Is the K factor relevant to me or should I cancel it out?

[Stephen Benskin]

Part 1 or Part 5b - Average Luminance Range, Exposure Placement, and k1. I had to split this part up into constituent parts. It was becoming way too long an installment. Plus, I was having writers block with a transition.

This part is finally getting into the application of the theory as it begins to define the standard exposure model.

 

[RalphLambrecht]

This a lone is worth alone being on APUG. Thank You, Steve!

 

[Stephen Benskin]

Thanks Ralph. It's way more work than I originally figured. I don't know how you were able to write an entire book.

 

[RalphLambrecht]

Thanks Ralph. It's way more work than I originally figured. I don't know how you were able to write an entire book.

The same way you eat an elephant. I'm already working on the next!

 

[BetterSense]

OK, does the K factor adjust the SLOPE of the (base-2 log X) curve or merely the intercept? I'm pretty sure the K factor is an even offset to the EV/LV scale, correct?

In otherwords, if you have a step wedge with each wedge 1 stop apart, and you spotmeter every step, and write the values down, and then increase the K value on the meter, and re-read the step wedge, would every step of the wedge increase in reading by the same amount? Or would different brightnesses be affected differently e.g. EV 1 reading would increase by .1 stops but EV 15 reading would increase by .5 stops?

 

[Stephen Benskin]

OK, does the K factor adjust the SLOPE of the (base-2 log X) curve or merely the intercept? I'm pretty sure the K factor is an even offset to the EV/LV scale, correct?

In otherwords, if you have a step wedge with each wedge 1 stop apart, and you spotmeter every step, and write the values down, and then increase the K value on the meter, and re-read the step wedge, would every step of the wedge increase in reading by the same amount? Or would different brightnesses be affected differently e.g. EV 1 reading would increase by .1 stops but EV 15 reading would increase by .5 stops?

K factor adjusts the placement of the exposure back and forth along the log-H axis. All points move equal distances with changes in K (exposure). You might be thinking about how flare affects exposure proportionally. I've attached an example of what I hope helps to illustrate the distinction.

 

[BetterSense]

So it looks like changing K basically has the same effect as using a different EI.

 

[Stephen Benskin]

Exactly.

 

[RalphLambrecht]

Then allow me to ask (and it's most likely is in your excellent presentation of facts):

What change in EI is required for a change from one C or K value to another?

 

[BetterSense]

I just adjusted my incident meters so that they gave EV 15 (1/125 @ f/16) with 100 speed film in daylight. I considered tweaking it till they gave 1/100th @ f/16, but I decided that I like EV 15 being pinned to sunlight. I have no idea what k-factors these correspond to, since I derive my calibration formula empirically directly from the output of my light sensor, without calculating 'lux' or anything else in between.

 

[RalphLambrecht]

I'm sure that works pretty well in Dallas, Texas.

 

[Stephen Benskin]

This installment covers film speed, exposure placement, and the constant k1. Film speed is a fairly involved subject on its own. I've spent most of the time since the last installment trying to figure out how to present the information as concise as possible with some semblance of a logical order. I hope I've connected most of the dots. If you bear with it, there are a few little known concepts about film speed and exposure. I've referred to some of the concepts from time to time in my posts, but this installment should flesh out the reasoning a little better.

I think I have two more installments to go at best. One is connecting everything to a Zone System perspective, and the other is applying the different values of K and seeing how they affect exposure in addition to applying flare to the exposure equation.

 

[Stephen Benskin]

Defining K, part 5d - Calculating Flare and Exposure

This installment does just that. It shows how to calculate exposure and flare and how to incorporate flare into the exposure equation. It should have been included with the last installment, but 5c was too long and involved without it.

 

[Stephen Benskin]

The attachment is an excerpt from Safety Factors in Camera Exposure by C.N. Nelson. I'm posting it because it helps to support some of the concepts I expressed in part 5C concerning film speed. While none of the ideas are new, they aren't well known. Yet they present a concept of film speed that is almost radically different than most of what is presented in today's popular literature.

Some of the concepts presented are:

- The film speed number can and usually is adjusted through the use of a film speed constant.
- B&W speed point isn't necessarily where the shadow exposure is supposed to fall.
- There is a direct correlation between the fractional gradient method and the current ISO method only under specific developmental conditions.
- Under those conditions, the current ISO method is just a simplified version of the fractional gradient method with a change in the safety factor.
- The fixed density method doesn't produce accurate results with development that does not conform to the ISO parameters.
- When the processing conditions vary from the ISO parameters, an additional component should be introduced to adjust for the inaccuracies.
- Sensitometric exposure and camera exposure aren't the same thing, but the correct interpretation of the sensitometric exposure translates into usable information for the camera exposure.

I promise, a careful reading and some reflection can be quite illuminating. And I will email the full version of the paper to anyone interested.

 

[Mr Bill]

Hi Stephen, thanks for posting. I just took a brief read through "Safety Factors...", noticing the emphasis Nelson puts on the concept of development to an "average gradient", rather than a gamma value. And that this was, perhaps, key to accepting the film speed point based on a fixed density (relative to base+fog).

Changing topics...long time ago, when I was a kid, I used to wonder why Kodak made such a big deal about this idea they had, called "contrast index", as a replacement for gamma, or whatever. I happen to have a mid-1960s paper, titled "Contrast Index", where Nelson is one of the authors. In this paper, they refer to American Standard PH2.5-1960, the film speed standard, and the average gradient as per your excerpt from "Safety Factors...". It turns out that contrast index came about as a proposed improvement to the average gradient method of evaluating development, with broader application. Again, thanks for posting. Your 2-page excerpt has solved the mystery, for me, of why contrast index came about.

 

[Stephen Benskin]

Defining K - Comparing the Standard Model and Zone System Model of Exposure.

This installment is slightly out of sequence. I'm having some difficulty organizing the installment on comparing the two values of K, 1.16 and 1.30. While this installment will lack the benefit of a few points of theory from the missing installment, it shouldn't negatively affect it.

As many people use the Zone System and interpret exposure through it, I thought I'd attempt to combine Zone System concepts with the sensitometric concepts used in this series.

 

[Bill Burk]

Revisiting this discussion of K because some recent discussions are starting to shed light on the topic.

Some ideas I want to expand on: Did the first ASA speeds fit 18% gray cards better? Did the instructions for using the gray card have to change when ASA speeds were doubled? Are you really only supposed to use 18% gray card readings directly for flat copywork? Should you use the lumidisc (flat receptor) when trying to correlate incident readings with reflected readings? Is it illogical to use the lumisphere (dome receptor) when shooting flat copy (including gray cards)?

 

[Bill Burk]

Diapositivo pointed out a link to a discussion of K factor which adds a table of the different implied reflectance values which different K and C values correlate to.

http://dpanswers.com/content/tech_kfactor.php

 

[Bill Burk]

shutterboy brought up a link to Sekonic video showing expected difference between incident and 18% gray card readings.

http://www.sekonic.com/whatisyoursp...dts-software-profile-your-digital-camera.aspx.

and shutterboy also wrote Sekonic who wrote back...

Okay, I actually wrote to Sekonic and here is the response I got.

Dear Mr. or Ms. Subhro Kar,

Thank you for your inquiry.
I am Minoru Oda(Mr) from Sales Operation Sec.
of light meter in Sekonic Corporation in Japan.

As you already might know, exposure value is determined by calibration constants.
Our light meter has K=12.5, C=340(Lumisphere), C=250(Lumidisc).
Therefore, the incident reading and the reflected reading(with 18% gray) are different under the
condition of point light source and in dark room. However, the reason you mentioned is a little
wrong.

These calibration constants means that if you measure 16% gray with spot, the reading might be
same as incident reading of lumidisc.
However, in the measurement with lumisphere, the lumisphere receives more light than Lumidisc
does.
It is no meaning to compare the readings between Lumisphere (incident) and Spot(reflected) because
they are different system and receive different light.
If forced to say, if you measure 11.5% gray with spot, the reading might be same as incident reading
of lumisphere under the condition of point light source and in dark room with no reflection from
walls.
However, actually in the lighting in nature or studio, "Point light source" are impossible.
So, this is only for theory and formula.
We have traditionally　continued to use C=340(Lumisphere) from long time ago, because we have
recommended it and our customer has appreciated its performance in nature or studio.

Does it answer your inquiry?
If you have any question, please kindly let us know.

Sincerely yours,

Minoru Oda
Sales Operation Sec.

I particular like the information in the reply that you can match 11.5% with a point light source in a dark room.

It's funny to me because that is what I did.

 

[RalphLambrecht]

Sometimes I feel like I'm driving in the snow with all this system calibration. You turn right when you want to go left. Turn too far and if you're lucky you hit the hillside and wait for some helpful citizen to come along with a tow line to pull your radiator off. (OK, I was the helpful citizen. How was I supposed to know the brace behind the Fiero's bumper held the radiator. The driver was really nice considering).

I've heard that instead of being calibrated to 18% gray, most meter manufacturers include a "K-factor" which is approximately a 1/3 stop deviation away from 18% gray.

In the specifications for my meter, the Calibration Constant is 12.5 for reflected light. I have the opportunity to change this to any number I want.

Does the K factor improve the accuracy of my meter? Does it compensate for flare? Should I try to determine my own K factor?

Knowing you K faxtor will help you to use your meter to read in lux or lumen if you care to do so.For determining exposure,you can most likely live without knowing the K factor of your meter.

 

[Ricardo Miranda]

Revisiting this discussion of K because some recent discussions are starting to shed light on the topic.

Some ideas I want to expand on: Did the first ASA speeds fit 18% gray cards better? Did the instructions for using the gray card have to change when ASA speeds were doubled? Are you really only supposed to use 18% gray card readings directly for flat copywork? Should you use the lumidisc (flat receptor) when trying to correlate incident readings with reflected readings? Is it illogical to use the lumisphere (dome receptor) when shooting flat copy (including gray cards)?

Bill, I am just going to tell you this: too much fiddling and too much obsession with technical details don't lead to better photographs.
You read what you want to read from that.
But, I think I'm better off with just using a reflective meter, be it inbuilt in my 35mm cameras or a Leningrad 8 and not to worry so much.
I don't remember reading that Albert Eisenstadt, Andre Kertesz, Jane Bown or Vivian Maier were ever obsessed with technical details.
You can't fault aesthetically their photographs, nor even technically.
Lastly, don't try to have an heart attack. Life is too short for that.

 

[Kevin Caulfield]

Knowing you K faxtor will help you to use your meter to read in lux or lumen if you care to do so.For determining exposure,you can most likely live without knowing the K factor of your meter.

That sums it up for me. All we need to know.

 

[Bill Burk]

Bill, I am just going to tell you this: too much fiddling and too much obsession with technical details don't lead to better photographs.
You read what you want to read from that.
But, I think I'm better off with just using a reflective meter, be it inbuilt in my 35mm cameras or a Leningrad 8 and not to worry so much.
I don't remember reading that Albert Eisenstadt, Andre Kertesz, Jane Bown or Vivian Maier were ever obsessed with technical details.
You can't fault aesthetically their photographs, nor even technically.
Lastly, don't try to have an heart attack. Life is too short for that.

I'm doing this with Ansel Adams' question about K in mind.

He was concerned that K included an adjustment for "distribution of luminance levels in the scene*"

And I am interested in following his logical question, which is why I started this thread.

Since a certain part of K was chosen to adjust for average scene luminance level distribution (a typical distribution might include too much sky for instance), why is that specific adjustment within K included in a spotmeter where the luminance level is taken from a spot?

It seems likely that this part of the value of K was chosen by tradition. Now my meter, Sekonic L-758DR, specifies K same as nominal, K = 12.5. This makes me think I will find a published value for "distribution of luminance levels in the scene".

*ANSI PH3.49-1971 Appendix C, Exposure Parameters, C1. Relationship Between Sensitometric Exposure and Camera Exposure

 

[Bill Burk]

Also the dpanswers article pointed a mention of K in Ansel Adams "The Negative" on page 66.

On that page, Ansel Adams suggests you eliminate K when using a Weston meter when you use the Exposure Formula instead of the calculator dial.

Makes me think there are two different ways of using the Weston meter (not talking about Invercone right now)... Aiming generally at the subject and taking an average reading - then the average "distribution of luminance levels" matters. Or aiming at a gray card, where there is a single luminance to be evaluated.

I also like to keep in mind that deliberately changing a metering habit, such as pointing downward to eliminate sky... is likely to impact the "distribution of luminance levels" to the point where a different "calibration" is required.

 

[BrianShaw]

Your just learning that, Bill????