Long Lenses & Low Light

[Highbow]

For most all of my bird photography the Z100-400 S F4 is just so fitting for my needs. I can use either my 1.4 or 2.0 TC for additional range on my Z6iii . The idea of carrying the extra weight of a longer lens is less and less appealing as I approach 70 years of age. Sure if money was not an object I could invest in more glass but photography is still my hobby not what supports my income.

 

[Bill Ferris]

In addition to all of the above, another factor I look at before buying any lens or using a lens I own is its Minimum Focus Distance or Close Focus Distance. For my two Nikon 300mm F mount lenses, the 300mm f4 pf has a MFD of 4.6 ft; the 300mm f2.8 G II has a MFD of 7.22 ft. And its weight can also be a factor too. Chasing butterflies is a lot easier with the 300mm f4 than the 300mm f2.8.

Thank you, Joe. That's a great point you raise about MFD.

There are many factors we take into consideration when choosing a lens. Low light performerance is just one and, depending on the photography a person does, not necessarily a top priority. When your subjects are closer than 5-feet in distance, MFD is among the top priorities.

Thanks for adding your perspective.

 

[bleirer]

The entrance pupil is the lens's aperture.

Apparent aperture, no? Rather than the actual physical size of the diaphragm? Since we are deep in the weeds I agree a more basic explanation of these concepts would help me understand better.

 

[Nimi]

Is there a reliable resource where published T-stops for a range of lenses is archived and updated? If not that would be a great resource for someone to curate, a la Bill Claff's "photons to photons" site.

Never seen one. For many years the two worlds didn't overlap so conversion wasn't necessary. Canon produced the same lenses in different housing for stills and cine, as did Tokina, Sony, Zeiss, so that's probably the nearest comparison. Rule of thumb is about 5-10%, so an f/2.8 is T3. That's one of the reasons cine lenses are always introduced as sets with constant T (except the more extreme focal lengths).

 

[Bill Ferris]

Yes, I don’t necessarily disagree, but I suppose it depends on the framing required, size of the subject, how close can one get to the subject etc? For instance, I enjoy photographing deer/foxes towards the end of the day & the 400 2.8 wins hand down based on my direct experience having owned both a 400 2.8 & 600 f4 at various times. Ditto capturing barn owls in flight as the 1 stop is a huge advantage towards the end of the day (of course, getting closer to get the required framing isn’t an issue for me re: the above). However, yes, there are times when a 600 is more advantageous, especially when it comes to small birds etc.

In terms of raw light gathering, the difference between the 400/2.8 and 600/4 is so minimal (10%) as to be negligible. I would imagine the f/2.8 is an advantage when fast, accurate focus in weak light is critical. And the 400's MFD advantage must also be an advantage for some subjects.

 

[bcgnz]

In terms of raw light gathering, the difference between the 400/2.8 and 600/4 is so minimal (10%) as to be negligible.

This is incorrect though? In terms of raw light gathering, the 400/2.8 collects double the amount of light. It’s only when shooting at the same distance AND cropping/teleconvertering the 400/2.8 image to the same composition that the amounts of light drawn are equal.

 

[SCoombs]

From the standpoint of practical experience, I maintain that an f2.8 lens even with a shorter focal length tends in certain situations to produce lower noise than, say, an f6.3 lens at a greater focal length even when heavy cropping is involved.

This past fall I was shooting HS soccer games in what was at times relatively poor late afternoon shaded lighting. I primarily used my 70-200/2.8 but also used the 180-600. I got into a little bit of a debate on a forum with some people insisting that I should essentially not be using the 70-200 for soccer because the focal length was inadequate and that a 400 or 600mm lens, even at a much higher aperture, would yield better results. I responded that other sports photographers I see covering these games also use their 70-200s and that the 2.8 aperture is worth the crop vs. the 6.3 aperture of the 180-600. There was still disagreement, so I took a game to go out of my way to test this.

In good light, yes, the 180-600 was better.

In worse light, I found that even cropping the 200mm in to the equivalent FOV of the 600mm yielded better, less noisy results. By the calculations offered in this discussion, we would say that the 200/2.8 has an entrance pupil of or about 71 vs. the longer lens with an entrance pupil of 600/6.3 or 95 - but my subjects absolutely looked better even cropping from all the way across the field on the 70-200 vs. using the 180-600 at f6.3.

 

[Doug Herr]

This is incorrect though? In terms of raw light gathering, the 400/2.8 collects double the amount of light. It’s only when shooting at the same distance AND cropping/teleconvertering the 400/2.8 image to the same composition that the amounts of light drawn are equal.

Bill Ferris is correct and you are also correct, each in the correct context. At maximum aperture and at the same distance, the 400mm f/2.8 and the 600mm f/4 having nearly identical entrance pupils put nearly the same amount of light on the subject. The 400mm f/2.8 collects twice the light as you have pointed out, but it's over a wider field with a smaller subject. The additional light gathered by the 400mm f/2.8 outside the field of view the 600 would produce is discarded when the image is cropped or TC'd to match the 600's field of view.

 

[bcgnz]

Bill Ferris is correct and you are also correct, each in the correct context. At maximum aperture and at the same distance, the 400mm f/2.8 and the 600mm f/4 having nearly identical entrance pupils put nearly the same amount of light on the subject. The 400mm f/2.8 collects twice the light as you have pointed out, but it's over a wider field with a smaller subject. The additional light gathered by the 400mm f/2.8 outside the field of view the 600 would produce is discarded when the image is cropped or TC'd to match the 600's field of view.

Perfectly said.

 

[Bill Ferris]

This is incorrect though? In terms of raw light gathering, the 400/2.8 collects double the amount of light. It’s only when shooting at the same distance AND cropping/teleconvertering the 400/2.8 image to the same composition that the amounts of light drawn are equal.

A 600mm f/4 lens has a 150mm diameter entrance pupil. A 400mm f/2.8 lens has a 143mm entrance pupil. The 600's slight aperture advantage translates to collecting 10% more photons from any subject in the frame. The noise difference may or may not be discernible to a viewer.

 

[Bill Ferris]

Apparent aperture, no? Rather than the actual physical size of the diaphragm? Since we are deep in the weeds I agree a more basic explanation of these concepts would help me understand better.

It's an image of an aperture as seen when looking through the front of the lens, a virtual entrance pupil.

 

[bcgnz]

A 600mm f/4 lens has a 150mm diameter entrance pupil. A 400mm f/2.8 lens has a 143mm entrance pupil. The 600's slight aperture advantage translates to collecting 10% more photons from any subject in the frame. The noise difference may or may not be discernible to a viewer.

By the same token, a 800mm f/5.6will be able to take the head of lion while the 400mm f2.8 will be able to take the entire body of the lion, while maintaining the same density of light on the subject.

 

[Bill Ferris]

Your referenced Appendix in the above blog explains how the magnification of an optical lens is intrinsically linked to the numerical aperture and working f-number. This explains the fundamental relationship between the Focal Length and Effective Window, which determines the fastest fStop of an optical lens.

This one from Hogan's blog us more directly on point: https://www.strollswithmydog.com/pupils-apertures/

The diagram simplifying a lens as a black box with entrance and exit pupils conveys the point.

mmmm... a tough read. I don't think i'm alone in struggling to find the scientific credibility in optical physics for this statement - italics added: "We saw above how changing sensors size resulted in a change in focal length, which when used at the same f/ratio results in different lens diameters and different noise levels." This is a common mistake encountered across the internet.

What did you think of the sample photos demonstrating that a lens with a larger entrance pupil collects more light from a subject than another lens at the same f-stop but a smaller entrance pupil? The night sky scenes illustrate the advantage pretty clearly.

Sound principles of physics explain why cinematographers pay for extra fast lenses to capture the maximal light in dimly lit scenes. For example, the late Stanley Kulbrick invested heavily in adapting the expensive and rare Zeiss Planar 50mm f0.7 to a modified camera to film the infamous candlelit scene in Barry Lyndon - fStop is as critical as in scientifically valid comparisons of optical lenses.

I'm not disputing f-stop as an indicator of exposure: light energy per unit area of the medium.

I'm simply calling attention to the role a lens's virtual entrance plays as the collector of light from a subject in the scene. It's total light energy that determines SNR and the resulting noise level in a photo.

 

[Bill Ferris]

For most all of my bird photography the Z100-400 S F4 is just so fitting for my needs. I can use either my 1.4 or 2.0 TC for additional range on my Z6iii . The idea of carrying the extra weight of a longer lens is less and less appealing as I approach 70 years of age. Sure if money was not an object I could invest in more glass but photography is still my hobby not what supports my income.

Thanks for sharing a great example of how a collection of factors come together to determine which lens best meets the unique needs of its owner.

 

[Luskentyre]

In terms of raw light gathering, the difference between the 400/2.8 and 600/4 is so minimal (10%) as to be negligible. I would imagine the f/2.8 is an advantage when fast, accurate focus in weak light is critical. And the 400's MFD advantage must also be an advantage for some subjects.

Thanks Bill. I understand the point you are making I.e. the 400 f2.8 would really be a 600 f4.2 & the 600 f4 would really be a 400 f2.7 if they were extrapolated by keeping their respective entrance pupil diameters (& therefore the slight advantage to the 600 by virtue of the larger entrance). However, it is worth noting that they have different use cases/MFD, and can only be directly compared if getting closer to the subject is not an option and/or have the same framing requirements. For instance, I am generally able to get as close to the subjects (within reason) I shoot with the 400 2.8 (deer/owl/fox) & therefore the 400 2.8 has a significant advantage vs the 600 towards the end of the day. However, I do agree that if one cannot get closer and/or is relying on cropping to the same FOV, then the 600 has a slight advantage vs the former. Not sure if I’m missing anything & more than happy to be corrected.

 

[SCoombs]

I am either badly misunderstanding what is being claimed here or what is being claimed here has got to simply be wrong, and this can be demonstrated with an extremely simple test.

Two quick scenes captured from my desk chair, each done at 500mm at f5.6 and at 200mm at f2.8. Shutter speed and ISO are kept the same across all photos. (Ignore the exposure bias of +.7: this was just there from the last time I was shooting outdoors, but for this test I just went to full manual and picked an ISO of 3200 and stuck with it). The 200mm shots were then cropped into be more or less the same size. I first did the water bottle, but then quickly thought maybe I should do something white because I planned on measuring the luminance of the target and thought it may be less finicky to pick an objective point to do this with off of a more uniform color and so I shot the UPC codes.

I ended up not taking the time to measure the luminance because very, very clearly there is more light on the targets with the 2.8 apertures and there is no need of looking at any numbers to prove this.

Now the entrance pupils here are about the same size as those given in several of the examples cited above: the 500mm at 5.6 is about 89mm, while the 200mm at 2.8 is about 71mm. As I understand the claim above, this should mean that there is more light on the subject with the 500mm lens and it would only be by looking at the full frame that we would see the same amount of light from the f2.8 lens. Yet very, very clearly this is not the case.

You can only see EXIF info for this image if you are logged in.	You can only see EXIF info for this image if you are logged in.
You can only see EXIF info for this image if you are logged in.	You can only see EXIF info for this image if you are logged in.

 

[bcgnz]

I am either badly misunderstanding what is being claimed here or what is being claimed here has got to simply be wrong, and this can be demonstrated with an extremely simple test.

Two quick scenes captured from my desk chair, each done at 500mm at f5.6 and at 200mm at f2.8. Shutter speed and ISO are kept the same across all photos. (Ignore the exposure bias of +.7: this was just there from the last time I was shooting outdoors, but for this test I just went to full manual and picked an ISO of 3200 and stuck with it). The 200mm shots were then cropped into be more or less the same size. I first did the water bottle, but then quickly thought maybe I should do something white because I planned on measuring the luminance of the target and thought it may be less finicky to pick an objective point to do this with off of a more uniform color and so I shot the UPC codes.

I ended up not taking the time to measure the luminance because very, very clearly there is more light on the targets with the 2.8 apertures and there is no need of looking at any numbers to prove this.

Now the entrance pupils here are about the same size as those given in several of the examples cited above: the 500mm at 5.6 is about 89mm, while the 200mm at 2.8 is about 71mm. As I understand the claim above, this should mean that there is more light on the subject with the 500mm lens and it would only be by looking at the full frame that we would see the same amount of light from the f2.8 lens. Yet very, very clearly this is not the case.

View attachment 110119	View attachment 110120
View attachment 110121	View attachment 110122

ISO also needs to be adjusted when you crop…

“In order to factor ISO into equivalence – thus fixing differences in brightness and noise performance – the equation is: (ISO) x (crop factor^2). As you can see, you need to square your crop factor first.”

Equivalence Also Includes Aperture and ISO

photographylife.com

 

[Bill Ferris]

The criterion of fStop - discussed as "Lens Speed" - is not misleading as a fundamental guide to selecting a telephoto for wildlife photography.

It's the comparison of f-stops to determine which of several lenses of different focal lengths will be best in low light that isn't reliable. If all are used to make the same photo, of the same subject, in the same light, from the same perspective, the fastest lens may not make the cleanest image.

A comparison of their respective entrance pupil sizes will reliably resolve that question.

Obviously, every photographer investing in expensive glass pays equal consideration to the Magnification(s) and field of view with respect to composition. This often includes pairing the telephoto with one of more Teleconverters to increase the magnification (with penalties on the Lens speed) Whether a prime or zoom, these criteria determine what can be described as the active focal length(s). [edited for clarity]

Regular readers of BCG discussions etc will be familiar with the overlapping exposure costs of Teleconverter-Pairings and Cropping to tighten the frame on a wildlife subject. Steve has published several lucid explanations about these subjects in his books and videos.

Whether one crops around the subject in post, in-camera (using a built-in crop mode or an actual crop format sensor) or adds a TC, the resulting image will be made with the same total light energy and have the same level of noise.

In summary, a teleconverter increases the magnification and narrows the FoV and it maintains Pixels/Duck. Cropping reduces Pixels/Duck and tightens FoV. Both solutions to fill the frame have a noise penalty, and often a noticeable impact on image quality. Hence, the wise rule to invest in telephoto(s) that minimize your cropping/pairing with TC's.

While a cropped image has more prominent noise than the uncropped original, the subject in the uncropped image encompasses the same total number of photons, same SNR, and same noise level as it does in the cropped or TC'd photo.

Cropping or pairing a lens with a TC doesn't add more noise to the subject. These practices make noise that was already present more visible.

Thus, Lens Speed, Field of View (FoV) and Magnification are the 3 critical variables to compare before investing in Telephotos. Obviously, together with image quality, the ergonomics, and especially costs, equally determine trade offs in decisions. In practice, or rather ideally, the choices require two or more lenses and/or zooms, which are complementary for different compositions and exposure challenges.

Of course, lens speed is an equation for calculating the size of the lens's virtual entrance pupil; its aperture. A comparison of apertures across a selection of lenses of different focal lengths, used on different format cameras, with or without TCs will reliably indicate low light performance.

 

[SCoombs]

ISO also needs to be adjusted when you crop…

Someone is going to have to explain that, but even granting it, I don't see how that could possibly change the takeaway here. After all, ISO has nothing to do with how much light is actually entering the lens: it's just an amplification of whatever light enters the lens. If what you're saying here is that we need to decrease the ISO when we use a wider aperture and crop, how is that not just acknowledging that the wider aperture is allowing more light in not just overall but also per unit area, since we need to turn that amplification down to get the same result?

 

[Luskentyre]

I am either badly misunderstanding what is being claimed here or what is being claimed here has got to simply be wrong, and this can be demonstrated with an extremely simple test.

Two quick scenes captured from my desk chair, each done at 500mm at f5.6 and at 200mm at f2.8. Shutter speed and ISO are kept the same across all photos. (Ignore the exposure bias of +.7: this was just there from the last time I was shooting outdoors, but for this test I just went to full manual and picked an ISO of 3200 and stuck with it). The 200mm shots were then cropped into be more or less the same size. I first did the water bottle, but then quickly thought maybe I should do something white because I planned on measuring the luminance of the target and thought it may be less finicky to pick an objective point to do this with off of a more uniform color and so I shot the UPC codes.

I ended up not taking the time to measure the luminance because very, very clearly there is more light on the targets with the 2.8 apertures and there is no need of looking at any numbers to prove this.

Now the entrance pupils here are about the same size as those given in several of the examples cited above: the 500mm at 5.6 is about 89mm, while the 200mm at 2.8 is about 71mm. As I understand the claim above, this should mean that there is more light on the subject with the 500mm lens and it would only be by looking at the full frame that we would see the same amount of light from the f2.8 lens. Yet very, very clearly this is not the case.

View attachment 110119	View attachment 110120
View attachment 110121

I am either badly misunderstanding what is being claimed here or what is being claimed here has got to simply be wrong, and this can be demonstrated with an extremely simple test.

Two quick scenes captured from my desk chair, each done at 500mm at f5.6 and at 200mm at f2.8. Shutter speed and ISO are kept the same across all photos. (Ignore the exposure bias of +.7: this was just there from the last time I was shooting outdoors, but for this test I just went to full manual and picked an ISO of 3200 and stuck with it). The 200mm shots were then cropped into be more or less the same size. I first did the water bottle, but then quickly thought maybe I should do something white because I planned on measuring the luminance of the target and thought it may be less finicky to pick an objective point to do this with off of a more uniform color and so I shot the UPC codes.

I ended up not taking the time to measure the luminance because very, very clearly there is more light on the targets with the 2.8 apertures and there is no need of looking at any numbers to prove this.

Now the entrance pupils here are about the same size as those given in several of the examples cited above: the 500mm at 5.6 is about 89mm, while the 200mm at 2.8 is about 71mm. As I understand the claim above, this should mean that there is more light on the subject with the 500mm lens and it would only be by looking at the full frame that we would see the same amount of light from the f2.8 lens. Yet very, very clearly this is not the case.

View attachment 110119	View attachment 110120
View attachment 110121	View attachment 110122

My (admittedly limited) understanding is that by cropping the 2.8 to match the 5.6, you have a) less pixels on the subject & b) more visible noise within the image. If you are happy with the above constraints, then 2.8 is the way to go...

 

[Bill Ferris]

I am either badly misunderstanding what is being claimed here or what is being claimed here has got to simply be wrong, and this can be demonstrated with an extremely simple test.

Two quick scenes captured from my desk chair, each done at 500mm at f5.6 and at 200mm at f2.8. Shutter speed and ISO are kept the same across all photos. (Ignore the exposure bias of +.7: this was just there from the last time I was shooting outdoors, but for this test I just went to full manual and picked an ISO of 3200 and stuck with it). The 200mm shots were then cropped into be more or less the same size. I first did the water bottle, but then quickly thought maybe I should do something white because I planned on measuring the luminance of the target and thought it may be less finicky to pick an objective point to do this with off of a more uniform color and so I shot the UPC codes.

I ended up not taking the time to measure the luminance because very, very clearly there is more light on the targets with the 2.8 apertures and there is no need of looking at any numbers to prove this.

Now the entrance pupils here are about the same size as those given in several of the examples cited above: the 500mm at 5.6 is about 89mm, while the 200mm at 2.8 is about 71mm. As I understand the claim above, this should mean that there is more light on the subject with the 500mm lens and it would only be by looking at the full frame that we would see the same amount of light from the f2.8 lens. Yet very, very clearly this is not the case.

View attachment 110119	View attachment 110120
View attachment 110121	View attachment 110122

If I'm understanding correctly, each set of photos was made in the same ambient light from the same distance and the photos made at 200mm were cropped to match the framing of the photos made at 500mm. Is that correct?

If so, the brighter images weren't made with more light, they're simply lighter due to the ISO used. The 500mm lens was working with a larger entrance pupil and captured 57% more light. If the images are processed to have the same lightness, the 500mm photos should have less noise.

 

[Tiago C]

Would love to see photo comparisons as examples from Bill

 

[SCoombs]

If I'm understanding correctly, each set of photos was made in the same ambient light from the same distance and the photos made at 200mm were cropped to match the framing of the photos made at 500mm. Is that correct?

If so, the brighter images weren't made with more light, they're simply lighter due to the ISO used. The 500mm lens was working with a larger entrance pupil and captured 57% more light. If the images are processed to have the same lightness, the 500mm photos should have less noise.

That is correct, but all the ISO is doing is amplifying the light that was captured by the camera. If the 500mm lens captured more light, then why is amplifying that light by the same amount not producing a brighter image?

If I stand next to a canon and record the sound of it firing and someone else stands a few hundred feet away and records the same sound and then we play that sound back and turn the volume dial to the same place, the sound coming from the recording made up close is going to sound louder because more sonic energy moved the microphone there than the microphone that was further away, and then we are applying the same amount of amplification to each.

I went ahead and decreased the exposure on the 2.8 image by two stops, which should give them the same lightness, and indeed by my eye it does (I also sampled the RGB values in a few places and found they are roughly equivalent). I then cropped each in close, keeping the crop the exact same size relative to the overall picture (i.e., the same number of MP) and put it on an area that was uniform in color and was as close as I could get to the same level of lightness and the f/5.6 shot clearly has more noise.

You can only see EXIF info for this image if you are logged in.

You can only see EXIF info for this image if you are logged in.

 

[Dinusaur]

I am with @SCoombs; to me the simplicity of his test results settles the score of which of the two lenses let more light in under the same or similar shooting parameters - crop or no crop.

 

[Bill Ferris]

That is correct, but all the ISO is doing is amplifying the light that was captured by the camera.

ISO does not "amplify light".

If the 500mm lens captured more light, then why is amplifying that light by the same amount not producing a brighter image?

ISO communicates a relationship between the target lightness (the selected ISO) and exposure to the camera. The camera processes the JPEG to give it a corresponding lightness. If you wanted the photos made with the f/2.8 lens to have the same overall lightness as the photos made with the f/5.6 lens, an ISO of 800 would have done the job.

If I stand next to a canon and record the sound of it firing and someone else stands a few hundred feet away and records the same sound and then we play that sound back and turn the volume dial to the same place, the sound coming from the recording made up close is going to sound louder because more sonic energy moved the microphone there than the microphone that was further away, and then we are applying the same amount of amplification to each.

I went ahead and decreased the exposure on the 2.8 image by two stops, which should give them the same lightness, and indeed by my eye it does (I also sampled the RGB values in a few places and found they are roughly equivalent). I then cropped each in close, keeping the crop the exact same size relative to the overall picture (i.e., the same number of MP) and put it on an area that was uniform in color and was as close as I could get to the same level of lightness and the f/5.6 shot clearly has more noise.

View attachment 110134

View attachment 110136

Those crops do not appear to be of the same areas in the photos. The details do not match.