• Please use real names.

    Greetings to all who have registered to OPF and those guests taking a look around. Please use real names. Registrations with fictitious names will not be processed. REAL NAMES ONLY will be processed

    Firstname Lastname

    Register

    We are a courteous and supportive community. No need to hide behind an alia. If you have a genuine need for privacy/secrecy then let me know!
  • Welcome to the new site. Here's a thread about the update where you can post your feedback, ask questions or spot those nasty bugs!

A shock about lens transmission

Doug Kerr

Doug Kerr

Well-known member
By the transmission of a lens we mean the fraction of the light captured by its entrance pupil that is actually delivered to the focal plane. This primarily recognizes that fact that light is lost by reflection at each air-to-glass or glass-to-glass interface, and a bit by actual absorption in the elements.

Based on some of the (made long ago) assumptions in the equations for exposure metering standards, I had come to think that perhaps a value of 90% was to be expected in a typical case - maybe even less, given modern advances in lens coatings.

This morning, in connection with some measurements I was taking, I needed to get some idea of the transmission of the lens I was using (a Canon EF 24-105 mm f/4L IS).

I found that DxO Labs nicely determines this in their modern DxOMark lens testing suites, so I checked their finding for that lens.

I was shocked to learn that at the maximum focal length (where the transmission is highest), the value was about 62%. At the minimum focal length, it was about 58%.

Note that, with through-the-lens exposure metering systems, this does not make the automatically-determined exposure incorrect. But it does mean, that for any given scene luminance and exposure index, the exposure recommendation is over 1/2 stop greater than for a lens with transmission of, for example, 90%.

Looking at it another way, this f/4.0 lens has a T-stop value of about T5.1. That is, its effect on exposure is as if its maximum aperture were f/5.1.

Best regards,

Doug
 
B

Bart_van_der_Wolf

pro member
Tom dinning said:
Should we panic now, Doug, or wait for the jury to come out?
Click to expand...

Hi Tom,

No need for panic, but feel free to do so if you like.

There are two sides to the light-loss issue. The first is obvious, we need a longer exposure time to achieve the same amount of photons to register as image data. That may be bad news if one is limited by exposure time due to subject motion or risk of camera shake. So, for specific applications it may be useful to compare lenses and also base one's choice on the T-stop differences.

The second issue is that that percentage of light is (mostly) not really lost. Instead, it bounces around between lens groups and elements, and will add to a lower image contrast by what's known as "veiling glare". Veiling glare will add a locally fixed amount of diffused light to the actual image forming rays, and thus (as a percentage) mostly hurt the shadow areas of our images, by loss of contrast and ability to discriminate details (even with attempted software fixes).

Only the light that is reflected by the first air-to-lens surface is reflected back into the scene, and is lost for image forming. The other rays that do make it through, may be lost by internal reflections (they never leave the lens element itself), or get absorbed by the edge mounts of the lens elements (but for that the lens element edges need to be made black). A tiny amount is lost by absorption of the lens element itself (depends on its crystal composition), and is somewhat wavelength dependent. Absorption will convert to heat or fluorescence.

I can understand that Doug is surprised. We know (some of us anyway) that all un-coated air-to-lens surfaces (that's 2 per lens or group) reflect something like 1.5% to 2% for each surface. Coating of lens surfaces can reduce those reflections to less than 0.5% each. But apparently there are lenses that perform much worse that optimal. Again, not the end of the world, but it usually won't help image quality either.

Cheers,
Bart
 
Tom dinning

Tom dinning

Registrant*
".... won't help image quality..."
Now I am panicking.
Here I was thinking I had total control over the quality of my pictures and it turns out its out of my hands.
I give up.
Is there a hobby, craft, art form I can take up that I can control the outcome?
 
B

Bart_van_der_Wolf

pro member
Tom dinning said:
".... won't help image quality..."
Now I am panicking.
Click to expand...

Hi Tom,

Again, don't panic .

don-t-panic-wallpaper-21-cool-hd.jpg

Here I was thinking I had total control over the quality of my pictures and it turns out its out of my hands.
Click to expand...

Welcome to reality, you don't have full control, but you may have options to choose from (assuming available budget). Knowing what to look for, which is what this thread is about, definitively helps.

I give up.
Is there a hobby, craft, art form I can take up that I can control the outcome?
Click to expand...

No, not if it involves physics. It's very hard to bend the laws of physics as we understand them. By better understanding them though, we might find a way around them, or accept our (own) limitations, and accept what we can control.

Cheers,
Bart
 
B

Bart_van_der_Wolf

pro member
Doug Kerr said:
Based on some of the (made long ago) assumptions in the equations for exposure metering standards, I had come to think that perhaps a value of 90% was to be expected in a typical case - maybe even less, given modern advances in lens coatings.

This morning, in connection with some measurements I was taking, I needed to get some idea of the transmission of the lens I was using (a Canon EF 24-105 mm f/4L IS).

I found that DxO Labs nicely determines this in their modern DxOMark lens testing suites, so I checked their finding for that lens.

I was shocked to learn that at the maximum focal length (where the transmission is highest), the value was about 62%. At the minimum focal length, it was about 58%.
Click to expand...

Hi Doug,

That's indeed a significant hit on the original scene data we try to record. This usually happens with lens designs that involve many lens elements/groups. That's why sometimes simple optical designs have high transmission and high contrast scores. On the other hand, more elements of special (rare-earth) composition are often required to manage complicated residual lens aberrations (typical in wide angle lenses and zooms).

The Zeiss Otus range of lenses is an example of state-of the-art complicated designs with many different lens materials (with different refractive index and color dispersion characteristics) and (a-spherical) shapes, where each element compensates for the errors of other elements, but also adds its own errors that need to be corrected with yet other elements, which ...

Here's a nice interview with one of the lead designers of Zeiss, Dr. Hubert Nasse, where some of these issues are addressed (relatively casually, fit for the intended audience).

Cheers,
Bart
 
Last edited:
Asher Kelman

Asher Kelman

OPF Owner/Editor-in-Chief
So Bart,

If I'm interested in getting the max contrast from a lens and am not concerned much about chromatic aberrations or some distortions that will generally be correctable in software, what are the ways to avoid lenses producing a vail of diffuse light over everything, losing us detail in the shadows?

Or does one go for lenses like the Otus or the sigma Art lenses which push the limits having different materials and even more elements to correct the errors introduced by the the last correction element?

(...and Tom, no we are not talking about putting a finger in a dam to save society, I'm just interested in simplicity v. the search for perfection and a major loss of light transmission!)

Really the big issue with old lenses for me is flare, otherwise most problems are correctable!

Asher
 
B

Bart_van_der_Wolf

pro member
Asher Kelman said:
So Bart,

If I'm interested in getting the max contrast from a lens and am not concerned much about chromatic aberrations or some distortions that will generally be correctable in software, what are the ways to avoid lenses producing a vail of diffuse light over everything, losing us detail in the shadows?
Click to expand...

Hi Asher,

While Chromatic Aberration will also somewhat affect contrast, to reduce veiling glare and improve transmittance, one would look for a lens design with the smallest number of air-to-lens element surfaces, and high quality coatings. A single group of lens elements consisting of kitted lens elements counts as having only 2 air-to-lens element surfaces.

There's another factor, but that is harder to derive from looking at a cross-section of a lens design, and that's the shape of lens elements. Digital cameras, unlike film, typically have a filter-stack (IR-filter and AA-filter) and a sensor cover-glass that could be relatively reflective, and thus reflect part of the projected image back to the lens. It depends on the AR-coating and shape of the rear lens surface if that reflection can wreak havoc (no-panic ;) ) by re-entering the lens or back into the mirrorbox where reflective surfaces may be exposed from the rear.

Or does one go for lenses like the Otus or the sigma Art lenses which push the limits having different materials and even more elements to correct the errors introduced by the the last correction element?
Click to expand...

These designs attempt to improve wide open aperture performance by correcting for all sorts of aberrations. Because that takes many lens elements/groups, One can assume they understand that high quality AR-coatings are essential (otherwise they would defeat the goal of wide aperture lenses, lots of light to be transmitted for exposure). Lenses with narrower wide-open aperture will require less aberration correction, and can be designed with fewer lens elements/groups.

Really the big issue with old lenses for me is flare, otherwise most problems are correctable!
Click to expand...

There is veiling glare (a locally diffuse light veil from unfocused light), and there is glare (more or less sharply outlined reflections from lightsources on the lens (or un-coated lens filter) surfaces, usually showing the aperture shape, but can also look like coma). The flare takes on the color of AR-coating filtered light, and thus changes color depending on angle of incidence.

They both lower the technical image quality. Veiling glare is a property of lens and construction design and coating quality and cannot be affected by the user. Flare on the other hand can be somewhat reduced by avoiding bright lights in the image' s field of view, and the use of deep lens hoods.

Cheers,
Bart
 
A

Arthur Haselden

New member
The best thing you can do is use a good prime lens. Primes usually have far fewer elements than zooms. They have to correct a much simpler range of options.
 
Asher Kelman

Asher Kelman

OPF Owner/Editor-in-Chief
Arthur Haselden said:
The best thing you can do is use a good prime lens. Primes usually have far fewer elements than zooms. They have to correct a much simpler range of options.
Click to expand...

Art,

Thanks for chiming in! glad to have you around! Perhaps with LF cameras, we have more control of getting rid of these unwanted reflections. BTW, if there's a scratch on a lens, one can just use a think magic marker or ink pain to fill it in. Main thing, it gets rid of the stray odd reflections and really the glass behaves as a good lens, once more!

Asher
 
M

Michael Nagel

Well-known member
I am surprised by the surprise this caused.
The fact that cinematographic lenses use T-Stop instead of f-Stop is intriguing enough.
The coating is there to increase contrast and transmission.
For a telecommunications engineer, attenuation within an optical fiber is nothing new though not relevant for photography...

Best regards,
Michael
 
You must log in or register to reply here.
Top