What is wrong with Depth of Field and Hyperfocal Distance? While both concepts are based on sound mathematical principals, they are fundamentally flawed by an assumption and a...

selmslie Loc: Fernandina Beach, FL, USA

While both concepts are based on sound mathematical principals, they are fundamentally flawed by an assumption and a misunderstanding:

1. That Depth of Field (DoF) is a range of distances from the camera over which the image is equally sharp.
2. That everyone understands and agrees on the definition of the Circle of Confusion (CoC).

The first assumption is close but there is a difference between being equally sharp or acceptably sharp.

There will be only one distance at which an image is perfectly sharp. Consider a point source of light such as a star. If you set the focus at infinity, the star should project a point of light on your sensor that is smaller than a single pixel. The same would apply if the point source of light were much closer to the camera, for example 20 feet from the camera or much closer. When in focus, the point on your sensor could still be smaller than a pixel. If it were slightly out of focus it might be as large as a whole pixel and your camera would not know the difference.

How big is a pixel? A full frame 24x36 mm sensor that produces an image of 4000x6000 pixels or 24 MP – 24 mm divided by 4000 (or 36 mm divided by 6000) is 0.006 mm. Disregard for the moment that this may be either a red, blue or green pixel and that your lens might not be capable of projecting a point source that precisely. A CoC of 0.006 mm is much smaller than you normally need. If you were to aim at a star with a 50 mm lens at f/11, you might get the point of light to stay within a single pixel if your focus was anywhere from about 125 feet to infinity.

A common assumption for the size of the CoC is about 0.03 mm for a full frame sensor or film with the same format. That would mean that a point source of light could actually spread out to cover a diameter of 5 pixels and still be acceptably sharp. The star that is in focus at infinity will cover less than 5 pixels with a 50 mm lens at f/11 if you focus at any distance longer than about 25 feet, the hyperfocal distance for a 0.03 mm CoC.

Why 0.03 mm? Someone with normal eyesight can resolve a detail that is about 2 minutes of arc or 30 cycles of black/white stripes per degree. That works out to about 0.0291 mm for a 24x36 mm film or digital format.

The common assumption is that you will be making a print that is about 8x12 inches using the entire image and viewing it from about 10 inches with normal eyesight – or 8x12 feet and viewing it from 10 feet or any other ratio that keeps the print dimension proportional to the viewing distance. But if you crop the original image, view it from a distance that is different, with eyesight that is different from normal, use a loupe or magnifying glass, then 0.03 mm may no longer be appropriate.

So you need to know something about how your image will be viewed before deciding on the size of the CoC. If you are going to display it within the parameters described in the previous paragraph the 0.03 mm will work. If you will be displaying the whole image on a smart phone, tablet or printing it smaller than 8x12 inches then 0.03 mm will be more than adequate. But if you plan to enlarge it, crop it, pixel peep, print it large and look at it closer than normal viewing distance, you will need to use a smaller CoC.

imagemeister Loc: mid east Florida

selmslie wrote:

While both concepts are based on sound mathematical principals, they are fundamentally flawed by an assumption and a misunderstanding:

1. That Depth of Field (DoF) is a range of distances from the camera over which the image is equally sharp.
2. That everyone understands and agrees on the definition of the Circle of Confusion (CoC).

The first assumption is close but there is a difference between being equally sharp or acceptably sharp.

There will be only one distance at which an image is perfectly sharp. Consider a point source of light such as a star. If you set the focus at infinity, the star should project a point of light on your sensor that is smaller than a single pixel. The same would apply if the point source of light were much closer to the camera, for example 20 feet from the camera or much closer. When in focus, the point on your sensor could still be smaller than a pixel. If it were slightly out of focus it might be as large as a whole pixel and your camera would not know the difference.

How big is a pixel? A full frame 24x36 mm sensor that produces an image of 4000x6000 pixels or 24 MP – 24 mm divided by 4000 (or 36 mm divided by 6000) is 0.006 mm. Disregard for the moment that this may be either a red, blue or green pixel and that your lens might not be capable of projecting a point source that precisely. A CoC of 0.006 mm is much smaller than you normally need. If you were to aim at a star with a 50 mm lens at f/11, you might get the point of light to stay within a single pixel if your focus was anywhere from about 125 feet to infinity.

A common assumption for the size of the CoC is about 0.03 mm for a full frame sensor or film with the same format. That would mean that a point source of light could actually spread out to cover a diameter of 5 pixels and still be acceptably sharp. The star that is in focus at infinity will cover less than 5 pixels with a 50 mm lens at f/11 if you focus at any distance longer than about 25 feet, the hyperfocal distance for a 0.03 mm CoC.

Why 0.03 mm? Someone with normal eyesight can resolve a detail that is about 2 minutes of arc or 30 cycles of black/white stripes per degree. That works out to about 0.0291 mm for a 24x36 mm film or digital format.

The common assumption is that you will be making a print that is about 8x12 inches using the entire image and viewing it from about 10 inches with normal eyesight – or 8x12 feet and viewing it from 10 feet or any other ratio that keeps the print dimension proportional to the viewing distance. But if you crop the original image, view it from a distance that is different, with eyesight that is different from normal, use a loupe or magnifying glass, then 0.03 mm may no longer be appropriate.

So you need to know something about how your image will be viewed before deciding on the size of the CoC. If you are going to display it within the parameters described in the previous paragraph the 0.03 mm will work. If you will be displaying the whole image on a smart phone, tablet or printing it smaller than 8x12 inches then 0.03 mm will be more than adequate. But if you plan to enlarge it, crop it, pixel peep, print it large and look at it closer than normal viewing distance, you will need to use a smaller CoC.

While both concepts are based on sound mathematica... (show quote)

That is why DOF and everything pertaining to it is such a nebulous subject !

..

speters Loc: Grangeville/Idaho

selmslie wrote:

While both concepts are based on sound mathematical principals, they are fundamentally flawed by an assumption and a misunderstanding:

1. That Depth of Field (DoF) is a range of distances from the camera over which the image is equally sharp.
2. That everyone understands and agrees on the definition of the Circle of Confusion (CoC).

The first assumption is close but there is a difference between being equally sharp or acceptably sharp.

There will be only one distance at which an image is perfectly sharp. Consider a point source of light such as a star. If you set the focus at infinity, the star should project a point of light on your sensor that is smaller than a single pixel. The same would apply if the point source of light were much closer to the camera, for example 20 feet from the camera or much closer. When in focus, the point on your sensor could still be smaller than a pixel. If it were slightly out of focus it might be as large as a whole pixel and your camera would not know the difference.

How big is a pixel? A full frame 24x36 mm sensor that produces an image of 4000x6000 pixels or 24 MP – 24 mm divided by 4000 (or 36 mm divided by 6000) is 0.006 mm. Disregard for the moment that this may be either a red, blue or green pixel and that your lens might not be capable of projecting a point source that precisely. A CoC of 0.006 mm is much smaller than you normally need. If you were to aim at a star with a 50 mm lens at f/11, you might get the point of light to stay within a single pixel if your focus was anywhere from about 125 feet to infinity.

A common assumption for the size of the CoC is about 0.03 mm for a full frame sensor or film with the same format. That would mean that a point source of light could actually spread out to cover a diameter of 5 pixels and still be acceptably sharp. The star that is in focus at infinity will cover less than 5 pixels with a 50 mm lens at f/11 if you focus at any distance longer than about 25 feet, the hyperfocal distance for a 0.03 mm CoC.

Why 0.03 mm? Someone with normal eyesight can resolve a detail that is about 2 minutes of arc or 30 cycles of black/white stripes per degree. That works out to about 0.0291 mm for a 24x36 mm film or digital format.

The common assumption is that you will be making a print that is about 8x12 inches using the entire image and viewing it from about 10 inches with normal eyesight – or 8x12 feet and viewing it from 10 feet or any other ratio that keeps the print dimension proportional to the viewing distance. But if you crop the original image, view it from a distance that is different, with eyesight that is different from normal, use a loupe or magnifying glass, then 0.03 mm may no longer be appropriate.

So you need to know something about how your image will be viewed before deciding on the size of the CoC. If you are going to display it within the parameters described in the previous paragraph the 0.03 mm will work. If you will be displaying the whole image on a smart phone, tablet or printing it smaller than 8x12 inches then 0.03 mm will be more than adequate. But if you plan to enlarge it, crop it, pixel peep, print it large and look at it closer than normal viewing distance, you will need to use a smaller CoC.

While both concepts are based on sound mathematica... (show quote)

I don't see there being any misunderstanding or wrong assumptions about DOF, it says what it is very clearly with not much or any other way of interpretation!

ballsafire Loc: Lafayette, Louisiana

selmslie wrote:

While both concepts are based on sound mathematical principals, they are fundamentally flawed by an assumption and a misunderstanding:

1. That Depth of Field (DoF) is a range of distances from the camera over which the image is equally sharp.
2. That everyone understands and agrees on the definition of the Circle of Confusion (CoC).

The first assumption is close but there is a difference between being equally sharp or acceptably sharp.

There will be only one distance at which an image is perfectly sharp. Consider a point source of light such as a star. If you set the focus at infinity, the star should project a point of light on your sensor that is smaller than a single pixel. The same would apply if the point source of light were much closer to the camera, for example 20 feet from the camera or much closer. When in focus, the point on your sensor could still be smaller than a pixel. If it were slightly out of focus it might be as large as a whole pixel and your camera would not know the difference.

How big is a pixel? A full frame 24x36 mm sensor that produces an image of 4000x6000 pixels or 24 MP – 24 mm divided by 4000 (or 36 mm divided by 6000) is 0.006 mm. Disregard for the moment that this may be either a red, blue or green pixel and that your lens might not be capable of projecting a point source that precisely. A CoC of 0.006 mm is much smaller than you normally need. If you were to aim at a star with a 50 mm lens at f/11, you might get the point of light to stay within a single pixel if your focus was anywhere from about 125 feet to infinity.

A common assumption for the size of the CoC is about 0.03 mm for a full frame sensor or film with the same format. That would mean that a point source of light could actually spread out to cover a diameter of 5 pixels and still be acceptably sharp. The star that is in focus at infinity will cover less than 5 pixels with a 50 mm lens at f/11 if you focus at any distance longer than about 25 feet, the hyperfocal distance for a 0.03 mm CoC.

Why 0.03 mm? Someone with normal eyesight can resolve a detail that is about 2 minutes of arc or 30 cycles of black/white stripes per degree. That works out to about 0.0291 mm for a 24x36 mm film or digital format.

The common assumption is that you will be making a print that is about 8x12 inches using the entire image and viewing it from about 10 inches with normal eyesight – or 8x12 feet and viewing it from 10 feet or any other ratio that keeps the print dimension proportional to the viewing distance. But if you crop the original image, view it from a distance that is different, with eyesight that is different from normal, use a loupe or magnifying glass, then 0.03 mm may no longer be appropriate.

So you need to know something about how your image will be viewed before deciding on the size of the CoC. If you are going to display it within the parameters described in the previous paragraph the 0.03 mm will work. If you will be displaying the whole image on a smart phone, tablet or printing it smaller than 8x12 inches then 0.03 mm will be more than adequate. But if you plan to enlarge it, crop it, pixel peep, print it large and look at it closer than normal viewing distance, you will need to use a smaller CoC.

While both concepts are based on sound mathematica... (show quote)

Why do you insist in playing this game of mental gymnastics is beyond my poor powers of understanding!!

imagemeister Loc: mid east Florida

speters wrote:

I don't see there being any misunderstanding or wrong assumptions about DOF, it says what it is very clearly with not much or any other way of interpretation!

LOL

selmslie Loc: Fernandina Beach, FL, USA

ballsafire wrote:

Why do you insist in playing this game of mental gymnastics is beyond my poor powers of understanding!!

One of the nice things about photography is that you don't need to understand the science behind it to create good photographs.

BebuLamar

Semile! I have a suggestion to use the size of the pixel as the COC.

selmslie Loc: Fernandina Beach, FL, USA

BebuLamar wrote:

Semile! I have a suggestion to use the size of the pixel as the COC.

As I explained above, that's much smaller than necessary. You would never look at an image that closely.

imagemeister Loc: mid east Florida

selmslie wrote:

As I explained above, that's much smaller than necessary. You would never look at an image that closely.

But, some try to ....

..

Gene51 Loc: Yonkers, NY, now in LSD (LowerSlowerDelaware)

selmslie wrote:

While both concepts are based on sound mathematical principals, they are fundamentally flawed by an assumption and a misunderstanding:

1. That Depth of Field (DoF) is a range of distances from the camera over which the image is equally sharp.
2. That everyone understands and agrees on the definition of the Circle of Confusion (CoC).

The first assumption is close but there is a difference between being equally sharp or acceptably sharp.

There will be only one distance at which an image is perfectly sharp. Consider a point source of light such as a star. If you set the focus at infinity, the star should project a point of light on your sensor that is smaller than a single pixel. The same would apply if the point source of light were much closer to the camera, for example 20 feet from the camera or much closer. When in focus, the point on your sensor could still be smaller than a pixel. If it were slightly out of focus it might be as large as a whole pixel and your camera would not know the difference.

How big is a pixel? A full frame 24x36 mm sensor that produces an image of 4000x6000 pixels or 24 MP – 24 mm divided by 4000 (or 36 mm divided by 6000) is 0.006 mm. Disregard for the moment that this may be either a red, blue or green pixel and that your lens might not be capable of projecting a point source that precisely. A CoC of 0.006 mm is much smaller than you normally need. If you were to aim at a star with a 50 mm lens at f/11, you might get the point of light to stay within a single pixel if your focus was anywhere from about 125 feet to infinity.

A common assumption for the size of the CoC is about 0.03 mm for a full frame sensor or film with the same format. That would mean that a point source of light could actually spread out to cover a diameter of 5 pixels and still be acceptably sharp. The star that is in focus at infinity will cover less than 5 pixels with a 50 mm lens at f/11 if you focus at any distance longer than about 25 feet, the hyperfocal distance for a 0.03 mm CoC.

Why 0.03 mm? Someone with normal eyesight can resolve a detail that is about 2 minutes of arc or 30 cycles of black/white stripes per degree. That works out to about 0.0291 mm for a 24x36 mm film or digital format.

The common assumption is that you will be making a print that is about 8x12 inches using the entire image and viewing it from about 10 inches with normal eyesight – or 8x12 feet and viewing it from 10 feet or any other ratio that keeps the print dimension proportional to the viewing distance. But if you crop the original image, view it from a distance that is different, with eyesight that is different from normal, use a loupe or magnifying glass, then 0.03 mm may no longer be appropriate.

So you need to know something about how your image will be viewed before deciding on the size of the CoC. If you are going to display it within the parameters described in the previous paragraph the 0.03 mm will work. If you will be displaying the whole image on a smart phone, tablet or printing it smaller than 8x12 inches then 0.03 mm will be more than adequate. But if you plan to enlarge it, crop it, pixel peep, print it large and look at it closer than normal viewing distance, you will need to use a smaller CoC.

While both concepts are based on sound mathematica... (show quote)

Great explanation, but the problem is not in the concept of Depth of Field or Hyperfocal Distance - the problem is that people interpret the information in their own way, and have inflated expectations that if the chart says something will be in focus from 4ft to 7ft they expect it to be in PERFECT, TACK SHARP FOCUS throughout that distance range. They reject the notion (or just don't understand) that there is a difference of "acceptable" focus and "peak" focus, and that it gradually falls off from the plane of peak focus.

Another popular misconception is the distribution field of focus in front of and behind. You will hear and read that if you focus 1/3 into your landscape scene, everything from the focal point to 1/3 the depth of fieldin front will be in focus, and 2/3 of the field behind. This is an ill-informed rule of thumb, as the distribution of focus varies widely with focal length, subject to camera distance and aperture. A 55mm lens can have an asymmetrical field of 4% in front and 96% behind at F/16 and 30 ft focus distance, and it can also have a nearly equal distribution field of sharpness at F2.8 and 4 ft subject to camera distance.

selmslie Loc: Fernandina Beach, FL, USA

Gene51 wrote:

Great explanation, but the problem is not in the c... (show quote)

Yes, your first statement is the same as my first point. There is only one plane of sharpest focus and everything becomes progressively less sharp as you move away from it.

Anyone can disprove the 1/3-2/3 myth by playing with a DoF calculator. It only happens for a a subject distance equal to 1/3 of the hyperfocal distance. At any other subject distance the 1/3-2/3 split does not happen.

repleo Loc: Boston

selmslie wrote:

1. That Depth of Field (DoF) is a range of distances from the camera over which the image is equally sharp.

I have only ever heard of DoF as the range of 'acceptable focus' - never heard of it as 'equally sharp'.

selmslie Loc: Fernandina Beach, FL, USA

repleo wrote:

I have only ever heard of DoF as the range of 'acceptable focus' - never heard of it as 'equally sharp'.

That’s fortunate for you.

Not everyone understands this and it’s why so many people rely on hyperfocal distance for landscape photography. They will be disappointed when they try to make large prints.

10MPlayer Loc: California

ballsafire wrote:

Why do you insist in playing this game of mental gymnastics is beyond my poor powers of understanding!!

I among many others find it informative and appreciate the time he took to do it. If it's not your idea of a fun read then you don't have to bother. <snark> Science is hard </snark>

E.L.. Shapiro Loc: Ottawa, Ontario Canada

I usually don't like to tell folks to "Google" things or just supply a link if I feel I can supply an logical and practical answer. Most people know to Google on their own In this cases, however, I'm gonna recommend a book entitled "Focusing the View Camera". Although it contains quite a portion on view camera operation, this author included what I consider a "doctoral thesis" on depth of field. It's all there; detailed definitions of DOP, Hyper-focal Distances and Circles of Confusion. It is packed with every pertinent mathematical formula, diagram, depth of field chart that one could imagine and more. It's a master class in optics in 102 pages plus a pocket envelope with additional charts.

The hyper-focal charts are cross-referenced as to aperture and focal length.

Another accompanying book by the same author is "The Ins and Out of Focus"- even more detailed DOP information.

I hope they are still in publication.

Author:
Harold M. Merklinger
P.O. Box 494
Dartmouth, Nova Scotia Canada B2Y 3Y8

If you want to fully understand the science, this is the guy!