Re: It's only an explanatory graphic ...

Press Correspondent wrote:

Jeff wrote:

AnthonyL wrote:

Jeff wrote:

AnthonyL wrote:

and which part of your explanation did I say I did not understand?

Well, if you're comfortable with this explanation, then the triangle is obvious. Get yourself a piece of triangular graph paper -- label each corner with ISO, f-ratio, and shutter speed, and label each grid line moving away from the corner with a one-stop change in the corresponding value.

Constant exposure is maintained by moving along any grid line parallel to a side of the triangle.

That's the exposure triangle.

Did that answer your question?

The problem with that model is that it is very 2 dimensional. You can move along the ISO to Av points, or the Av to Tv points, but it doesn't show readily the three elements interacting. My model shows all three elements being pushed pulled, or if you hold the length of one element fixed, the other two must change.

With Auto ISO now becoming more prevalent a better visual model I think would be more helpful.

Let S be ISO, t be shutter duration, N be f-Number. We'll call image brightness B where

B = log2(N^2/t) - log2(S)

The first term is the definition of exposure value, the second describes the effect of ISO on image brightness. What we want is B = log2(L/k) where L is scene luminance and a meter calibration constant. Manufacturers using slightly different values of k. Taking a logarithm,

B = 2*log2(N) - log2(t) - log2(S)

This how these *four* elements interact, and why a triangular graph paper labeled by powers of two establishes an 'exposure triangle' for a given level of incident light.

If you're trying to incorporate the third dimension of incident light, then you really do have a three dimensional object and can't express it as a triangle in two dimensions. You need an 'exposure pyramid'.

You are trying to project a 3-dimensional formila to a plane. This is only possible as an illustration. Also the use of logarithms represents nothing, but the type of scale you have chosen. It is a complication that makes understanding dependent on the knowledge of logarithms tgat is not necessary.

I think he has to in order to force this somewhat silly model to fit. Jeff knows it. Most of those expressions are those standardized definitions of exposure value needed to set it up. He's trying to show the need to incorporate EV-L or scene luminance. That's not going to be a triangle anymore, and if you continue, Jeff does a fairly good job of describing an accurate model I think. It's all not very useful, but does answer the question, don't you think?

Anyway, it looks just to be a fun attempt at meeting Anthony's request for a model. I certainly wouldn't want to take the time to close those expressions to an equation just to lose the logarithmic nature.

-- hide signature --

Cheers, Craig

Follow me on Twitter @craighardingsr : Equipment in Profile - f/22 Club Member

I reserve the right to make mistakes in reasoning and logic as well as to change my mind anytime I wish. I also ask forbearance with respect to my typos. Please take a look at my gallery here at DPR.