50mm APO-Summicron - reported aperture smaller than 50mm Summilux-ASPH?

[farnz]

Advertisement (gone after registration)

9 hours ago, IkarusJohn said:

I had understood that the original post suggested that the APO was 1/2 stop under exposing at f/2 compared to the ASPH.  That is a different issue to vignetting.  Or have I missed something?

I hope that Jaap will correct me if I've misunderstood him. I think that the point he was making that was that vignetting caused by the optical design of 'lens 1' will determine the amount of light reaching and reflecting off the white exposure spot on the shutter blades and if 'lens 2' produces different vignetting owing to a different optical design then a different Exposure Value could result despite the same f/stop etc having been used for both lenses.

Pete.

[Advertisement]

The Leica Q3 is available for pre-order at B&H Photo Video and Adorama

[Jared]

11 hours ago, jaapv said:

BTW this is the metering pattern of digital M cameras it is limited towards the edges by the FOV of the metering cell. This cannot be influenced by vignetting. 

I would add that Lindolfi is head of a university physics laboratory. 

I’ll look for that thread, thanks!

[lct]

The 50/2 apo has more vignetting at f/2 than the 50/1.4 asph v1 at the same aperture and the darkening it causes cannot change anything to the nominal aperture of the lens unless i'm missing some basics.

[wizard]

vor 7 Minuten schrieb lct:

The 50/2 apo has more vignetting at f/2 than the 50/1.4 asph v1 at the same aperture and the darkening it causes cannot change anything to the nominal aperture of the lens unless i'm missing some basics.

+1. And that is true for many other lenses, too. Vignetting is most pronounced at the maximum aperture, and decreases on stopping down. It is usually gone or at least no longer significant at around 3 stops down from full aperture.

Edited July 25, 2023 by wizard

[Jared]

12 hours ago, IkarusJohn said:

No profiles that I am aware of; but if there were, the problem is what exactly?  I certainly did nothing in LR to either image.

The discussion of t-stops is all very interesting, but we are not shooting video, and these lenses are not designed for it.  We use f/stops.  The two images, taken on the SL with the same ISO, same f/stop and same shutter speed, taken from a tripod (no lens hoods) using the Fotos App as a remote, then downloaded to LR and exported should show a real life difference.  If there was any.

At least, for me, using these lenses, I see no cause for concern.  When using my digital M cameras and my SL, I do use the in camera meter; with my M-A, I rely on what is on the lens barrel, so such variation is important, particularly when using slide film.

Profiles aren't a "problem" per se. They are generally a good thing. However, they potentially change the exposure of an image across much of the image frame. I don't happen to know how Lightroom handles 'M' lenses adapted to the SL, though, and don't have an SL to test. By default, Lightroom applies the lens profile to even raw images, but some Leica lens profiles have minimal corrections.

Profiles do two things: they correct distortion and they correct the lens' natural vignetting by dividing all areas of the image aside from the center by a fractional value, thus brightening the areas away from the center till you have an evenly illuminated result. For them to work well, Lightroom needs to know both the distance the lens was focused at and the aperture used. Obviously, in the case of 'M' lenses, neither of these is available. In the case of M bodies, the vignetting correction is therefore minimal (if any) since overcorrection would look unnatural. I'm not sure what happens with an M lens adapted to the SL. I'm sure the profile is applied by default, but it may do very little to the image.

Obviously, this will usually not matter much in terms of every day use, but it's good to know how they work since vignetting corrections increase noise away from the center of the frame, sometimes by a very significant amount. Lots of camera manufacturers are now leaning on profiles to make pretty large corrections since this allows for less expensive and/or more compact lens designs to perform like larger, more complex designs. The Leica Q series is a good example. Profiles correct a lot of distortion and vignetting on that 28mm Summilux. Leica 'M' lenses, by contrast, can't rely too heavily on profiles since Leica film cameras don't even know what lens is attached, and the digital cameras don't know the set aperture or subject distance.

In the context of this discussion--why the APO is metering half a stop "slower" than the 'Lux ASPH when both are at f/2... I really don't know. If Jaapv is correct (and I assume he is, though I haven't yet read through the link he provided) and a center spot is responsible for the entire EV calculation, then my suggestion that vignetting is the primary cause is problematic. While vignetting with the APO does start in the center of the field at f/2, it doesn't really become significant till you are about 10mm off axis based on Leica's graphs, so vignetting only works as an explanation if the meter reading is incorporating at least some data from the edges of the field. Traditionally, center weighted metering was just that--the center of the field is emphasized, but the entire field is used to calculate exposure. If in the case of the M10 the meter can't even see the edges of the field, then even pretty heavy vignetting would be ignored.

If that's the case, then there is something else going on. I doubt it is simple f/stop vs t/stop since that would require greater exposure on the APO at all apertures rather than just wide open. I'd like to take actual brightness measurements across different parts of the frame with the APO and a couple other lenses at a fixed shutter speed and ISO to see whether the meter is recommending an extra half EV of exposure but the actual values in the center are similar, or whether there really is a half EV of exposure difference and whether that is true only in the exact center or across different parts of the frame. At this point, at least for me, it's intellectual curiosity, not anything that would affect my usage. I've got about 8 years under my belt using the 50mm APO on both film bodies and digital M's, so I'm not worried about real-world behavior. Unfortunately, I'm on vacation at the moment so only have a couple lenses with me so I won't be able to satisfy my curiosity for a little while.

[wizard]

One more thing: the diameter of any aperture, e.g. f2, is purely a mathematical value calculated in relation to the focal length of the lens. So for a 50mm lens to have f2, the diameter of the aperture at f2 needs to be 25mm. If a 50mm lens has a 25mm diameter opening at full aperture, it is f2. However, the value of the aperture's diameter at full opening says nothing about how much light is actually transferred through the lens to the film or sensor. Rather, actual light transmission depends on the type of glass used, the number of glass elements in the lens, the coatings etc., which means that two f2/50mm lenses of different construction may well have a different light transmission, resulting in more or less light reaching the film or sensor at the same set aperture.

I doubt, however, that those differences between the Summilux asph. and the Summicron apo, if they exist, will account for 1/2 EV stop of exposure difference. In my view, this would only be possible if one lens is a modern lens and the other lens is a lens from, say, pre WW II.

Edited July 25, 2023 by wizard

[IkarusJohn]

Advertisement (gone after registration)

6 hours ago, farnz said:

I hope that Jaap will correct me if I've misunderstood him. I think that the point he was making that was that vignetting caused by the optical design of 'lens 1' will determine the amount of light reaching and reflecting off the white exposure spot on the shutter blades and if 'lens 2' produces different vignetting owing to a different optical design then a different Exposure Value could result despite the same f/stop etc having been used for both lenses.

Pete.

Yes, I understood Jaap’s point.  That’s why my sample images were taken in M setting, with the same ISO, aperture and shutter. Whatever the meter was deciding was therefore irrelevant, the only difference being the change in lenses.

If I had used an aperture priority setting, or auto ISO, then the meter might have played a part, but the images don’t actually show significant difference in vignetting.  It would also depend if I was using an M with metering off the shutter blades or live view off the sensor, to address Jaap’s point.

I agree that lens profiles in camera will adjust for vignetting, but aren’t we getting even more esoteric?  For there to be a noticeable difference, I would need to be using the APO at f/2, with lens identification turned off on the camera (and presumably in LR) and make a point of using aperture priority and metering off the shutter - so with an earlier M camera than M11, and not in live view.

Doesn’t that seem a little obscure, when the OP seemed to suggest that his APO images are darker than his ASPH images at f/2?

[jaapv]

7 hours ago, wizard said:

One more thing: the diameter of any aperture, e.g. f2, is purely a mathematical value calculated in relation to the focal length of the lens. So for a 50mm lens to have f2, the diameter of the aperture at f2 needs to be 25mm. If a 50mm lens has a 25mm diameter opening at full aperture, it is f2. However, the value of the aperture's diameter at full opening says nothing about how much light is actually transferred through the lens to the film or sensor. Rather, actual light transmission depends on the type of glass used, the number of glass elements in the lens, the coatings etc., which means that two f2/50mm lenses of different construction may well have a different light transmission, resulting in more or less light reaching the film or sensor at the same set aperture.

I doubt, however, that those differences between the Summilux asph. and the Summicron apo, if they exist, will account for 1/2 EV stop of exposure difference. In my view, this would only be possible if one lens is a modern lens and the other lens is a lens from, say, pre WW II.

The glass types in the lenses is bound to differ and the transmission of optical glass varies widely:

 https://www.edmundoptics.com/knowledge-center/application-notes/optics/optical-glass/

and we haven’t even started on coatings, reflection and dispersion. 

[Chazphoto]

Hmm, interesting reading.  I'd like to clarify a few things.

1.  I tried hard to make a distinction in my earlier posts between what I observed and what I thought caused what I was seeing.  I'm quite content to have the latter characterised as fantasy 🙂 (particularly someone as august as Jaap) and the discussion seems to have veered round to saying that vignetting cannot be the cause, because the meter wouldn't be influenced by it, which was certainly something I wasn't sure about.

2.  What I observed was not a darker picture with the APO, but a choice of a longer shutter speed, by half a stop (for the same ISO, same aperture setting and more or less the same EV of scene).  Other than testing the same thing more rigorously (a steadily lit image and a tripod) and more frequently (I only took a few exposures) I'm reasonably confident on the observed facts.  If anyone wants to argue about the observer influencing the observed, we can take that up in the philosophy class.

3.  Re IkarusJohn's test, is it correct that by fixing the exposure on both lenses, what you're showing is the absence of (much) observable vignetting on those two photographs?  It looks like a good way of testing one of the theories for why I observed the slower shutter speed with the APO, but it isn't the same test, is it?  

4.  More generally, it was (is) my understanding that f-stops are consistent across lenses (I can't recall who said something different above).  The point being that they are a fraction of the focal length of that lens, so if you were setting f4 on the 135mm Telyt, it should admit the same amount of light as f4 on the SEM21.

5.  At present, it does not seem as though we have a consensus explanation of why my observations happened (or would happen with all 50mm APO-Summicrons, if this wasn't just on the two lenses I looked at).  I think that we do have a general view that this is not really relevant in terms of usage of that lens.

Cheers

 

Chazphoto

 

[jaapv]

1 hour ago, Chazphoto said:

if you were setting f4 on the 135mm Telyt, it should admit the same amount of light as f4 on the SEM21.

 

That assumption is incorrect. It only denotes that the mathematical relation between the aperture and focal length is the same. It does not  tell us anything about the amount of light passing through the lens other than a more or less accurate derivative. 

[jaapv]

There is another consideration 

If the camera on automatic sets the shutter speed in half-stop increments ( not sure about the M10/11 here), a measured EV will set the exposure to the nearest shutter speed. However if it is inbetween even the smallest variation will tip the balance to the lower or higher speed, leading to a discrepancy of half a stop. 

[wizard]

vor 10 Stunden schrieb jaapv:

The glass types in the lenses is bound to differ and the transmission of optical glass varies widely:

 https://www.edmundoptics.com/knowledge-center/application-notes/optics/optical-glass/

and we haven’t even started on coatings, reflection and dispersion. 

I hear you, that's exactly what I have alluded to in my above post. Typically, however, light transmission in modern optical devices does not differ that much, see e.g. binoculars, where light transmission is usually indicated to be 90 to 95%, depending on the model considered. While 5% may seem significant to some, it is almost impossible to perceive with your eyes. Half a f-stop would be much more than just 5%.

[Chazphoto]

One minor correction and one comment more form me.

I wrote something above about the aperture not necessarily being the same as the max width of light admitted by the lens and it has been bugging me since.  A bit of checking suggests that I was wrong, so I wanted to correct that.  The f-stop is calculated from the apparent width of the iris that light passes through at the entrance pupil.  Apparently, that is not the same as its physical diameter, because the glass in front of the aperture blades can magnify or reduce the aperture.  

As a comment on Jaap's post:

8 hours ago, jaapv said:

If the camera on automatic sets the shutter speed in half-stop increments ( not sure about the M10/11 here), a measured EV will set the exposure to the nearest shutter speed. However if it is in between even the smallest variation will tip the balance to the lower or higher speed, leading to a discrepancy of half a stop. 

I assume that the M10 shutter is stepless, so the reported 1/360th (at f2 on the APO) is an approximation and the half-stop comparison to a reported 1/500th (at f2 on the Lux) may mean that the discrepancy between the f-stops is less than half a stop.  

Regards

 

Chazphoto

[jaapv]

49 minutes ago, wizard said:

I hear you, that's exactly what I have alluded to in my above post. Typically, however, light transmission in modern optical devices does not differ that much, see e.g. binoculars, where light transmission is usually indicated to be 90 to 95%, depending on the model considered. While 5% may seem significant to some, it is almost impossible to perceive with your eyes. Half a f-stop would be much more than just 5%.

Yet, in the article I linked to earlier, the author found half-stop differences in Nikon lenses. 

[wizard]

vor 3 Stunden schrieb jaapv:

Yet, in the article I linked to earlier, the author found half-stop differences in Nikon lenses. 

Was that the article on optical glass? Or the DXOMARK article? I checked both, but could not find any indication of half-stop differences in Nikon lenses in any of them. As I said before, if one lens is, say, 70 or 80 years old and an other lens is a current lens, half-stop differences may indeed be possible due to significantly differing light transmission capabilities, but when comparing modern lenses of the same focal length and basically the same optical construction, I severely doubt that those differences exist.

[jaapv]

Sorry, my mistake  

 

[wizard]

vor 1 Stunde schrieb jaapv:

Sorry, my mistake  

Ok, thanks. However, this guy merely explains the difference between f-stops and t-stops. f-stops are based on the above discussed mathematical relation between the focal length of any given lens and the maximum diameter of its aperture opening. t-stops are based on the actual light transmission capabilities of a lens. Since a certain amount of light is lost when light travels through an optical system, the maximum t-stop of any given lens will almost always be lower than its maximum f-stop. The difference between maximum f-stop and maximum t-stop of a given lens can amount to half an f-stop.

The above applies to all optical systems, and in particular to all lenses. So what the guy said is that when considering any single lens, there may be a difference of about half an f-stop between the mathematically derived f-stop value and the actual t-stop. He did not say that there is a difference of about half an f-stop when comparing different lenses from the same manufacturer, which is what the OP has stated. In fact, if the OP's findings were correct, then the Summilux 50 asph. would be more or less an f1.4 lens, whereas the Summicron 50 apo lens would effectively be an f2.4 lens (half an f-stop down from the indicated f2 equals 50% less light). And that is something I find hard to believe, the more so since the Summicron 50 apo is the more modern lens.

[jaapv]

He has a long list of lenses based on DXO with significant and varying differences between t  stop  and f stop  Both Canon and Nikon. He also compares similar lenses. Half a EV value is not unusual. He also blames reflection and coating like I mentioned before.  I fear it is not a matter of belief. 

[Advertisement]

Hallo,

Look here: Leica M

[wizard]

vor 10 Minuten schrieb jaapv:

I fear it is not a matter of belief. 

Well then, only Leica will know at this point. Or has somebody made t-stop measurements of both the Summilux 50 asph. and the Summicron 50 apo?

[jaapv]

Unfortunately DXO has measured just one M lens.