Crop factor - misleading

[unpopular]

Under that calculation, sure, it should work - assuming that the figures used were accurate (they weren't, the web site even says they aren't), that there was no atmospheric distortion (there was, there always is - especially in that climate), that you didn't round anywhere (you likely did), etc.

 

[480sparky]

In order to test a 300mm lens at 300mm, you first need to verify it actually is 300mm, don't you? Until you do that, you're using a flawed system.

 

[unpopular]

^^ yep. even if it's a hair off, that calculation will be wildly inaccurate, especially with the flawed standards employed.

 

[Derrel]

I thought that the old JCII (the oval sticker people...for those old enough to remember...Japan Camera Inspection Institute, I think it was) and its associated manufacturers hewed to a standard where lens focal lengths needed to be within 10% of the marked length....and so the "278mm three hundred millimeter" lens Gavjenks mentioned in an earlier hypothetical would easily be within tolerance under that set of standards. NOT SURE if this is still followed these days...

I've read a lot of the Popular Photography test reports over the decades, and on zoom lenses, the minimum focal length is often just a small bit longer than the marked minimum length, and the maximum focal length is often shorter than the marked length. The percentage of error is usually greater for the long end of the zoom, while on the short end, the actual measured lengths are usually much closer in both percentage, and in absolute length in millimeters.

 

[Gavjenks]

In order to test a 300mm lens at 300mm, you first need to verify it actually is 300mm, don't you? Until you do that, you're using a flawed system.

Uh, that's the whole point of the procedure, to verify whether it is 300mm. The calculations are to tell you what the actual FL is of the lens you're holding. And if it comes out to 300mm, then the marking is correct. If not, it isn't.

Under that calculation, sure, it should work - assuming that the figures used were accurate (they weren't, the web site even says they aren't), that there was no atmospheric distortion (there was, there always is - especially in that climate), that you didn't round anywhere (you likely did), etc.

It was off by like... a lot. 60mm or so, IIRC. Several orders of magnitude more than rounding would have anything to do with (the website or mine). Distortion, however? Still seems unlikely, but no I don't know for sure.
So okay, whatever, trust the manufacturers then *shrug*
(Although you have to ask yourself, do you really think they'd go to the trouble of designing a minimally flawed lens at the right price point, find out that it's 273mm, then redsign the whole thing until it hit 300? Or do you think they'd just round it...? Considering all the lenses for sale are advertised in multiples of 100 at the tele end... seems pretty likely the latter is true even without testing, but maybe that's just me)

Anyway, I may actually build a laser test bench sometime in the future for unrelated projects (making homemade lenses), and could try again with that as long as I'm doing it anyway.

^^ yep. even if it's a hair off, that calculation will be wildly inaccurate, especially with the flawed standards employed.

No, it'd have to be HUGELY off. The moon would have to appear ~20% smaller than it should to yield that difference. A difference of something like 100 pixels (i.e. very much within precision of all measuring instruments involved)
The atmosphere may very well be enough to change it by 20%, I admit. But whether it is or not, that's not what I would call "a hair."

From what I'm reading now, it appears to be almost completely negligible above about 45 degrees from the horizon, btw / much less than the disparity I found.

 

[unpopular]

I'm confused what you mean by 60mm. Are you saying that the projected image of the moon is 60mm smaller than what you expected?

 

[Gavjenks]

No, I mean the angle that I got from my calculations corresponds to published theoretical angles of view that you should expect for a sensor the size of mine and a 240mm lens, when my lens was turned to the "300mm" marking
(or roughly that, but I'm pretty sure I remember it being more than 50mm off, because I remember thinking that it would have not even rounded to 300 if you went to the nearest 100mm)

 

[sk66]

Focal length is from the rear nodal point to the focal plane. BUT the rear nodal point does not necessarily fall within the physical boundaries of the lens (it is NOT the "nodal point" axis for taking pano's)
The primary factors of focal length is the FOV and magnification power ("reproduction ratio," which is the inverse of the focal length). It does not determine the projected image circle size. The projected image size is determined by the distance of the last set of elements from the focal plane. For example, Canon EF-S lenses use a shorter distance and project a smaller image circle (the "S" is for "small image circle" according to wikipedia... it could also stand for "short back focal distance" IMO).

Yes, manufacturer's often "round" focal lengths to the nearest "normal/common" focal length for simplicity sake, and a few mm of FL can make a noticeable difference in FOV (UWA's). But none of this accounts for 19mm looking like 29mm (or anything approximating that IMO). The only thing I can think of is an actual lens issue... Like maybe it's of a design that also moves the rear elements and that cam/gear is broken/sticking.