conurus

[conurus]

Option 1: everything stays the same as it now is

	24-85	f/3.5-f/5.0
	70-200	f/3.5-f/5.0
	645 Distagon 35	f/3.5
	645 VS 45-90	f/5.0

Contax and Canon has different rounding rules but we could add a minuscle amount to the max aperture to force Canon to round up instead of round down.

Option 2: display f/4.5 instead of f/5.0 at the tele-end

	24-85	f/3.2-f/4.5
	70-200	f/3.5-f/4.5
	645 Distagon 35	f/3.2
	645 VS 45-90	f/4.5

[conurus]

Wow, the results seem to be a landslide so far...

So, you guys are not bothered when the lens marking says f/3.5 but the display says f/3.2, but you are bothered when the lens marking says f/4.5 but the display says f/5.0!?

[frank]

I guess that the f/3.2 side will cause the exposure a little underexposed and f/5.0 overexposed. If my guessing is correct, I would rather have tiny underexposed pictures.

Frank

[Hening]

Which one (for the 24-85) is closest to the truth, i.e. the physical size?

[conurus]

This is a very complicated question and I will try my best to address. My answer is divided into 3 sections for easier reading.

1. f-number of the Zeiss 24-85 according to the optical formula

According to the datasheet of the lens,

http://www.zeiss.com/C12567A8003B8B6F/EmbedTitelIntern/Vario-Sonnar3.5-4.5_24-85mm_e/$File/Vario-Sonnar3.5-4.5_24-85mm_e.pdf

The focal length at the tele-end is 82.4mm and the entrance pupil diameter at the tele-end is 17.6mm. The physical aperture f-number is always the focal length divided by entrance pupil diameter, so that is f/4.68. That is the f-number at the tele-end as dictated by the optical design, as specified in Zeiss' data sheet.

2. Industry practice is to 'cheat' and claim f/4.5 when the max aperture is f/4.76

f-numbers form a geometric progression with a multiplier of square root 2. So, to be mathematically precise, f/5.6 should be f/5.66. The f-number exactly halfway between f/4.0 and f/5.66 is f/4.76. All lens manufacturers claim f/4.5 when the half-stop between f/4 and f/5.66 is meant.

The third stops between f/4.0 and f/5.66 are f/4.49 and f/5.04.

Think of it this way: if you put your camera in 1/3 stop mode, the aperture goes from f/4.0, to f/4.5, f/5.0, f/5.6. In 1/2 stop mode, the aperture goes from f/4.0 to f/4.5, to f/5.6. Something is wrong here, isn't it? How can f/4.5 be both the 1/2 stop smaller than f/4.0 and the 1/3 stop smaller than f/4.0? Mathematically the f-number 1/2 stop smaller than f/4.0 is approximately f/4.8.

If you find this confusing, so do I, but industry practice is to call the lens f/4.5 when f/4.76 is meant.

The f-number of a lens changes slightly when the focusing distance change. Any discussion of a couple of tenths of an f-stop is immaterial and only of academic interest.

3. How Canon does it and how we are going to do it, going forward.

We reviewed a Canon 28-105/3.5-4.5 lens and observed the same: at the tele-end the lens data says the f-number is the half-stop between f/4.0 and f/5.6, so the lens is actually f/4.76 at the tele-end. Canon fiddles with the display and claim a slightly larger aperture than actual, showing f/4.5 both when the camera is configured to show half-stops and to show 1/3 stops.

Going forward we are adding the same aperture display tweak to converted Contax lenses. This change is supported by several factors. First, this poll shows most customers want the display to show f/4.5. Second, Canon does the same tweak to their own lenses, and lastly, we would like to end all future customer confusion. Hopefully your time pondering the f/5.0 phenomenon and our time explaining it could be channeled to more productive pursuits.

Hope this explanation helps.

[Hening]

Thank you, conurus, for this very exhaustive answer!