Never had a camera that sees it that way though
Upvote
0
RF 85mm f/1.2 or 135mm f/1.8?
- Thread starter a.jiang
- Start date
L
Loswr
Guest
Ever set one to 1/2 stop vs 1/3 stop increments and noticed the values are different?Never had a camera that sees it that way though
Doesn’t matter at the end of the day, because what your camera really sees are T-stops.
Upvote
0
One third of a stop is almost precisely one decibel. (A doubling or halving [i.e., a stop] is roughly 3 dB.)
Upvote
0
I was assuming that Canon uses the 1/3-stop scale.
f/1.2 = 1/3-stop faster than f/1.4
f/1.8 = 1/3-stop faster than f/2.0
In which case f/1.2 is exactly one stop faster than f/1.8.
Of course, the precise 1/3-stop numbers are:
2^(0/6) = 1.0
2^(1/6) = 1.12246204831
2^(2/6) = 1.25992104989
2^(3/6) = 1.41421356237
2^(4/6) = 1.58740105197
2^(5/6) = 1.78179743628
2^(6/6) = 2.0
f/1.2 = 1/3-stop faster than f/1.4
f/1.8 = 1/3-stop faster than f/2.0
In which case f/1.2 is exactly one stop faster than f/1.8.
Of course, the precise 1/3-stop numbers are:
2^(0/6) = 1.0
2^(1/6) = 1.12246204831
2^(2/6) = 1.25992104989
2^(3/6) = 1.41421356237
2^(4/6) = 1.58740105197
2^(5/6) = 1.78179743628
2^(6/6) = 2.0
Upvote
0
L
Loswr
Guest
Canon cameras use 1/3 or 1/2 stop increments, depending on the selected setting.
Regardless, in actual terms of the amount of light, the difference between f/1.2 and f/1.8 is more than one stop.
Upvote
0
Yeah, but when they market the 85mm as f/1.2, is that on the 1/3-stop scale, or the 1/2-stop scale ?
IOW, does it mean
2^(1/3) = 1.25992104989 (third-stop scale), or
2^(1/4) = 1.189207115 (half-stop scale) ?
Not if f/1.2 means one-third-stop faster than f/1.4, and f/1.8 means one-third-stop faster than f/2.
Then the difference is exactly one stop.
Reference:
https://en.wikipedia.org/wiki/F-number
(seems pretty good AFAICT)
Last edited:
Upvote
0
Wind back a bit.
An f-stop number is the focal length over entrance pupil diameter. E.g., 85mm/71mm=1.2 or so.
But as you focus, the entrance pupil size changes, so maximum f-stop isn't a constant but a variable.
More importantly, it changes size and shape depending on the angle. Once it's no longer circular, instead of the diameter, it'd make more sense to use not the diameter, but 4/pi* square root of the area. So a common "cat's eye" or "American football"-shaped entrance pupil off center would have a substantially smaller entrance pupil size even if the major axis is nearly the same. This is what gives mechanical vignetting: an RF 50/1.2 loses say 2 stops in the corners meaning at that point it is in some respects an f/2.4. You can average out the total average entrance pupil and get a T-stop as used in the film industry.
So, the f/stop is really nominal, and serves as a good proxy for what the maximal entrance pupil size is but doesn't tell you much else.
And finally, manufacturer patents show they're basically advertising anyway. Each brand has some amount of slop factor, it seems, that they actively target. You'll see a Canon patent with three different illustrative uses, each resulting in a lens of say f/1.51 or something. Why all f/1.51??? Well, probably because their internal rules, or maybe, an industry standard, let's them market this as f/1.4. (You also see this in focal lengths:"300mm" lenses are 284mm or whatever, while wide angles are always a tiny bit narrower than the spec claims.)
So don't worry too much about what the exact number is. Due to vignetting, especially, NO lens I know of (at least less than telephoto) is even within 2/3-stop of its nominal f-stop, wide open, by the time you get to the corners.
Last edited:
Upvote
0
L
Loswr
Guest
It’s irrelevant. They market it as f/1.2. The ‘real’ f/number for the lens (based on the Canon patent) is 1.24, and the ‘real’ focal length is 83.2 mm. It actually transmits T = 1.3, which is better than its predecessor, the EF 85mm f/1.2 II that was T = 1.5.
If you look at the 1/3-stop scale on the wiki page you posted, f/1.2 is 1/3-stop faster than f/1.4 as you say, but if you look at the 1/2-stop scale on the same wiki page, f/1.2 is also 1/2-stop faster than f/1.4. How can f/1.2 be both 1/3- and 1/2-stop faster than f/1.4 at the same time? It can't and it's not. That's why the wiki page indicates those are 'typical f-number scales', i.e. approximations.
With 1/2-stop increments, f/1.2 is one full stop faster than f/1.7, as shown in the graphic I posted and in your wiki link. Therefore, the difference between f/1.2 and f/1.8 is more than one stop.
The f/numbers printed on barrels and used as camera settings are all approximations anyway, just like those tables. And they don't always match up. For example, here is the half-stop progression from f/1.4 to f/8 from the wiki page, and below that is the progression on my R3 when I progressively close down the EF 85/1.4 from wide open to f/8 with the camera set to use 1/2 stop increments:
| 1.4 | 1.7 | 2 | 2.4 | 2.8 | 3.3 | 4 | 4.8 | 5.6 | 6.7 | 8 |
| 1.4 | 1.8 | 2 | 2.5 | 2.8 | 3.5 | 4 | 4.5 | 5.6 | 6.7 | 8 |
So below f/5.6, the intermediate half stop increment is actually shown as a 1/3 stop setting, but even then not consistently shifted (e.g. the R3-shown value between f/2 and f/2.8 is the 1/3-stop increment closer to the narrower f/2.8, but the value between f/4 and f/5.6 is the 1/3-stop increment closer to the wider f/4). As I said, approximations and really the 1/6-stop differences don't matter.
What is not an approximation is the actual math. As an example, a shot at 1/100 s, f/1.2, ISO 100 has EV = 7.17, and a shot at 1/100 s, f/1.8, ISO 100 has EV = 8.33. 8.33 – 7.17 = 1.16, and 1.16 is greater than 1.
In real world terms, 0.16 EV is irrelevant.
Upvote
0
Just for some bokeh fun, I compared shots between RF 50 f/1.2, 85 f/1.2 and 135 f/1.8 lenses. All shot from same position, with the 50 and 85 cropped to match the 135.
Interesting to note the 50 and 85 have pretty much the same warm tone and the 135 is a bit cooler (on my monitor).
Interesting to note the 50 and 85 have pretty much the same warm tone and the 135 is a bit cooler (on my monitor).
Last edited:
Upvote
0
Upvote
0
85mm/1.2=71mm or so. 135mm/1.8=75mm. So their physical aperture, or "entrance pupil" is basically the same, so a crop of the 85/1.2 down to a 135mm angle of view will have just about exactly the same bokeh as the 135. And indeed, the 135 vs cropped 85 show more or less exactly the same thing.
While f/1.2 sounds big, on a 50mm lens that's "only" a 40mm aperture. And bokeh is a question of area, or aperture SQUARED, so the 85 and 135 actually have about 3x the bokeh, even if you crop the 50mm shot to the 85 or 135 field of view.
This is why I think it's not a silly request for Canon to make a "halo" 35/1.0 for RF. It sounds insane but that's only a 35mm aperture, same as a 50/1.4. It's not huge at all.
Last edited:
Upvote
0
Yep, it was interesting to see how that worked. The 50 definitely can't hang with the 85 & 135, so I see what you're saying about a 35 f/1.
Upvote
0
It’s a nice comparison video!
If AF and tracking is important and the FL is both just as usable, go with the 135, if shallow dof and light gathering is the most important, go with the 85.
Now, if Canon only made a 100 f1,4 with best of both that would be my choice.
Upvote
0
Similar threads
