March 05, 2015, 11:20:07 PM

Show Posts

This section allows you to view all posts made by this member. Note that you can only see posts made in areas you currently have access to.

Messages - dougkerr

Pages: [1] 2 3 ... 6
Canon General / Re: DOF and Sensor Size
« on: July 01, 2014, 07:04:31 PM »
The question really has to be, "What is the effect of sensor size on DoF all other factors being equal."

And we have to decide what "all other factors being equal" is to mean.

A reasonable set of definitions starts with:

• The focal lengths on the two cameras will give comparable field of view.

• The two cameras will be focused at the same distance.

But there is one further wrinkle. Depth of field isn't an actual optical phenomenon. It s how we choose to look at an optical phenomenon. The actual question is, "considering the setup, over what range of object distances will the image not be blurred by more than a certain amount". And we have to decode what that certain amount is.

We do that, numerically, by placing a limit on the diameter of the blur figure (circle of confusion) that is created on the focal plan by a point on the subject when the subject is not in perfect focus. I call that the circle of confusion diameter limit (COCDL). It often (but misleadingly) called the "circle of confusion". (That's like calling the largest tire size you care will accept the "tire".)

There are two approached we might use to pick a COCDL. We might pick the COCDL that is essentially the diameter of a blur figure that the eye can just recognize as blurred. This of course requires us to assume that the image is "printed" at a certain size and viewed from a certain distance.

Often this might lead us to choose a COCDL of perhaps 1/1400 of the diagonal size of the sensor.

Or we might say that the COCDL should be chosen as the diameter of the blur circle that will just degrade the resultion of the camera from its resolution for an in-focus subject.

This might, for example, cause us to choose a COCDL or perhaps twice the sensel pitch of the sensor.

In any case, as we visualize moving from one sensor size to another, we need to decide how we will change our COCDL to follow.

Now, having made all those preparations, we are prepared to calculate the "smaller sensor" and "larger sensor" DoF values and compare them.

But for either approach to choosing our COCDL, and if we use the"degradation of resolution" approach to choosing our COCDL there is no dramatic difference in sensel pitch between the two cameras, then the DoF will be less for the larger-sensor camera. How much less depends on exactly what we do about the COCDL.

Best regards,

Doug Kerr

EOS Bodies / Re: Educate me about AF systems
« on: June 12, 2014, 06:52:09 PM »
So another assumption from me is that the AF points are created through software. But are there electrical components that will prevent Canon from improving the AF system at a later stage with more AF points?
In the "traditional" Canon phase-detection AF system, each "AF point" is handled by an explicit physical "little camera" located in the AF detector assembly (in the "basement" of the camera), each one having two little sensors along with (at least) two little lenses.

More of these cannot be created through firmware changes.

Dual-pixel AF operation is another thing altogether.

Best regards,


EOS Bodies / Re: An EOS Canon Medium Format System [CR1]
« on: September 28, 2013, 01:48:28 PM »
Perhaps a good "subfamily" moniker for a "medium format" EOS-family subfamily would be "EOS-M".

Oh, no—Canon has already squandered that on the "baby EOS" subfamily.

They could call it EOS-B, for "bigger than full size".

Best regards,

Doug Kerr

EOS Bodies / Re: Crop sensors need cropped lenes
« on: August 10, 2013, 06:05:52 PM »

A FF lens being used with an APS-C sensor results in a slower system, lens+sensor. Since the APS-C sensor only captures 40% of the light projected by the lens on the image plane, the combination is slower than if the same lens had been used with a FF sensor. The lens is of course the same and doesn't know what sensor you place behind. But a lens on its own will not produce any pictures. You need a sensor as well. Really, I cannot see the controversy.

Not so.

The effect of a lens on exposure depends on its f-number. For, say, an f/4.0 lens, the luminance on the sensor for some part of the scene of course does not depend on the sensor overall size.

If we have two cameras, one with a larger sensor that the other, both with sensor systems that we operate at an ISO SOS of ISO 400, then for a given scene, the same exposure result will be obtained with the same shutter speed and f-number.

Indeed, the total amount of light captured across the entire sensor is less for a smaller sensor than for a larger one. But the light per unit area (the property to which the sensor responds, for any given ISO sensitivity) will remain the same.

Best regards,



My view would be that with any lens designed to be used on a crop sensor, most of the photons entering the lens will hit the sensor, subject of course to the rectangular sensor reading a round lens. In contrast only 40% of the photons that would have hit the sensor on a full frame body will be used in a crop body (assuming a 1.6 crop) with a FF lens.

I'm getting more convinced that long lenses specifically designed for crop bodies would have a weight benefit. I still can't see Canon or Nikon making them. But maybe someone like Sigma might just do it to give themselves a USP.

What you are describing is called a speed booster, and it is available today for mirror less cameras.

If, as you suggest above, you take, say, a 300 mm f/4 lens and project the otherwise wasted FF-photons onto the smaller APS-C sensor, the smaller sensor would still record the same FOV and the same amount of light as the FF sensor. This is because the aperture still has the same absolute size 300/4=75 mm.

However, by doing this you also decrease the focal length of the lens by a factor 1.6. So, while the absolute size of the aperture stays constant, its relative size increases to about f/2.5.

So, there you have it. The same FOV and the same amount of light captured by APS-C and FF. But, you have to decrease the FL with the crop sensor and keep the absolute aperture the same. This of course increases the relative size of the aperture. Hence the f/4 on FF is equivalent to f/2.5 on APS-C.

Unfortunately, if you want to capture a certain amount of light from a distant object the optics need to be of a certain size. You cannot fool the laws of physics just by cropping the sensor. Otherwise, why would e.g. the astronomers keep building larger and larger telescopes if they could just achieve the same thing by cropping their imaging sensors.

This is why its rather pointless to design APS-C only tele lenses. For WA lenses on the other hand its a different story as you can position the lens elements closer to the sensor due to the smaller mirror. Or lack of a mirror.

First recall that DoF is not a distance that comes solely from optical theory (like the distance to the image for a certain object location); it is a description of the range of object distances over which a certain arbitrarily-adopted criterion of "blurred-ness" is not exceeded, with the camera focused at a certain distance.

Often, for consistency, that criterion is a blur circle whose diameter is a certain fraction of the overall image size (maybe its diagonal size). This relates to a consistent diameter blur circle on an image produced at a consistent size (such as a consistent size print).

That having been said, when we compare the DoF attained for two values of some attribute of the camera (such as sensor size), "all other factors being equal", we must adopt and announce what we mean by  "all other factors being equal".

Here is one set of such that we might adopt:

a.  Focal length such that the field of view is the same in both setups.  (Aha!)

b. Camera focused at the same distance.  (Of course.)

c. Same f-number.  (Of course.)

d. Criterion for "acceptable" blurring the same in terms of blur circle diameter as a fraction of the image size (diagonal will do).

Now, if we are comparing two cameras, with "B" having an image size 2x that of "A":

1.  Under rule a, we must use a lens of twice the focal length in B as in A.

2. Under rule c, our acceptable diameter of blur circle is twice for B as it is for A.

Now, with regard to point 2, that means that we are more tolerant of blur in B than in A. Thus this consideration alone would lead to a greater depth of field for B than A.

This works essentially proportionally to the blur circle diameter criterion, and thus to image size (sensor size).

Remember, depth of field is not an creature of optical theory alone. It is a creature of what amount of blurring we consider "acceptable". If we increase the amount we consider "acceptable", then our focus distance can be more "off" and we still consider the result acceptable.

Now, to point 1. Because of the optical situation involved (and I will not attempt to describe this in detail here), for a greater focal length the "incorrectness" in focus distance to cause a certain diameter blur circle is less. Thus this consideration alone would lead to a lesser depth of field for B than A.

This works essentially proportionally to the square of the focal length, and thus (because of rule a) to the square of the image size (sensor size).

So as we move from A to B, we find that issue 2 gives an increase in the depth of field, and issue 1 gives a larger decrease in the depth of field (because it varies inversely as the square of the image size).

Thus, the overall effect is that (under the rules stated above), for an increase in sensor size we have a net decrease in depth of field.

Quod erat demonstrandum.

Best regards,


Canon General / Re: Clarification....Fine Art
« on: July 06, 2013, 01:33:58 PM »
Traditionally, "fine art(s)" was a term that served to distinguish such arts as sculpture, painting, and musical composition and performance from "practical" arts such as carpentry, furniture making, watchmaking, cuisine, and the practice of law. It often distinguished a division of a university devoted to somewhat that range of disciplines.

The term is perhaps best left to that usage.

If we were to agree what "fine art photography" was, what could we do with that wisdom? Perhaps know in what division of the State Fair we should try and enter our work? How to describe a print that was stolen to the insurance company?

Best regards,


EOS Bodies / Re: EOS 70D a New Benchmark in ISO Performance?
« on: July 01, 2013, 11:14:48 PM »
I'm waiting for the Nattering Nabobs of Negativism Band of Banding-Hunting Brothers to mention that it has obvious pattern banding at +10stops DEC & +1000 fill-light, as confirmed by Professor Professorson, PhD E.E., when pixel-voyeured at 800%...

"Nattering Nabobs of Negativity".

  -Spiro Agnew, written by William Safire

Best regards,

Doug Kerr

In a traditional cinema film camera, what we would call "shutter speed" controls how long the shutter is open for each frame, a different matter from how often the frames occur. It is actually usually expressed in terms of shutter angle in degrees, with 360° meaning a shutter time that is exactly the interframe time.

It is very common to use a 180° shutter angle, which at a frame rate of 24 f/sec is an exposure time of 1/48 sec.

This, combined with the aperture, affects the exposure.

This also effects the degree of visible motion blur, although the impact of this is a bit different in motion picture work than in still photography. Shorter shutter speeds impart a "crisper" view (some would say "more urgent") to the image of a moving object, even though it is "moving" as the eye sees it. Thus the cinematographer may choose a shutter angle partly for aesthetic reasons.

This is a nice discussion of the matter:

Shutter Angle - How to make your DSLR video look like film

We often have a similar opportunity in digital cameras in video mode. At 30 f/sec, a shutter time of 1/60 sec corresponds to a "180°" shutter angle in a film cinema camera.

Best regards,



Two Mark IIIs same settings show different available images on 4gb card

Both Set to large L 5760X3840 but one camera shows 500+ available images on an empty formated 4 gig card and the other shows 350. Thoughts?

Is that Mark III of any particular model?

Same settings for sharpening?

Same JPG quality setting?

Same "ISO" setting?

Both cameras at the same elevation (here in New Mexico you don't do anything without knowing the elevation)?

Best regards,


EOS Bodies - For Stills / Re: Real iso's?
« on: May 29, 2013, 11:00:26 AM »
Just a little background here.

"ISO" is an intialism that stands for an international standards body.

There is no property of photographic film or of a digital imaging chain whose actual designation is "ISO".

The ISO publishes various standards that give ways of stating what we might call the "sensitivity" of a photographic film or of a digital imaging chain.

With regard to film, the basic sensitivity metric, determined as prescribed by the appropriate ISO standard, is called the ISO speed. The name of course comes from the fact that with film of a higher sensitivity, an appropriate exposure (for given scene luminance and given aperture) is given by a shorter exposure - a "faster" exposure process.

To provide continuity with practice in the film realm, the basic metric for the sensitivity of a digital imaging chain is also called the ISO speed.  (It is of course determined in a wholly different way.)

Putting aside some complications, the basic "equation" we use (for example, in an exposure meter or automatic exposure control system) to arrive at a "recommended" photographic exposure (that is, combination of exposure time and aperture) for a given scene luminance, taking into account the sensitivity of the film or digital imaging chain (as its ISO speed) is the same for either medium.

But, especially taking into account the greater sophistication of modern exposure metering and control systems, it was observed that, in the case of digital cameras, this equation typically led to a lesser exposure than was "optimal" - typically about 1/2 stop "short". This gives a less-good noise result than we might actually enjoy.

Camera manufactures could have "tweaked" the operation of their exposure control systems to "take advantage" of this in two ways:

a. Change the "equation" used for automatic exposure control.

b. Keep the equation the same but "rate" the sensor system at a lesser ISO speed than would be determined by the ISO test method.

Had they done (a), the result would have been that the internal automatic exposure system would have produced a different exposure for a given shot than would have been "recommended" by a properly-calibrated free-standing exposure meter. This would have led to complaints about "inaccuracy" of the exposure control system of the camera.

So they did (b) instead.

But of course this led knowledgeable enthusiasts to complain that the "ISO speed ratings" of the digital camera were incorrect.

So the ISO defined a new metric for the sensitivity of a digital camera chain, the ISO standard output sensitivity (ISO SOS). This is a different measure than the ISO speed. In fact, again putting aside some complications, for a given sensor chain, the ISO SOS is about "1/2 stop" less than the ISO speed. (That is, the ISO SOS is about 0.7 times the ISO speed.)

And that "solved" the problem!

Today, the sensitivity of many digital cameras is stated in terms of the ISO SOS, not the ISO speed (and that often is stated in the "specifications", although it is sometimes hard to find).

Now for a second chapter of this story, i must refer to the matter of exposure index. For our purposes we can say that exposure index is "what we tell the exposure meter is the ISO speed of the film of digital camera chain".

Going back to traditional analog exposure meters, we recall that if we want to bump the exposure the meter recommends for some reason (compensate for backlighting, perhaps), we can set the "ISO speed" dial to a lower value than the actual ISO speed of the film or digital sensor. And of course, what we set on that little dial is the exposure index (and, as I just illustrated, it is not always the ISO speed of the film or digital sensor chain).

Now, returning to recent developments in the description of digital camera sensitivity, when the ISO introduced the ISO SOS, it also introduced another new "metric", the ISO Recommended Exposure Index (ISO REI).

This is a metric that works the same way as the ISO speed or ISO SOS (as a parameter of the exposure equation). It is defined as (and I paraphrase) "the exposure index that the camera manufacturer feels will, as an input to the exposure equation, give a 'desirable' exposure result in many cases."

In many case, digital camera manufacturers now indicate that the sensitivity "ratings" of their camera (at its various "ISO" settings) are in terms of ISO REI. That is, there is not any way that these values can be "correct" of "incorrect".

As a practical matter, in most of these cases, the "ISO" ratings are actually intended to be the ISO SOS values (and often turn out to closely conform to that definition), but the manufacturers do not care to commit to that.

A more extensive discussion of this matter is given here:

Best regards,


EOS Bodies - For Stills / Re: F/8 Autofocusing impressions
« on: May 18, 2013, 01:11:30 PM »
... everyone who had Canon tele-converters could not have this option!

Smells like Canon is doing this on purpose ... because what I can't understand is if the 6D can AF down to -3EV, then it should also be able to AF with an f/11 lens (in bright daylight, that is).

The AF aperture limit issue that we normally discuss does not evolve light levels. It is based on optical-geometric considerations.

An AF detector is like a rangefinder, with two sensors looking at the scene  through two "virtual sub apertures" located on opposite edges of the real exit pupil.

To attain a certain accuracy, these sub-apertures must be located a certain distance apart. This distance is like the "baseline length" of a classical rangefinder  (I do not mean "rangefinder camera").

The distance between the two sub apertures is built into the optical system of each AF detector pair. If they are farther apart than the diameter of the exit pupil for the maximum aperture of the lens in place, then no light passes to the detector pairs and AF fails. (So typically the Canon cameras block any attempt when the aperture is not "suitable").

But of course the whole situation is more complex than that, and their are various steps that can be taken to "cheat around" the optical-geometric limitations.

Some of the principles are described here:

Best regards,


Technical Support / Re: Teleconverter and DOF
« on: March 01, 2013, 01:08:21 PM »
When comparing DoF issues, we must be very careful about what the conditions of the two situations are. We may say, "all other things being equal", but what does that mean? Same focal length? Same field of view?

For now. let's compare a situation of the same lens with and without a 2x focal length converter. Let's assume that focus is at the same distance, the focal length setting of the base lens is the same (and thus the focal length of the combination is twice that), and the f-number of the base lens is the same (so the f-number of the combination is twice that). And we will assume a consistent circle of confusion diameter limit (COCDL)

Then, in the configuration with the focal length converter in place, the depth of field will be substantially less than for the configuration with no focal length convertor.

If the setting of the base lens is f =  100 mm and f/3.5, with focus at a distance of 10 m, then, using a COCDL of 0.031 mm, the full depth of field is:

No focal length converter:  10.62 m

2x focal length converter:  4.50 m

But perhaps for our purposes it would be best to think in terms of a constant field of view (a constant focal length). Thus, with the focal length converter in place, we would set the focal length of the basic lens to (for our example) 25 mm. Then the results would be:

No focal length converter:  10.62 m

2x focal length converter:  67.98 m


If we look at out-of-focus blur performance (which is what we are sometimes more interested in than depth of field), we also see interesting results.

Suppose that in both cases we assume focus at 10 m and are interested in the blur on a background object at 50 m. We will always use the base lens at its maximum aperture, f/2.0.

First suppose that with no focal length converter aboard, we set the focal length of the base lens to 50 mm, and we leave it set thus when we add the focal length converter. Then the results are (in terms of the diameter of the circle of confusion (blur figure) produces for a point on the background object:

No focal length converter:  0.10 mm (3.24 times our reference COCDL)

2x focal length converter:  0.20 mm  (6.52 times our reference COCDL)

This time, when we mount the focal length convertor, we reset the focal length of the base lens to 25 mm (to hold the actual focal length and thus the field of view). Now the results are:

No focal length converter:  0.10 mm (3.24 times our reference COCDL)

2x focal length converter:  0.05 mm  (1.62 times our reference COCDL)

Best regards,


PowerShot Cameras / Re: G15 or G1X for my mom???
« on: January 19, 2013, 09:39:48 AM »
My wife has excellent results from her PowerShot SX150 IS, which is very compact. Of course with the very small sensor, the noise performance becomes an issue at higher ISO sensitivities.

A problem for me is that it has no viewfinder of any sort, but the chimp panel is quite nice.

Doug Kerr

EOS Bodies - For Stills / Re: Solar eclipse questions...???
« on: May 19, 2012, 12:05:09 PM »
Here is another recent forum discussion about solar photography:

I'm heading to Albuquerque to be in the center of the annular eclipse path...  the forecasted sky cover is higher than I would like (27%) but I'll take my chances.  I'll be using two stacked ND filters with a combined optical density of 3.9
About 13 stops - often recommended for this purpose.

Best regards,


EOS Bodies - For Stills / Re: Solar eclipse questions...???
« on: May 19, 2012, 12:02:58 PM »
Hi, Don,
One source tells me to use the filter for direct shots of the sun but to remove it during the total eclipse...
Good advice for a total eclipse event. Note however that this time there will be no total eclipse.

I don't know what the best advice is during the "complete annular" phase.

Best regards,

Doug (Just barely outside the path of "full annular" eclipse, in Alamogordo, N.M.)

Pages: [1] 2 3 ... 6