Inconsistent results with exposure using a grey card

[Dave_H]

I have an old Minolta Spotmeter F which I would like to calibrate against my Canon EOS-20D so they give identical metering. Today it was a clear sunny day so I set up a digital Gray Kard and took several shots to see how much difference there was between the two meters, but quickly noticed something else which has me a little baffled.

I took the first shot of the grey card (which was placed flat on the lawn outside) from a distance of around 18 inches or so, with a small amount of the grass around it. I had a quick look at the histogram, and then took the next shot slightly closer with the grey card completely filling the frame of view (including a bar code label on the card). The third shot was closer still, just enough so that the label was out of the field of view and the entire grey area filled the frame. Flicking through the 3 shots, the reading on the histogram changed on all 3 images, and the closer I got to the card, the dimmer the image.

Now I always thought that if you set the camera to a fixed exposure and took a meter reading, the final image would be correctly exposed, regardless whether you were 50 feet, 10 feet or 1 foot from the subject. In these tests, the camera was set to manual, daylight colour balance, 1/125 sec at f/16. Image format was RAW. The lens I used was a Canon 100mm f/2.8 macro lens.

I had another look at them in Photoshop, with no processing done to them. The RGB values were:

1st image 197, 198, 200
2nd image 180, 181, 183
3rd image 155, 155, 159

I've sat down tonight and thought about this, and the only explanation I can think of is that focusing has influenced the amount of light reaching the sensor. Even though there wasn't much difference in distance between shots (several inches) possibly that is the reason. When taking the photos, this wasn't apparent as the meter reading in the viewfinder didn't appear to indicate this difference at all.

Can anyone here confirm this, or perhaps offer an explanation as to why this happened? It appears that I may have to place the card much further away from the lens when continuing with my calibration exercise.

Thanks guys,
Dave

 

[Helen B]

In simple terms, effective f-number increases as the lens gets further from the image plane (to focus on closer subjects), so you can take two images at the same marked f-number but different effective f-numbers. One answer is to not focus the lens. Also make sure that the lighting on the grey card is even, with no reflections and no shadows from you or the camera.

For example at 1:1 macro, the effective f-number is twice that at infinity, if the marked f-number stays the same. You can look on it like this:

Marked f-number = lens focal length / diameter of entrance pupil*

Effective f-number = image distance** / diameter of entrance pupil

This formula for effective f-number is simplified - it is not correct, but it is close. Just ask if you want more complexity and accuracy.

* The entrance pupil is the image of the iris, as seen from the front of the lens
** The image distance is the distance from the lens' rear nodal point to the image - think of it as the distance between a very simple single-element lens of the same focal length and the image.

 

[480sparky]

Is your camera set on Auto Exposure Bracketing?

 

[Big Mike]

As you got closer to the grey card, were you blocking more light from hitting it? Thus making it darker?

A grey card is entirely characterized by diffuse reflection, which means it picks up (and reflects) light from all directions.

 

[Helen B]

A grey card is entirely characterized by diffuse reflection, which means it picks up (and reflects) light from all directions.

Yes, but most do not do it evenly: they aren't perfectly diffuse. A Kodak grey card, for example, can easily have a two-stop difference between the diffuse 45/0 reflectivity and the 45/45 in line reflectivity (incident/measured angle).

 

[Dave_H]

You've confirmed my suspicions about lens distance from the film plane Helen. Back in the days of film I used to have a set of bellows for an Olympus OM-1 and recall having to make an allowance for the lens being so much further from the film, but I never thought it would make such a noticeable difference with just a slight adjustment in focusing. I'm surprised I didn't notice any difference on the exposure meter in the camera and I might repeat the tests again next time the sun is out just to check that.

Sparky: Yes, I did suspect auto bracketing initially but it was definitely off. That was the first thing I thought of too...

Mike: I was very careful in placing the card in a position that didn't reflect any glare directly from the sun, and made sure I didn't stand in a position that altered the light on the card. I placed the card on the grass rather than a footpath in case light reflected from the concrete influenced the results.

I tried uploading the images here but didn't have much success so I've put them on my ISP server if you wanted to have a look. They are jpgs reduced in size, only around 90k each. (excuse the dust on the sensor, I've got an Arctic Butterfly on order to deal with it)

http://homepages.slingshot.co.nz/~hfradio/test1.jpg
http://homepages.slingshot.co.nz/~hfradio/test2.jpg
http://homepages.slingshot.co.nz/~hfradio/test3.jpg