I used the example with Ross only to react to the sentament that Davis is creating myth or misinformation about the incident meter. The statement he makes is true----- using a full sunlight incident reading to determine exposure can result in shadows that are horribly underexposed.
I chimed in, but not as one who uses an incident meter with any regularity at all as I am completely in the "reflective" camp------metering the intensity of the reflected light that will actually be striking the film is far more intuitive to me and more important to my way of doing things. But in the BTZS way of doing things, it is obvious that the incident meter can be used with that system quite successfully. Any answer I gave toward those questions would be totally biased toward reflective readings with a spot meter.
So, the concept of "place" and "fall" is entirely different, to my mind, between the two methods, I place a luminance value, you would place a light intensity. I know exactly what to expect from that value in my negative and how it should render on the paper, the same idea with the "fall" of the highlight. I can see in my mind a print value for the "placed" luminance and the luminance of the "fall", and knowing the luminance of all the values in between helps to complete the visualization. For me, I can't predict or visualize the same for those luminances when only "placing" the light intensity of the shadow or just knowing where the intensity of the sunlight "falls".
As to question #2, again, relative to the spot meter, I would compromise next to nothing-----within the limits------of my tested system. Any luminances that "fall" below the "placed" value, are entirely precictable and thus ok with me, and any luminances that "fall" higher on the scale after the planned development scheme that favors the textured high value are also predictable. As far as other tools, dodge and burn of course (always, no negative is that perfect! when it comes to meeting the visualization, IMO), perhaps intensification of the negative, perhaps pre-exposure of the negative, etc......
Bottom line, I feel I can achieve the most optimum negative to meet the desired print by reflective metering than with incident, but that is just my feelings on it.
Considering Camera and lens flare, the normal 7 stop range is reduced to 6 stops. So taking reading on any where is "assumed" to be 5 stops by the meter.
With that, reducing a stop for shadows when reading is based on shadow(meter placement is critical) and opening one stop for highlight with bring it to an effective six stops.
Roll film(Negative): I personally have good shadow detail.
Again, testing and placement of meter are essential and critical.
IMO Davis is using an assumption and a special case to make his point.
Originally Posted by CPorter
First, Davis is assuming that deep shadow detail is important in most shots, that seems true for you and I'm sure a fair number of people at APUG, but it is far from a given in photography as a whole.
Second Davis is describing a special case measurement, pointing directly at the light source, that measurement is typically only used with a second measurement (as in duplexing) or with modification (as you might with your spot meter placing zone III). Ruling out the normal way to use the incident meter may help win a debate but it doesn't help us understand what's really happening or help us make better pictures.
If instead we allow the normal cases into the argument we easily get good results, no muss, no fuss.
An incident meter used in the follow-the-manufacturer's-directions manner, dome pointed at the camera, meter held at Ross's nose, one reading taken, will in the grand majority of situations reasonably place Ross's face nicely on a print. That is assuming we pose Ross a bit in your scene to avoid mottling the light on his face.
A duplexed incident meter reading using the readings from say frames a&b in your example would provide plenty of info to decide on how to adjust contrast and the average of the two readings would give you a camera setting that should protect both shadows and highlights quite nicely.
I have a problem with the affirmation I underlined.
Originally Posted by Mahler_one
The nature of incident reading is that it leaves aside the reflectance of the subject. In the same light, with the same incident reading, a middle grey subject will be rendered middle grey, a bright subject will be rendered bright, and a dark subject will be rendered dark.
This happens "automatically" by the sheer fact that the bright object reflects more light than the middle grey objects and thus renders brighter on the slide than the middle grey object, given the same exposure. It works very well until the "whiteness" or "blackness" of the subject is within the dynamic range of the film.
With reflective reading the photographer must be aware of how different the subject is from middle grey and adjust the exposure based on that, thus "placing" the subject in the film curve. With incident metering the photographer doesn't need to do the thinking. The subject will be "placed" in the film curve by its own reflective power "automagically".
Provided it is uniformly lit by a source. Since, blackest black is approximately 4.5 stops to whitest white. So incident metering will render blacks and whites satisfactorily.
Originally Posted by Diapositivo
It needs some adjustment when there are shadows and highlights.
Mark, I'm only responding to the Davis quote you referenced earlier. His statement is accurate with regard to a single incident measurement in full sun and the corresponding poor exposure to be had in the shadows----there's no myth involved with that statement, it's a fact. I'm making no other challenge.
Originally Posted by markbarendt
I believe dome in gives the local reading of light falling on the subject when local source of illuminate is considered.
In real-world(landspace) the shadow may have global source. So, reading dome out may be appropriate.
Invercone in: flat, single dimension subjects such as copying, overhead direct shots.
Invercone out: directional 3 dimensional light.
A landscape reading with the invercone recessed will be erroneous.