Pages: [1]   Go Down

Author Topic: Hyperfocal distance question  (Read 3822 times)

Edalongthepacific

  • Full Member
  • ***
  • Offline Offline
  • Posts: 118
Hyperfocal distance question
« on: August 22, 2012, 12:08:57 pm »

If a 85mm portrait lens is used on a ASP C sensor the crop factor (1.5) would give the lens an equivalent field of view of a 128mm lens. To determine the distance from a person we would need to be to obtain a satisfactory focus when using f/8 we would could use the formula: H= (f x f)/N, that is: 128 x 128/8 = 2048 mm or 6.7 feet (disregarding circle of confusion). So we must be at least 6.7 feet from our person before we could use the lens as stated. Is my math & understanding correct?
Logged

fike

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 1413
  • Hiker Photographer
    • trailpixie.net
Re: Hyperfocal distance question
« Reply #1 on: August 22, 2012, 02:36:38 pm »

This doesn't look right to me. You can't ignore circle of confusion.

http://www.cambridgeincolour.com/tutorials/hyperfocal-distance.htm
Their calculator shows 148 feet, but they don't tell us what circle of confusion they use. After some calculating, they used 0.02mm.  This seems to be the general preference for many of the calculators because people can't be bothered to figure out what they want. 

http://en.wikipedia.org/wiki/Hyperfocal_distance
Their formula is easy to follow. 

In the past, when I did some print comparisons, I came up with 0.01mm circle of confusion for high quality APS-C work printed at 240 DPI and viewed up close (less than 1 foot). Depending on intended conditions and enlargement factor this can be anywhere between 0.01mm and 0.02mm (more if you have very large viewing distances like a billboard). 

You don't need to multiply the focal length by the crop factor.  The crop factor is considered in the circle of confusion you choose.  With smaller formats (smaller pixels) you need a lower circle of confusion to achieve the same quality. So, your result should be about 148 feet with a circle of confusion of 0.02mm. 

Logged
Fike, Trailpixie, or Marc Shaffer

Edalongthepacific

  • Full Member
  • ***
  • Offline Offline
  • Posts: 118
Re: Hyperfocal distance question
« Reply #2 on: August 22, 2012, 02:45:47 pm »

Thank you. The CoC, is this an aspect of the accuracy of the lens's focus or an aspect of sensor quality or both?
Logged

fike

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 1413
  • Hiker Photographer
    • trailpixie.net
Re: Hyperfocal distance question
« Reply #3 on: August 22, 2012, 02:51:29 pm »

It is a function of the sensor, output resolution and viewing distances. Take a look at this, and come back with questions.

http://www.cambridgeincolour.com/tutorials/depth-of-field.htm
Logged
Fike, Trailpixie, or Marc Shaffer

Edalongthepacific

  • Full Member
  • ***
  • Offline Offline
  • Posts: 118
Re: Hyperfocal distance question
« Reply #4 on: August 22, 2012, 08:20:05 pm »

So every lens comes to a sharp focus at some point in space but, hopefully, on the surface of the sensor. There is a range of acceptable focus on either side of the focus point known as the depth of field. The depth of field lies within the CoC. If the CoC can be seen on a print at a reasonable viewing distance, the focus lies outside the DOF.
Logged

Glenn NK

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 313
Re: Hyperfocal distance question
« Reply #5 on: August 22, 2012, 10:22:12 pm »

If a 85mm portrait lens is used on a ASP C sensor the crop factor (1.5) would give the lens an equivalent field of view of a 128mm lens. To determine the distance from a person we would need to be to obtain a satisfactory focus when using f/8 we would could use the formula: H= (f x f)/N, that is: 128 x 128/8 = 2048 mm or 6.7 feet (disregarding circle of confusion). So we must be at least 6.7 feet from our person before we could use the lens as stated. Is my math & understanding correct?

In any formula using the FL, one should use the actual FL of the lens.  The crop factor does not change the FL one iota.

This is a common confusion - a smaller sensor simply doesn't "see" as much image as a FF sensor would.  In fact, a FF sensor doesn't "see" everything that a lens transmits - the image projected by a lens is circular, so a considerable amount of potential image is not recorded by the rectangular sensor/film.

The optical properties of a lens cannot be changed by the sensor/film.

Glenn
Logged
Economics:  the study of achieving infinite growth with finite resources

Edalongthepacific

  • Full Member
  • ***
  • Offline Offline
  • Posts: 118
Re: Hyperfocal distance question
« Reply #6 on: August 23, 2012, 02:50:06 am »

Hyperfocal distance (H) = ((focal length squared) / (f-stop x CoC)) + (focal length).

The most confusing part of this formula is the blur circle or circle of confusion (CoC). Imagine that for any depth of field a lens focuses most precisely at some point within that depth. There is an acceptable focus for some distance on either side of that point. But the image is not formed out in space in front of the camera. The image is formed on the censor. The image comes to a focus on the censor within a measurable circle known as the circle of confusion (CoC). “The depth of field is the region where the CoC is less than the resolution of the human eye (or of the display medium).” If the CoC is too large and becomes visible on an 8 x 10 photograph when viewed at a reasonable viewing distance, the focus is outside the depth of field and is not acceptably sharp. The CoC is a factor of the sensor, output resolution and media viewing distance.  
« Last Edit: August 23, 2012, 02:52:34 am by Edalongthepacific »
Logged

torger

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 3267
Re: Hyperfocal distance question
« Reply #7 on: August 24, 2012, 09:31:44 am »

Hyperfocal distance (H) = ((focal length squared) / (f-stop x CoC)) + (focal length).

The most confusing part of this formula is the blur circle or circle of confusion (CoC).

I think that too. For hyperfocal calculations on modern digital cameras the most useful measure of CoC I think is ~2x airy disc, which means a larger CoC the more diffraction you got (smaller aperture).

The reason for this choice is that at the largest magnification level the blur at the DoF edge is about the same as the diffraction blur in the plane of focus. If you need more DoF (closer hyperfocal distance) you stop down more and then you get more diffraction and larger CoC to get in balance. The old traditional large CoC I think makes the hyperfocal concept next to useless.

I've written more about this at: http://www.ludd.ltu.se/~torger/photography/focus-landscape.html
Logged

fike

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 1413
  • Hiker Photographer
    • trailpixie.net
Re: Hyperfocal distance question
« Reply #8 on: August 24, 2012, 09:43:49 am »

I think that too. For hyperfocal calculations on modern digital cameras the most useful measure of CoC I think is ~2x airy disc, which means a larger CoC the more diffraction you got (smaller aperture).

The reason for this choice is that at the largest magnification level the blur at the DoF edge is about the same as the diffraction blur in the plane of focus. If you need more DoF (closer hyperfocal distance) you stop down more and then you get more diffraction and larger CoC to get in balance. The old traditional large CoC I think makes the hyperfocal concept next to useless.

I've written more about this at: http://www.ludd.ltu.se/~torger/photography/focus-landscape.html

Anders, that is an interesting treatise.  I have been trying to envision a graph that illustrates the competing attributes of DOF and diffraction as aperture increases/decreases.  Something like sharpness on the vertical and aperture on the horizontal, but this is a three variable problem (DOF, aperture and some value relating to sharpness) so the graph becomes something planar and a bit harder to visualize for most people.
Logged
Fike, Trailpixie, or Marc Shaffer

torger

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 3267
Re: Hyperfocal distance question
« Reply #9 on: August 25, 2012, 05:46:05 pm »

the most useful measure of CoC I think is ~2x airy disc

I was a bit unclear; 2x airy disk *radius* that should be, or CoC = airy disk diameter. It is an approximation, but a good enough such I think. If the edge should be 100% indistinguishable from plane of focus some value 0.6 x airy disk diameter would probably be closer to the thruth, but if we are at that detail level there are probably other small factor to take into account to like wavelength distribution etc. One could probably also want a smaller factor for large apertures and a larger when there is much diffraction due to change in image quality expectation. So if one wants to make the "perfect" hyperfocal table there is more to do :-)
Logged

u2jimbo

  • Newbie
  • *
  • Offline Offline
  • Posts: 22
    • exploring photography
Re: Hyperfocal distance question
« Reply #10 on: August 28, 2012, 12:35:12 pm »

Hi everyone!

I hope I'm not interfering with edalongthepacific's original question...  However, this topic directly relates to my recent effort to begin learning how to use the hyperfocal distance to my advantage (I am really attracted to photos I see where there is a subject in the extreme foreground in very sharp focus and the the horizon is also in very sharp focus - so I am trying to learn the technique).  I also am trying to appreciate the role defraction plays in undermining these efforts.

I began by reading the tutorials on Cambridge in Colour.  Then went outdoors to test my understanding.  I used a Depth of Field calculator on my iPhone to (using 0.03 for the Circle of Confusion factor) provide me with the hyperfocal distance for f8 on my Canon 5DmkII wi/ an 18-35mm, then used a tape measure to ensure my foreground subject  was at that precise distance.  I took the shot. I then took shots at f22 for comparison.

Back home, in comparing the two, the one shot at f22 displayed a much broader range of sharp focus than the f8 shot.

Is anyone else surprised?  Am I missing something in my setup process?

Thanks
Jim
Logged

fike

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 1413
  • Hiker Photographer
    • trailpixie.net
Re: Hyperfocal distance question
« Reply #11 on: August 28, 2012, 01:23:09 pm »

Hi everyone!

I hope I'm not interfering with edalongthepacific's original question...  However, this topic directly relates to my recent effort to begin learning how to use the hyperfocal distance to my advantage (I am really attracted to photos I see where there is a subject in the extreme foreground in very sharp focus and the the horizon is also in very sharp focus - so I am trying to learn the technique).  I also am trying to appreciate the role defraction plays in undermining these efforts.

I began by reading the tutorials on Cambridge in Colour.  Then went outdoors to test my understanding.  I used a Depth of Field calculator on my iPhone to (using 0.03 for the Circle of Confusion factor) provide me with the hyperfocal distance for f8 on my Canon 5DmkII wi/ an 18-35mm, then used a tape measure to ensure my foreground subject  was at that precise distance.  I took the shot. I then took shots at f22 for comparison.

Back home, in comparing the two, the one shot at f22 displayed a much broader range of sharp focus than the f8 shot.

Is anyone else surprised?  Am I missing something in my setup process?

Thanks
Jim

First off, Great Job in running your own tests and experiments.  Based on your output format (small prints, online display, etc...) and your personal subjective assessment, it COULD be okay to shoot f/22 all the time.  Doing experiments is how you decide that and learn.

Now for a question:  what focal length did you plug into the calculator?  You don't mention the actual focal length, and I want to make sure you setup the test correctly.  If done correctly, and your image was framed so that it only contained the 1/2*hyperfocal distance to infinity, I would expect the f/8 image to be in better focus than the f/22 image.  At f/8 and 18mm, your hyperfocal distance is a little bit over a meter.  At 35mm the hyperfocal distance is closer to 5 meters.

Why would people NOT shoot at f/22 all the time?
Some do, but most don't.  Shooting at f/22 is far more prone to lens artifacts like chromatic aberration or vignetting and to phenomena like diffraction.  Most modern digital cameras start seeing a decrease in sharpness caused by diffraction somewhere around mid-apertures.  For full frame it might not come in until f/16.  For current APS-C cameras it is more like F/10 or f/11 while the micro four thirds cameras see it begin at closer to f/8.  These are broad generalizations I am making and depending on your output format (large prints or small online viewing) diffraction issues may never matter to you. 

(in some cases people may not want to focus a camera, particularly a wide angle lens for street photography, so they will set the aperture really slow and set the hyperfocal distance and shoot away without a care in the world.  Some old cheap cameras didn't have apertures or a focus ring at all and this is how you took all the images.)

Further muddying this discussion is the fact that lenses are theoretically sharpest at their widest (fastest) aperture. Practically speaking they are often sharpest at their fastest aperture -1 stop; so a f/2.8 lens is probably going to be sharpest at f/4.  This isn't always true, but it is a reasonable guideline to follow until you have time to do your own tests on your equipment.  This may seem confusing because there is so much less depth of field at f/4 than at f/8 or f/22, but this peak sharpness is only in a very thin depth of field.  For example, at f/1.4 a portrait may have a person's eyes in focus (great) while their ears and nose are out of focus (maybe not so great).  Stopping that portrait down to f/22 will make the whole face in focus, but the eyes won't be as perfectly in focus as they were at f/1.4.  If you print that image at 13x19 inches, this could matter a lot. conversely, if you are making a thumbnail image, it won't matter at all. 

There are lots of complexities here that I haven't touched on, but remember everything is a tradeoff and keep doing experiments.
Logged
Fike, Trailpixie, or Marc Shaffer

u2jimbo

  • Newbie
  • *
  • Offline Offline
  • Posts: 22
    • exploring photography
Re: Hyperfocal distance question
« Reply #12 on: August 29, 2012, 12:05:30 am »

fike:
Thank-you for your detailed response to my post!

The details of the test setup are: Canon 16-35 f2.8; focal length: 16mm; f/stop: f8 (thought this would be the aperture that combined the sharpest focus for this lense at this focal length with the greatest DofF before hitting the defraction limit based upon other articles and posts); the calculated hyperfocal distance was: 3.55 ft.; the Near Limit of Focus was 1.78 ft.; the Far Limit of Focus was 6043 ft.; and the Total Depth of Focus was: 6041 ft.;  the CofC was 0.030mm; I focused at 3.55 ft. using the tape measure to validate.


I think that is everything.
Jim
Logged

fike

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 1413
  • Hiker Photographer
    • trailpixie.net
Re: Hyperfocal distance question
« Reply #13 on: August 29, 2012, 08:50:23 am »

I think you should retry your experiment.  It should have worked, and I think the f/8 shot should have appeared sharper than the f/22 shot.  Were the shots handheld or tripod mounted?  Shutter speed? Di you use mirror lockup, by chance?


fike:
Thank-you for your detailed response to my post!

The details of the test setup are: Canon 16-35 f2.8; focal length: 16mm; f/stop: f8 (thought this would be the aperture that combined the sharpest focus for this lense at this focal length with the greatest DofF before hitting the defraction limit based upon other articles and posts); the calculated hyperfocal distance was: 3.55 ft.; the Near Limit of Focus was 1.78 ft.; the Far Limit of Focus was 6043 ft.; and the Total Depth of Focus was: 6041 ft.;  the CofC was 0.030mm; I focused at 3.55 ft. using the tape measure to validate.


I think that is everything.
Jim
Logged
Fike, Trailpixie, or Marc Shaffer

u2jimbo

  • Newbie
  • *
  • Offline Offline
  • Posts: 22
    • exploring photography
Re: Hyperfocal distance question
« Reply #14 on: August 29, 2012, 11:29:12 am »

fike:
Yes, I set the shot up using a tripod and mirror-lockup.  I triggered the exposure with a cable release.  The exposure was 200 th of sec..  ISO: 100.

Your comment that it should have worked gives me hope.  It gets confusing - I have received some opinion (off-line) that shooting at f22 (small apertures) will always outperform the technique I am trying to learn...that defraction is a neglible effect compared to the benefits of substantial DoF.  Reminds me of Ansel Adams' f64 club.  How did he deal with defraction?

I'll try again.
Jim
Logged

fike

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 1413
  • Hiker Photographer
    • trailpixie.net
Re: Hyperfocal distance question
« Reply #15 on: August 29, 2012, 11:36:24 am »

fike:
Yes, I set the shot up using a tripod and mirror-lockup.  I triggered the exposure with a cable release.  The exposure was 200 th of sec..  ISO: 100.

Your comment that it should have worked gives me hope.  It gets confusing - I have received some opinion (off-line) that shooting at f22 (small apertures) will always outperform the technique I am trying to learn...that defraction is a neglible effect compared to the benefits of substantial DoF.  Reminds me of Ansel Adams' f64 club.  How did he deal with defraction?

I'll try again.
Jim

Ansel Adams was shooting large format negatives, so diffraction was not easily measured--also because he was shooting film, the imprecision of grain that made precisely observing these types of phenomena much harder..  Furthermore, the standards for sharpness and available equipment in the 1930s and 40s were lower than today.  Not everything he produced looks tack-sharp when enlarged to very large sizes that are comparable to our enlargements today. If you look at his exhibitions, not too many of the images are enlarged to what we could call very large sizes.  Lots of 18x24, 13x19, and 8x10 stuff in there. A 18x24 made from an 8x10 negative isn't enlarged nearly as much as a 13x19 made from an APS-C sensor.  Lots more to say about Mr. AA and depth of field and the technology of the day, but I am really out of my expertise area when we start talking technical details with large formats.
Logged
Fike, Trailpixie, or Marc Shaffer

Edalongthepacific

  • Full Member
  • ***
  • Offline Offline
  • Posts: 118
Re: Hyperfocal distance question
« Reply #16 on: September 02, 2012, 02:37:05 am »

I think this discussion has been great. I suppose, since digital "film" is cheap (after we buy the computer, monitor, monitor calibration unit, printer, camera, lens, memory and case) we may just as well take a fair number of shots at different f-stops and at different points of focus. Also, doesn't Photoshop allow photos at differing depths of field to be merged? I also do not understand why camera adjustments can't be made on an iPod or laptop using a wireless connection. That way both focus and f-stop could be arranged with a few touches.
Logged
Pages: [1]   Go Up