Testing lens transmitence

**Lowell Goudge** · 12-18-2022, 01:15 PM - **1 Like**

Curious thought, has anyone on the forum tested the light transmittance of a lens. I read a post today (I forget which one) saying that one lens had a full FStop difference in transmittance, so it set my mind to thinking, I often get the impression that some lenses seem to need more light than others, for any given condition, but this is more an impression than anything else.

In thinking about it, I had an idea, using a modification of the test I did many years ago to plot the green button exposure error of different camera bodies when using non A lenses.

In that test I would shoot at every aperture the southern block wall of my house, which gets full sun. I would then plot the grey scale value of the histograms vs aperture.

I could do the same test here, but add a control. For each test shot I would take a photo of the same part of the wall with a control lens and a different body. let’s say for sake of argument my M100/4 macro. Being a simple lens it should have very good transmission, but also by comparing the grey scale of the control shot with the macro, to every other lens at the same aperture I should also start to see differences that are largely due to the transmission factor of the other lens.

Any comments.

Thanks

Edit note, I suggest this as I do not really have a calibrated light source

pschlute · 12-18-2022, 02:00 PM

Originally posted by Lowell Goudge

Any comments.

My initial thought is that a camera is not perhaps the best device to test this with. Apart from the obvious 1/2 stop or 1/3 stop setting, the main issue is going to be the ISO rating. While I understand that Pentax cameras have always been very close in this regard, I have seen reports that different camera from other makes can vary up to one stop even within the same brand for the same ISO setting.

Regarding green button exposure with non-A lenses you have another issue. The meter itself can only read what is transmitted through the focus screen (unless using solely LV). Anyone who has owned a camera from the K10D era will tell you that the new brighter screen from that and similar models could be woefully inaccurate depending on the f-stop used, when using manual aperture lenses.

Personally I find that the adjustments available to us in the digital era make it pretty irrelevant if a lens tends to over or under expose based on the same f-stop setting.

**Lowell Goudge** · 12-18-2022, 02:51 PM - **1 Like**

Originally posted by pschlute

My initial thought is that a camera is not perhaps the best device to test this with. Apart from the obvious 1/2 stop or 1/3 stop setting, the main issue is going to be the ISO rating. While I understand that Pentax cameras have always been very close in this regard, I have seen reports that different camera from other makes can vary up to one stop even within the same brand for the same ISO setting.

Regarding green button exposure with non-A lenses you have another issue. The meter itself can only read what is transmitted through the focus screen (unless using solely LV). Anyone who has owned a camera from the K10D era will tell you that the new brighter screen from that and similar models could be woefully inaccurate depending on the f-stop used, when using manual aperture lenses.

Personally I find that the adjustments available to us in the digital era make it pretty irrelevant if a lens tends to over or under expose based on the same f-stop setting.

I agree with all your comments, this is why my intent would always be to use a single lens and body combo for the reference shot, and a second body always used to take the test shot.

The reason for two bodies is to ensure that the reference shot always produces an image at the same time, so any minor variations in lighting from time to time can be nulled out, and the second body would remove errors that result from camera to camera variation. I’m not looking for minor changes of a few values of greyscale, but whole stops, or at least 1/2 stops considering that a jpeg image has a linear range of about 5 1/5 stops centred around the 128 neutral value. I have measured this in the past

**UncleVanya** · 12-18-2022, 05:05 PM - **1 Like**

DXOMARK gives T stop information. You could use to aid your control testing this if any of your lenses are in their list.

Here’s a Sony 100-400 f4.5-5.6. T stops are T5.2-6.4. Sony FE 100-400mm F4.5-5.6 GM OSS - DXOMARK

The 55/1.8 ZA lens has a T-stop of T1.8 - essentially no difference from the f stop.

Many other FE primes are off by 0.1T stops.

Note the FE 100-400 is far from a full stop off in transmission. At f5.6 that would be about T7.8

JohnMc · 12-18-2022, 05:43 PM

Transmission or transmitence? Spectral light characteristics which will vary based on time and date/season atmospheric changes could potentially affect your outcome.

**GUB** · 12-18-2022, 07:23 PM

I took a shot the other day with a flash onto a white sheet to compare the tint of various lenses. Very rough and ready. But thought a more controlled setup could do the job here. (ie clean the dust of the lenses first!!). Also probably accurate only for the centre few lenses. Histograms are of the dark Tak 50mm 1.4 7 element (sn finishing 763) and the 8 element above it. In Gimp I don't know how to convert that reading to stops but there will be a way. Of course the difference would be half this - light going in and out of the tested item. Random idea I know but looks effective.

**Lowell Goudge** · 12-19-2022, 05:15 AM

On my photo editor (Corel PSP) the histogram gives you the average and median values, from 0-255 . 128 is where you would expect most cameras to meter. As I said, I performed a test quite a while back where I set the lens at one stop, and using the shutter speed, varied exposure both sides of neutral exposure. What I got was about 5 1/2 stops centred around 128 with about 45 grey scale per stop. Beyond this band the stop on both sides then went 15 , 7, 3 greyscale respectively

Since the idea is to work in the middle of the exposure range with the test, it is easy to convert difference in greyscale with stops (as long as your editor gives a numeric value

**.a.t.** · 12-22-2022, 12:48 PM

Originally posted by Lowell Goudge

Curious thought, has anyone on the forum tested the light transmittance of a lens.
...

Lowell, I'm not sure if this is what you're asking, but I've compared the wide open "brightness" of some lenses using GIMP histograms.

For a consistent light source, I use a blank, white screen on a computer monitor in a darkened room.

angerdan · 12-22-2022, 02:41 PM

Originally posted by Lowell Goudge

Edit note, I suggest this as I do not really have a calibrated light source

You don't need a calibrated light source - just a steady one.
To measure the t-stop of a lens, i'd recommend this link:
https://photo.stackexchange.com/a/70660

**Lowell Goudge** · 12-22-2022, 06:13 PM

Originally posted by angerdan

You don't need a calibrated light source - just a steady one.
To measure the t-stop of a lens, i'd recommend this link:
https://photo.stackexchange.com/a/70660

Since my source is natural daylight that is why I will use a second camera for a reference shot, to be able to adjust for changes. Artificial lights leave me a little suspicious

---------- Post added 12-22-22 at 08:15 PM ----------

Originally posted by .a.t.

Lowell, I'm not sure if this is what you're asking, but I've compared the wide open "brightness" of some lenses using GIMP histograms.

For a consistent light source, I use a blank, white screen on a computer monitor in a darkened room.

Perhaps, but some lenses have minimum focus distances that make this impractical.

12-18-2022, 01:15 PM - 1 Like	#1
Lowell Goudge Site Supporter Join Date: Jan 2007 Location: Toronto Photos: Gallery \| Albums Posts: 17,891	Testing lens transmitence Curious thought, has anyone on the forum tested the light transmittance of a lens. I read a post today (I forget which one) saying that one lens had a full FStop difference in transmittance, so it set my mind to thinking, I often get the impression that some lenses seem to need more light than others, for any given condition, but this is more an impression than anything else. In thinking about it, I had an idea, using a modification of the test I did many years ago to plot the green button exposure error of different camera bodies when using non A lenses. In that test I would shoot at every aperture the southern block wall of my house, which gets full sun. I would then plot the grey scale value of the histograms vs aperture. I could do the same test here, but add a control. For each test shot I would take a photo of the same part of the wall with a control lens and a different body. let’s say for sake of argument my M100/4 macro. Being a simple lens it should have very good transmission, but also by comparing the grey scale of the control shot with the macro, to every other lens at the same aperture I should also start to see differences that are largely due to the transmission factor of the other lens. Any comments. Thanks Edit note, I suggest this as I do not really have a calibrated light source

12-18-2022, 02:51 PM - 1 Like	#3
Lowell Goudge Site Supporter Join Date: Jan 2007 Location: Toronto Photos: Gallery \| Albums Posts: 17,891 Original Poster	Originally posted by pschlute My initial thought is that a camera is not perhaps the best device to test this with. Apart from the obvious 1/2 stop or 1/3 stop setting, the main issue is going to be the ISO rating. While I understand that Pentax cameras have always been very close in this regard, I have seen reports that different camera from other makes can vary up to one stop even within the same brand for the same ISO setting. Regarding green button exposure with non-A lenses you have another issue. The meter itself can only read what is transmitted through the focus screen (unless using solely LV). Anyone who has owned a camera from the K10D era will tell you that the new brighter screen from that and similar models could be woefully inaccurate depending on the f-stop used, when using manual aperture lenses. Personally I find that the adjustments available to us in the digital era make it pretty irrelevant if a lens tends to over or under expose based on the same f-stop setting. I agree with all your comments, this is why my intent would always be to use a single lens and body combo for the reference shot, and a second body always used to take the test shot. The reason for two bodies is to ensure that the reference shot always produces an image at the same time, so any minor variations in lighting from time to time can be nulled out, and the second body would remove errors that result from camera to camera variation. I’m not looking for minor changes of a few values of greyscale, but whole stops, or at least 1/2 stops considering that a jpeg image has a linear range of about 5 1/5 stops centred around the 128 neutral value. I have measured this in the past Last edited by Lowell Goudge; 12-18-2022 at 03:01 PM.

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
When you are testing a lens, remember that the lens is also testing you.	Digitalis	Pentax SLR Lens Discussion	46	01-29-2022 03:14 PM
Nature Testing out my new lens	rm2	Post Your Photos!	3	05-20-2021 12:21 PM
Nature Testing 1,2,3, Testing.	Tonytee	Photo Critique	7	07-05-2018 06:30 PM
On Lens Testing (Or, Is This 50mm Lens Any Good?)	carpents	Pentax SLR Lens Discussion	18	06-21-2007 05:45 AM

12-18-2022, 02:00 PM	#2
pschlute Moderator Loyal Site Supporter Join Date: Mar 2007 Location: Surrey, UK Photos: Gallery Posts: 8,216	Originally posted by Lowell Goudge Any comments. My initial thought is that a camera is not perhaps the best device to test this with. Apart from the obvious 1/2 stop or 1/3 stop setting, the main issue is going to be the ISO rating. While I understand that Pentax cameras have always been very close in this regard, I have seen reports that different camera from other makes can vary up to one stop even within the same brand for the same ISO setting. Regarding green button exposure with non-A lenses you have another issue. The meter itself can only read what is transmitted through the focus screen (unless using solely LV). Anyone who has owned a camera from the K10D era will tell you that the new brighter screen from that and similar models could be woefully inaccurate depending on the f-stop used, when using manual aperture lenses. Personally I find that the adjustments available to us in the digital era make it pretty irrelevant if a lens tends to over or under expose based on the same f-stop setting.

12-18-2022, 05:05 PM - 1 Like	#4
UncleVanya Loyal Site Supporter Join Date: Jul 2014 Photos: Gallery \| Albums Posts: 28,459	DXOMARK gives T stop information. You could use to aid your control testing this if any of your lenses are in their list. Here’s a Sony 100-400 f4.5-5.6. T stops are T5.2-6.4. Sony FE 100-400mm F4.5-5.6 GM OSS - DXOMARK The 55/1.8 ZA lens has a T-stop of T1.8 - essentially no difference from the f stop. Many other FE primes are off by 0.1T stops. Note the FE 100-400 is far from a full stop off in transmission. At f5.6 that would be about T7.8

12-18-2022, 05:43 PM	#5
JohnMc Pentaxian Join Date: Sep 2020 Posts: 658	Transmission or transmitence? Spectral light characteristics which will vary based on time and date/season atmospheric changes could potentially affect your outcome.

12-19-2022, 05:15 AM	#7
Lowell Goudge Site Supporter Join Date: Jan 2007 Location: Toronto Photos: Gallery \| Albums Posts: 17,891 Original Poster	On my photo editor (Corel PSP) the histogram gives you the average and median values, from 0-255 . 128 is where you would expect most cameras to meter. As I said, I performed a test quite a while back where I set the lens at one stop, and using the shutter speed, varied exposure both sides of neutral exposure. What I got was about 5 1/2 stops centred around 128 with about 45 grey scale per stop. Beyond this band the stop on both sides then went 15 , 7, 3 greyscale respectively Since the idea is to work in the middle of the exposure range with the test, it is easy to convert difference in greyscale with stops (as long as your editor gives a numeric value

12-22-2022, 12:48 PM	#8
.a.t. Site Supporter Join Date: Mar 2011 Location: yesterday Photos: Gallery Posts: 1,261	Originally posted by Lowell Goudge Curious thought, has anyone on the forum tested the light transmittance of a lens. ... Lowell, I'm not sure if this is what you're asking, but I've compared the wide open "brightness" of some lenses using GIMP histograms. For a consistent light source, I use a blank, white screen on a computer monitor in a darkened room.

12-22-2022, 02:41 PM	#9
angerdan Pentaxian Join Date: Feb 2015 Photos: Gallery \| Albums Posts: 2,643	Originally posted by Lowell Goudge Edit note, I suggest this as I do not really have a calibrated light source You don't need a calibrated light source - just a steady one. To measure the t-stop of a lens, i'd recommend this link: https://photo.stackexchange.com/a/70660

12-22-2022, 06:13 PM	#10
Lowell Goudge Site Supporter Join Date: Jan 2007 Location: Toronto Photos: Gallery \| Albums Posts: 17,891 Original Poster	Originally posted by angerdan You don't need a calibrated light source - just a steady one. To measure the t-stop of a lens, i'd recommend this link: https://photo.stackexchange.com/a/70660 Since my source is natural daylight that is why I will use a second camera for a reference shot, to be able to adjust for changes. Artificial lights leave me a little suspicious ---------- Post added 12-22-22 at 08:15 PM ---------- Originally posted by .a.t. Lowell, I'm not sure if this is what you're asking, but I've compared the wide open "brightness" of some lenses using GIMP histograms. For a consistent light source, I use a blank, white screen on a computer monitor in a darkened room. Perhaps, but some lenses have minimum focus distances that make this impractical.