is 24mm prime the same as 24mm on a 18-105 lens

[Gavjenks]

think I'll test for myself too.

I'm assuming I'll have to move the position of the camera based on the length of the lens, ie, have the lens opening at the same distance from a wall versus having the sensor at the same location.

Hmm, I was thinking the opposite and leaving the sensor in the same position.

Leaving the sensor in the same place is the correct method (easily accomplished by simply having the camera on a tripod in the body's base mount). That way you're holding the sensor to object distance constant, and only varying the lenses, which are the things we want to compare.

 

[runnah]

Yes.

 

[Helen B]

No.

You should have the front nodal points or the entrance pupils in the same position, not the sensor.

 

[Gavjenks]

No.

You should have the front nodal points or the entrance pupils in the same position, not the sensor.

Care to explain why?
Also, care to explain HOW, since the position of the nodal points is not published information for most lenses, short of using lasers and test benches and the like? (I'm not suggesting it can't be done fairly easily. But just throwing that out there and then going "kthxbai" is not very helpful)

Remember, the goal here is not to measure the exact focal length. To do that, you would indeed like to know where those points you mention are. But all we are really wanting to do here is simply determine--Yes or No--whether the focal lengths of the two lenses are exactly equal, or not, to answer the OP. How much they are different by is a higher bar to clear, which would be nice and luxurious, but is not necessary here.

So if you line up the sensor and the object, and the images don't end up showing the same field at the same written FL, then you know that the nodal points are not in the same places. That is sufficient to answer "No, they aren't the same FL in actuality."

If the field you see is the same, then (setting aside distortions) that is sufficient to answer "Yes, they are the same measured FL (whatever that is, which may not be the written one!)"

 

[Dao]

No.

You should have the front nodal points or the entrance pupils in the same position, not the sensor.

Care to explain why?
Also, care to explain HOW, since the position of the nodal points is not published information for most lenses, short of using lasers and test benches and the like? (I'm not suggesting it can't be done fairly easily. But just throwing that out there and then going "kthxbai" is not very helpful)

Remember, the goal here is not to measure the exact focal length. To do that, you would indeed like to know where those points you mention are. But all we are really wanting to do here is simply determine--Yes or No--whether the focal lengths of the two lenses are exactly equal, or not, to answer the OP. How much they are different by is a higher bar to clear, which would be nice and luxurious, but is not necessary here.

So if you line up the sensor and the object, and the images don't end up showing the same field at the same written FL, then you know that the nodal points are not in the same places. That is sufficient to answer "No, they aren't the same FL in actuality."

If the field you see is the same, then (setting aside distortions) that is sufficient to answer "Yes, they are the same measured FL (whatever that is, which may not be the written one!)"

I remembered reading something here in the past (maybe from Helen, but not 100% sure) regarding shooting panoramas with reference to the nodal point instead of using swivel mount on the tripod.

 

[Braineack]

I remember reading something about how when someone puts a 24mm lens on a FX the physical world changes and everything literally becomes further away.

 

[Helen B]

Gav, this is so simple. The object distance (lens to subject distance) is the distance from the front nodal point to the object in focus. The image distance is that between the rear nodal point and the image plane. The object to sensor distance (which is what is marked on the distance scale of most lenses, for practical reasons) is the sum of the object distance, internodal distance (which may be negative) and image distance. You can have lenses with the same focal length (hence the same image distance for the same object distance) but very different internodal distances.

The entrance pupil is the surface the lens 'sees' the world from. It is very easy to find.

Do you need more of this explaining or do you know enough to fill in the gaps? You haven't explained why your answer is correct, but you expect me to explain my answer - which should be obvious to anyone who has some knowledge of optics. I would like to read your explanation of why you believe you are correct.

 

[Dao]

Helen, I just read something regarding panoramas shooting. From what I read from that article, it seems like the best point should be the entrance pupil of the lens instead of the nodal points because there are no parallax on the entrance pupil and it is the center of perspective. Am I right?

 

[Helen B]

Yes, it's the entrance pupil (for a stitched panorama using the rotation of the whole camera as a unit*). At first glance it might seem that the front nodal point was the correct location of the no parallax point, but it is the entrance pupil that determines which light rays will form the image, particularly for points outside the plane of focus. Because parallax is a 3D thing, we need to consider those out-of-true-focus points. If you are comparing the angle of view of two lenses in 3D, it would be best to have the entrance pupils in the same place; if doing it with a plane surface at right angles to the optical axis and in focus, such as with a brick wall, then the front nodal points could be in the same place. For most practical purposes it won't matter, unless you really want to match views as exactly as possible or are naturally curious about what is going on.

*For a camera with a fixed film/sensor and rotating lens with slit shutter the correct point of rotation of the lens for a static, non-smeared image is the rear nodal point of the lens. This may mean that the entrance pupil moves during the exposure. If the image is exposed through a slit the movement of the entrance pupil usually doesn't seem to matter.

 

[Dao]

Thanks Helen, that explains it. 2D vs 3D! How can I not know?? <A light bulb just pop up above my head>

 

[astroNikon]

Thanks Helen, that explains it. 2D vs 3D! How can I not know?? <A light bulb just pop up above my head>

My light bulb lite, unscrewed itself, fell and smashed on the floor.
But I think I got it because I know how to change light bulbs in the dark.

 

[Gavjenks]

Gav, this is so simple. The object distance (lens to subject distance) is the distance from the front nodal point to the object in focus. The image distance is that between the rear nodal point and the image plane. The object to sensor distance (which is what is marked on the distance scale of most lenses, for practical reasons) is the sum of the object distance, internodal distance (which may be negative) and image distance. You can have lenses with the same focal length (hence the same image distance for the same object distance) but very different internodal distances.

The entrance pupil is the surface the lens 'sees' the world from. It is very easy to find.

Do you need more of this explaining or do you know enough to fill in the gaps? You haven't explained why your answer is correct, but you expect me to explain my answer - which should be obvious to anyone who has some knowledge of optics. I would like to read your explanation of why you believe you are correct.

Now that you point that out, I am not correct about merely lining up the sensors and objects.
However, neither are you correct about merely lining up the entrance pupils and objects.

1/FL = 1/(front node to object) + 1/(back node to sensor)

So lining up the objects and the front nodes is exactly as insufficient as lining up the sensors is. They both only end up holding constant 2 out of the 4 numbers that matter (object and sensor, or object and front node). No amount of lining up anything in fact will therefore work to automatically take care of such math for us, because one variable can't solve for two unknowns. It would require measuring the additional values and computing ratios no matter what.

However, we don't know where the rear nodal point is, so that entire method won't work, either, if you want to actually know the exact FL (or do we? Please explain how to find the rear nodal point if so). You can't just do arithmetic with the distance scale markings, because the entire point of the thread is that we don't trust the manufacturer's markings (Also lots of lenses don't have those anyway). So since we can't trust them not to have rounded off or fudged those numbers either, how else can we calculate the rear nodal distance?

You might be able to measure only the sensor, and then use back focal as your unknown to solve, to simple answer whether the FL's are the same (not what they are). But if you're going to measure something and compute anyway, then starting by lining up sensors, and then measuring the front nodal point and calculating would ALSO work, and would take exactly as much effort.

Instead of all of that, I suggest an entirely different, empirical, guess and check method:
1) Get a piece of pegboard
2) jam some dowel rods in a rectangle, making them fairly far apart (several inches at least)
3) Tape off a straight perpendicular line from the dowel rod on the ground and place camera on a stool or tripod centered over that line at a height amidst the height of the dowel rods.
4) Put on lens #1, and take photos and adjust along the tape line until you get the front two dowel rods right at the edges of the frame.
5) Put on lens #2, and do the same thing, until they're right at the edges of the frame.
6) Measure the parallax distance between, for example, the front right dowel and the back right dowel in the two photos. If the effective focal lengths are equal, then parallax should be equal when the framing is matched. If not, then not.

 

[amolitor]

If you hold the front node to object distance fixed, and then jam a lens into the picture, the back node to sensor distance will in fact nail itself down. You have to move the sensor around until your object is in focus.

Explaining to Helen that she is wrong usually goes pretty poorly. Not always, I expect, but usually.

 

[Gavjenks]

You have to move the sensor around until your object is in focus.

I guess that might work, but then that's an additional step that wasn't mentioned. That would be "line up the front nodal point and object for each lens, and then fiddle with the sensors until they are focused the same way," not simply "line up the front nodal point and the object." Moving the sensors without also moving the nodal point (and thus making a feedback loop of endless measuring for yourself) may not actually be very trivial at all... And even if it is, it's something you have to consciously think about and correct for, not just hit the AF button.

Plus, Helen still never mentioned how to actually measure the front nodal point, other than "it's very easy." That may well be, but googling it shows a number of methods that all require lasers. I can imagine some slapshod ways of eyeballing it that are easier than that, but I can't think of (or find) easy ways of accurately measuring it. Doesn't mean they don't exist, but if I can't find them as an experienced hobbyist who builds lenses and has looked up optical formulae many times before, then I imagine a beginner on the beginner's forum would not easily be able to find it either.

I think my alternative pegboard thing will still also work, and can be followed step by step in a few minutes (with materials handy) with lay knowledge, in the meantime.

 

[amolitor]

Well, if you just move the sensor around at random then your formula for focal length kind of falls apart. The point is that you've got fewer variables than you think you do in there. With an actual physical lens in play, once you've nailed down one of the two distances, the other asserts itself.