Analysis of T-Numbers
18-200mm versus 18-70mm
Prepared 2006-01-09 (146/47321) by Bill Claff
When I was taking images to compare the 18-70mm with the
18-200mm I noticed that in several instances I was getting different exposures
for the same image depending on the lens.
An F-Number is the ratio of the focal length of the lens to the diameter of the exit pupil.
(The exit pupil is the image of the physical aperture when viewed from the camera end of the lens.)
F-Number assumes 100% of the light entering the lens exits the lens.
But due to the properties of the glass and slight reflections at the glass/air and (to a far lesser extent) glass/glass boundaries this is not true.
T-Number (T is for transmission) is an F-Number that has been corrected for transmittance.
(It is more commonly used in cinematography than photography.)
With a similar number of groups and elements I had expected the 18-70mm (13 Groups of 15 Elements) and the 18-200mm (12 Groups of 16 Elements) to have similar light loss and for transmission not to be a factor.
So I shot a white test target using three lens (adding the 50mm f/1.8 as a standard) at the usual focal lengths at f/8
I analyzed a small center portion of each image to determine how much light reached the sensor.
(Small enough that I believe light falloff is not a factor.)
I adjusted my results on the (wrong but convenient) assumption that the 50mm f/1.8D Nikkor (5 Groups of 6 Elements) has 100% transmittance so it's T-Number is t/8.
Here are the results:
Each zoom lens has a sweet spot and then transmission falls off to each side.
Except for between about 20mm and 35mm the 18-200mm clearly outperforms the 18-70mm in light gathering.
What a pleasant surprise given the complexity of the 18-200mm lens!
BTW, I didn't test any other apertures because I have found in the past that CCD response as a function of aperture is quite linear.
So I would expect the delta between F-Number and T-Number to remain constant for a given lens and focal length.
(Oh yeah, all measurements were at infinity focus!)