was a lens parameter.

And for film body/lens combinations?

As a function of the diameter of the film negative (the
body or lens doesn't enter into the equation).

Got it. I heard the "click," and the concept fell
into place like a pachinko ball.

Snip some good stuff

The crop factor for the D100 is 1.5 - so you need to
multiply the hyperfocal distances (hfd) you've worked
out for the N2020 by 1.5 when you use your lenses
on the D100.

Example: If a particular lens has a hfd of 40 ft when used on
a full frame camera at a certain aperture, the same lens set to
the same aperture will have a hfd of 40 x 1.5 = 60 ft on the D100.

You just saved me a weekend with a tape measure
in the back yard. Thanks.

Sorry - the intuitive answer is right.
Yes the more you enlarge the softer the image gets but it's just
plain silly to assume any single size for the final image since as size of
the image increases, viewing distance also increases and they cancel each
other out.

--
http://www.chapelhillnoir.com
home of The Camera-ist's Manifesto
The Improved Links Pages are at
http://www.chapelhillnoir.com/links/mlinks00.html
A sample chapter from my novel "Haight-Ashbury" is at
http://www.chapelhillnoir.com/writ/hait/hatitl.html
"Roland Karlsson" wrote in message
...
"Tony Spadaro" wrote in news:dJCEc.85007
:

No. It remains the same.

Yes - this is the intuitive answer. But it is not
entirely correct. The formula for hyperfocal distans
is (just as pointed out in another post):

h = (f*f)/(N*c)
f = focal length, N = f-ratio, c = "circle of confusion" diameter.

The crucial factor here is c (circle of confusion). If you search
further on the net you will find that it is 1/1740 of the diagonal
of the sensor. Therefore, h will be bigger for a smaller sensors.

So - the intuitive answer is wrong. The hyperfocal distance
depends on how much you crop your image. Therefore, it is also only
valid for the 35 mm film camera if you don't crop the image.


/Roland

On Wed, 30 Jun 2004 23:27:45 +0200, Gisle Hannemyr
wrote:

"leo" writes:
When using the concept of hyperfocal distance, do we need to adjust
for the smaller sensor on 300D?

Yes.

It works out as follows:

hfd = (fl^2)/(fs*CoC)

hfd = hyperfocal distance (in mm)
fl = focal length,
fs = f-stop
CoC = Circle of Confusion (in mm)

The CoC is not a very precise term. It is a measure of the amount
of blur the human eye will consider acceptable. To some extent it
is subjective.

But the canonical value for CoC used in many CoC-spreadsheets is
1/1730 of the diagonal of the capture area (negative size or sensor).

A 35mm negatative has a 43.3 mm diagonal - which gives us a CoC for
35 mm film of 43.3/1730 = 0.025 mm.

A 300D has a sensor with a 27.3 mm diagonal. Doing the same division,
we find that the CoC for a 300D is 0.016. Btw, this means that the
CoC is about 1.5 times the diameter of a pixel.

Example: 50 mm lens at f/8:

35 mm: hfd = 50^2/(8*0.025) = 12 500 mm = 12.5 m = 41 ft
300D: hfd = 50^2/(8*0.016) = 19 531 mm = 19.5 m = 64 ft


Small sensors have greater depth of field than larger sensors.
Therefore the hyperfocal distance--the lens setting where you have
acceptable focus from infinity to half the hyperfocal distance--
becomes smaller not larger, You divide by the 35 mm equivance ratio
not multiply as you claim

See http://dpfwiw.com/exposure.htm#hyperfocal for the correct formula.

jpc

Tony Spadaro wrote:

Sorry - the intuitive answer is right.
Yes the more you enlarge the softer the image gets
but it's just plain silly to assume any single size for
the final image since as size of the image increases,
viewing distance also increases and they cancel each
other out.

What formula are you using to compute hyperfocal
distance that uses increasing image sizes and increasing
viewing distances? I don't see this anywhere.

The formula that I've seen uses focal length,
f ratio, and CoC. The first two are characteristics
of the lens, so image size does not enter. If the
CoC is taken to be the inverse of the resolution factor,
and the rf is taken to be 1525/d, where d is the
diagonal measurement of the sensor. In these
computations, the image size is held constant
at 25 cm, with an 8x10 image size.

In all the discussions I read so far in the last couple
of days, a family of CoCs can be generated for
different sensor/image/viewing distance combinations,
but to compare them across families is of no value.

I don't see how you can compute hyperfocal distance without knowing
something about the format size, resolution, focal length of the lens,
and its f/#, so isn't that already considering the sensor size?

leo wrote:

When using the concept of hyperfocal distance, do we need to adjust for the
smaller sensor on 300D?

--
Don Stauffer in Minnesota

webpage- http://www.usfamily.net/web/stauffer

But nobody is talking about a P&S with a 10mm normal lens - they are usued
for 4x6 prints at the largest and quite frankly no one cares how sharp the
picture is -- it's uncle Harry and aunt Matilda at the Grand Canyon.

--
http://www.chapelhillnoir.com
home of The Camera-ist's Manifesto
The Improved Links Pages are at
http://www.chapelhillnoir.com/links/mlinks00.html
A sample chapter from my novel "Haight-Ashbury" is at
http://www.chapelhillnoir.com/writ/hait/hatitl.html
"Gisle Hannemyr" wrote in message
...
"Tony Spadaro" writes:
Sorry - the intuitive answer is right.

No. The intuitive answer (i.e. that you don't need to take the crop
factor into account when figuring hyperfocal distance for a lens) is
wrong.

It may not be obvious that this is wrong as long as we are talking
about a 300D. This camera have 1.6x crop factor, and blur tolerance
is after all subjective. Even if you don't bother adjusting for the
crop factor, the results are not that that much «off». They may
look just fine to most viewers.

But it becomes blantatly obvious that ingnoring the crop factor don't
cut the mustard if we move into compact territory.

Let's take a canon G5 (4.9x crop factor). A «normal» lens for
this camera is a 10 mm.

If we just compute the hyperfocal distance for, say f/5.6, «as if» we
were dealing with a full frame sensor, we get a hyperfocal distance
equal to 1.97 ft, and everything from 1.02 to infinity should be in
acceptable focus.

Well, a simple test will verify that this is /not/ the case.

If we take the crop factor into consideration, and recompute, the
hyperfocal distance at 10mm f/5.6 becomes 9.7 ft - which tests
will verify is about right.

Yes the more you enlarge the softer the image gets but it's just
plain silly to assume any single size for the final image since as
size of the image increases, viewing distance also increases and
they cancel each other out.

This has nothing to do with it.
--
- gisle hannemyr [ gisle{at}hannemyr.no - http://folk.uio.no/gisle/ ]
================================================== ======================
«To live outside the law, you must be honest.» (Bob Dylan)

Don't get all bogged down in formulas. Take an 8x10 print and hold it
where you can see the entire print at one time -- this is proper viewing
distance for an 6x10. Now take an 11x14 and do the same -- it is farther
away, isn't it? A 16x20 will be farther away than the 11x14 and when you get
up to 30x40 you should be halfway across the room.

--
http://www.chapelhillnoir.com
home of The Camera-ist's Manifesto
The Improved Links Pages are at
http://www.chapelhillnoir.com/links/mlinks00.html
A sample chapter from my novel "Haight-Ashbury" is at
http://www.chapelhillnoir.com/writ/hait/hatitl.html
"M Barnes" wrote in message
...
Tony Spadaro wrote:

Sorry - the intuitive answer is right.
Yes the more you enlarge the softer the image gets
but it's just plain silly to assume any single size for
the final image since as size of the image increases,
viewing distance also increases and they cancel each
other out.

What formula are you using to compute hyperfocal
distance that uses increasing image sizes and increasing
viewing distances? I don't see this anywhere.

The formula that I've seen uses focal length,
f ratio, and CoC. The first two are characteristics
of the lens, so image size does not enter. If the
CoC is taken to be the inverse of the resolution factor,
and the rf is taken to be 1525/d, where d is the
diagonal measurement of the sensor. In these
computations, the image size is held constant
at 25 cm, with an 8x10 image size.

In all the discussions I read so far in the last couple
of days, a family of CoCs can be generated for
different sensor/image/viewing distance combinations,
but to compare them across families is of no value.

Tony Spadaro wrote:

Don't get all bogged down in formulas ....

Well, frankly, when a technical question
is asked -- "Is the hyperfocal distance
affected by sensor size?" -- then I expect
that the technical, formulaic if you will,
answer is the correct one.

Engineering education consists not only
in learning boring stuff -- formulae -- but
also in overcoming intuition, which is often
incorrect.

If you can't show logically how your answer --
that hyperfocal distance is not affected by
sensor size -- fits the published engineering
formulae and data, then I must assume that
it is incorrect since it has been contradicted
by other posters who can do so.

Take an 8x10 print and hold it where you
can see the entire print at one time -- this
is proper viewing distance for an 6x10.
Now take an 11x14 and do the same ....

But this is not relevant to the discussion at
hand. The viewing distance factor is considered
in the CoC formula, and applies to a family
of hyperfocal calculations. Can you show
with logical, formulaic methods how the
viewing distance of a print has an effect on
hyperfocal distance calculated for a sensor
of a given size?

I'm willing to listen to your arguments, but
I'm not willing to forego the use of formulae
in making technical calculations.

If you are saying that adjusting sensor size
causes the calculations to slide up and down
commensurately -- through the focal length
and focal ratio calculations, or perhaps the
CoC calculation -- and thus cancel out any
hyperfocal distance changes, I would need
to see this cranked through the accepted
formulae published all over the world and
accepted by everybody in the business.

I don't see how comparing print sizes has
anything to do with this discussion.

On Thu, 1 Jul 2004 08:59:07 -0500, "M Barnes"
wrote:

Tony Spadaro wrote:

Sorry - the intuitive answer is right.
Yes the more you enlarge the softer the image gets
but it's just plain silly to assume any single size for
the final image since as size of the image increases,
viewing distance also increases and they cancel each
other out.

What formula are you using to compute hyperfocal
distance that uses increasing image sizes and increasing
viewing distances? I don't see this anywhere.

The formula that I've seen uses focal length,
f ratio, and CoC. The first two are characteristics
of the lens, so image size does not enter. If the
CoC is taken to be the inverse of the resolution factor,
and the rf is taken to be 1525/d, where d is the
diagonal measurement of the sensor. In these
computations, the image size is held constant
at 25 cm, with an 8x10 image size.

Question:
Using that formula, does it work for *any* sensor size, or the one the
image size on the focal plane was designed for (in the case of the
lenses in question, 35mm)?
Or, to put it a different way, if you take a 35mm flm image at
hyperfocal distance, does cropping that image alter the hyperfocal
distance, or was the HD set when the pic was taken?

I'm wondering, if the CoC formula includes 1525d (and I'm assuming it
does), does d refer to the sensor, or the image on the focal plane,
and the sensor size that image is designed for?
I mean, in a DSLR, the lens uses a smaller part of the image on the
focal plane than 35mm film does. In effect, it crops that image. As I
ask above, does this really change the hyperfocal distance of that
lens?

In all the discussions I read so far in the last couple
of days, a family of CoCs can be generated for
different sensor/image/viewing distance combinations,
but to compare them across families is of no value.


Bill Funk
Change "g" to "a"

On Thu, 01 Jul 2004 10:06:08 -0500, Don Stauffer
wrote:

I don't see how you can compute hyperfocal distance without knowing
something about the format size, resolution, focal length of the lens,
and its f/#, so isn't that already considering the sensor size?

But, that begs a question:
Using DSLRs, which sensor size do you use? The one that's actually
there, or the one that everything else was designed for (35mm)?

leo wrote:

When using the concept of hyperfocal distance, do we need to adjust for the
smaller sensor on 300D?

Bill Funk
Change "g" to "a"