Optics/Depth of Focus question.

bbjones

Can anyone make sense of this for me?

I took two shots, one with my 17-55mm, and one with my brother-in-law's 24-70L. Both shots were on my 7D, both shots were at f/2.8. The 17-55mm shot was at 1/40th, and the 24-70L was at 1/50th.

The 17-55mm was dialed to 33mm, and the 24-70 was at 35mm.

Below are the two shots. To my eye, the 24-70 has a significantly more shallow depth of focus. But to my head, it shouldn't matter -- focal length, sensor, and aperture stop are the same. Why are they different?

These images are converted from the RAW at the default Lightroom settings, without noise reduction (the familiar 7D noise at 1600ISO is there for your enjoyment.)

Links to full size: 24-70mm 17-55mm

24-70mm:


17-55mm:


I'd love any insight into what's going on here optically!

JonaBeth Russell

From what I can see on my screen, it seems as though the 24-70 shot has more clarity and definition overall. I know the 17-55 is known to be "just like an L", but the fact is, it's not. Additionally, your focus subject (the pooch) appears to be a tad closer to the body of the person holding it in the second shot (17-55). You can see that the shirt is a little more in-focus than the first shot, but then again the entire second shot is not as sharp as the first.

Plus, maybe there is a little compression happening in the 24-70 shot?

Those are just my off-the-hip ponderings. I don't have the absolute answer.

jmphotocraft

I noticed the same thing when I switched from a 24-70L to a 17-55/2.8IS on my old 40D, the difference is real. I had both lenses together for a brief time and I did some tests which showed this. I think it has to do with two things, 1, I believe the 17-55 sits closer to the sensor, and 2, I believe the 17-55 projects a smaller image circle, so the light rays are less divergent. I could be wrong, but that makes sense to me.

I would say for sharpness, the 17-55 is every bit as sharp as the L. The biggest difference is the dust vacuum construction.

bbjones

jmphotocraft wrote: »

I noticed the same thing when I switched from a 24-70L to a 17-55/2.8IS on my old 40D, the difference is real. I had both lenses together for a brief time and I did some tests which showed this. I think it has to do with two things, 1, I believe the 17-55 sits closer to the sensor, and 2, I believe the 17-55 projects a smaller image circle, so the light rays are less divergent. I could be wrong, but that makes sense to me.

Well, glad I'm not crazy. And your explanation does make sense, except isn't that identical to saying that the effective focal length is different? Now, that could very well be, but it would mean that 35mm is not 35mm between lenses.

jmphotocraft wrote: »

I would say for sharpness, the 17-55 is every bit as sharp as the L. The biggest difference is the dust vacuum construction.

I'm not a pixel-peeper; to me, given an excellent level of sharpness available on every DSLR now with near-L or better optics, composition is much more important. (This silly test shot notwithstanding.) That said, I was not so blown away by the difference as to add the 24-70 to my wish list.

jmphotocraft

bbjones wrote: »

Well, glad I'm not crazy. And your explanation does make sense, except isn't that identical to saying that the effective focal length is different?

Not necessarily. I don't know the exact engineering explanation, but you could see it for yourself by conducting a more controlled test than the one above.

Now, that could very well be, but it would mean that 35mm is not 35mm between lenses.

Again not necessarily. Different lenses with the same focal length can actually have different characteristics, like angle of view, etc.

I'm not a pixel-peeper; to me, given an excellent level of sharpness available on every DSLR now with near-L or better optics, composition is much more important. (This silly test shot notwithstanding.) That said, I was not so blown away by the difference as to add the 24-70 to my wish list.

No, and for APS-C the 17-55 really is the correct choice between the two. 55mm is not significantly shorter than 70mm, you can make up the difference by taking about two steps closer or just leaning in. Meanwhile 17mm is much wider than 24mm, and you cannot usually back up enough to make up that difference. And 24mm x 1.6 = 38.4mm, which is really not a wide angle view at all.

Matthew Saville

So is the lesson here, that DOF and quality of bokeh are not 100% connected? I'd wager that the larger amount of glass in the 24-70 gives you slightly smoother bokeh.

I would, however, also wager that this was not a precise test, as it did not involve truly immobile subjects.

To definitively resolve this, grab a tripod, and take pictures of a truly static object. Zoom to the EXACT same focal length, in fact try a few different focal lengths.

Post THOSE results, and we'll know for sure...

=Matt=

jmphotocraft

Matthew Saville wrote: »

So is the lesson here, that DOF and quality of bokeh are not 100% connected? I'd wager that the larger amount of glass in the 24-70 gives you slightly smoother bokeh.

I would, however, also wager that this was not a precise test, as it did not involve truly immobile subjects.

To definitively resolve this, grab a tripod, and take pictures of a truly static object. Zoom to the EXACT same focal length, in fact try a few different focal lengths.

Post THOSE results, and we'll know for sure...

=Matt=

http://www.dgrin.com/showthread.php?p=1983397#post1983397

I don't have the images anymore to post, but I noticed the same thing.

bbjones

Matthew Saville wrote: »

So is the lesson here, that DOF and quality of bokeh are not 100% connected? I'd wager that the larger amount of glass in the 24-70 gives you slightly smoother bokeh.

I guess I don't think of depth of focus and bokeh quality as being connected at all. I'm not really looking at the quality of the bokeh, just the depth.

Matthew Saville wrote: »

I would, however, also wager that this was not a precise test, as it did not involve truly immobile subjects.

It's certainly not a high-precision test, but it's not a subtle difference. Look at the text on the sweatshirt; it's quite a huge difference in DOF.

You are right that a higher precision test could measure the effect more carefully, but that's not necessary to see the effect here. Or put another way, to quantify the effect would require a careful setup, but you can see it's there even without a careful setup.

And a more precise measurement still won't answer the optical engineering question: why should it be different at all?

ziggy53

I believe that there are a few misconceptions which explain most of the discrepancies in this case. (Below is representative only of rectilinear corrected lens designs.)

1) "Zoom lenses focal length markings and ratings are accurate."

False. (ref 1)

Zoom lenses in particular have a marketing rating for both zoom range and for focal lengths. While there is no true standard for focal length accuracy in a zoom lens, it's often to find zoom lenses misrepresent their true focal lengths by +- 10 percent (or so).

---

2) "Zoom lens focal length markings are accurate at any focus distance."

False. (ref 2)

Most, if not all, zoom lenses "breath" with focus settings and different zoom settings. Generally, a zoom will have [strike]less[/strike] larger FOV (therefore less focal length) at closer distances than at infinity focus, which tend to be the most accurate marking representations. Each zoom optical formulation has a different amount of breathing, so comparing two dissimilar lenses at close focus is bound to produce different results.

---

3) "All lenses are "flat field" lenses, meaning that the image center and out to the corners will yield the same level of sharpness at any given distance to subject and, in the case of zooms, the same focal length."

False. (ref 3)

Zooms in particular are often a compromise in field curvature, and many will display field curvature at different settings of both zoom and focal distance. Field curvature will indeed impact on the apparent DOF of a given 3-dimensional scene, especially at maximum aperture. The DOF discrepancy will often be most noticeable at closer focus distances.

---

So, ziggy, how do I know for sure what focal length I am using at what zoom setting? I'm glad you asked.

The easiest method is to determine the FOV accurately. Once you know the field of view/angle of view you can calculate the true focal length at that setting. Likewise, if you match two dissimilar lenses with the same field of view/angle of view on the same camera body, you should be at the same focal length.

---

References (IMO)

(1) http://www.bobatkins.com/photography/technical/measuring_focal_length.html

(2) http://www.bobatkins.com/photography/technical/focus_breathing_focal_length_changes.html

(3) http://tashley1.zenfolio.com/blog/2013/5/field-curvature---a-practical-guide

bbjones

Thanks, Ziggy!
That's really helpful. As usual, when your data doesn't match up, go back and look at your assumptions.

I wonder if some of this is purposeful. One of the great things about the 24-70 and 24-105 lenses is the background. Part of that is well designed blades and optics, but maybe part of it is smaller than expected DOF.

In any case, thanks for the insight!
-Brian

Matthew Saville

ziggy53 wrote: »

I believe that there are a few misconceptions which explain most of the discrepancies in this case. (Below is representative only of rectilinear corrected lens designs.)

1) "Zoom lenses focal length markings and ratings are accurate."

False. (ref 1)

Zoom lenses in particular have a marketing rating for both zoom range and for focal lengths. While there is no true standard for focal length accuracy in a zoom lens, it's often to find zoom lenses misrepresent their true focal lengths by +- 10 percent (or so).

---

2) "Zoom lens focal length markings are accurate at any focus distance."

False. (ref 2)

Most, if not all, zoom lenses "breath" with focus settings and different zoom settings. Generally, a zoom will have [strike]less[/strike] larger FOV (therefore less focal length) at closer distances than at infinity focus, which tend to be the most accurate marking representations. Each zoom optical formulation has a different amount of breathing, so comparing two dissimilar lenses at close focus is bound to produce different results.

---

3) "All lenses are "flat field" lenses, meaning that the image center and out to the corners will yield the same level of sharpness at any given distance to subject and, in the case of zooms, the same focal length."

False. (ref 3)

Zooms in particular are often a compromise in field curvature, and many will display field curvature at different settings of both zoom and focal distance. Field curvature will indeed impact on the apparent DOF of a given 3-dimensional scene, especially at maximum aperture. The DOF discrepancy will often be most noticeable at closer focus distances.

---

So, ziggy, how do I know for sure what focal length I am using at what zoom setting? I'm glad you asked.

The easiest method is to determine the FOV accurately. Once you know the field of view/angle of view you can calculate the true focal length at that setting. Likewise, if you match two dissimilar lenses with the same field of view/angle of view on the same camera body, you should be at the same focal length.

---

References (IMO)

(1) http://www.bobatkins.com/photography/technical/measuring_focal_length.html

(2) http://www.bobatkins.com/photography/technical/focus_breathing_focal_length_changes.html

(3) http://tashley1.zenfolio.com/blog/2013/5/field-curvature---a-practical-guide

BIG chunk of useful info here!!!! :-)

Qarik

so in the 24-70 shot the shirt logo is out of focus (far away) but the couch details on left side are in focus (closer). On the 17-55 shot it is reverse. I can't really say one has more DOF then the other..I think the easiest explanation here is that the focal point was not the same in both pictures...24-70 focus point was closer and the 17-55 was further.

Cavalier

So, after all the wonderful technical information in this thread.................can I ask - what kind of dog is that ?

Love the dog.