Talk:Lens speed

Disputing absolute "fast" designation
I am disputing this previously bold paragraph:


 * Lenses with f-stops below f/2.8 are generally considered fast, while those with f-stops above f/2.8 would be considered slow.

Because it contradicts the opening paragraph:


 * The "speed" of a lens is a term referring to the maximum aperture of a given lens relative to its focal length. |my italics|

Also lens speed is not truly "qualitative", see

A History of the Photographic Lens, Rudolf Kingslake

for the mathematical definition involving the sine squared of the steepest axial image ray.

Caltrop 23:02, 8 October 2006 (UTC)

the term "aperture" means opening. if you are referring to the f-stop, you are referring to the relative aperture. all "f-stops" are relative to the focal length, which is why they are written as a mathematical equation "f/x". however, i will rewrite the opening paragraph as i can see how it might be misleading.

as for the idea that lens-speed is also quantitative, perhaps i should say that in general use lens speed is quantitative, since i've never met anybody who goes around referring to lens speed in terms of a trigonometric polynomial.

i wrote this as a general interest article so that people who are not well-versed in the science of optics (or the language of photography) could learn what most photographers mean when they use the term "lens speed". if you would like to go into detail about Kingslake's optical formulas, perhaps you should write a new article called "lens speed (optics)" and place a link to it in this article.

--Johnnydc 01:33, 9 October 2006 (UTC)

I like this article
I do like this article and I am glad you wrote it and are working on it.

Here's the thing, I agree with you that dragging mathematical optics into a general interest and clear article like this is not a good idea.

However, a statement like "less than 2.8 fast, greater than 2.8 slow" is so dependent on the configuration of the lens and the size of the film or sensor.

If one does not want to give pure math, at least it would be handy to give some examples showing the dependencies.

What is the simplest way to guide readers into understanding that F 2.8 is not the watershed for fast and slow lenses all on its own? F 2.8 for a 50 mm "normal" lens on a 35 mm camera with a film diagonal of 43 mm is not intrinsically fast or slow, but is fast relative to a maximum aperture of 5.6 for the same setup.

Lens speed is key in portrait depth of field, low light, and long lenses. I think that these issues need to be brought out.

Caltrop 03:35, 9 October 2006 (UTC)


 * It mostly works because faster than about f/2.8, lenses are harder to make, and the price starts to go up fast. That is mostly true over the range of focal lengths commonly used for photography, and not so dependent on image diagonal relative to focal length. An f/1.4 lens might cost about twice what an f/2.0 lens costs, though one rarely uses it at f/1.4.  Even more, the f/2.0 lens might be better at 2.0 than the f/1.4 lens at 2.0. Gah4 (talk) 17:19, 12 August 2016 (UTC)

Let's keep working together on this great article.
Again, thanks for your collegiality in working to improve this important photography article.

I have marked this paragraph as lacking a citation:


 * In this case, a lens with a maximum aperture of f/3.5 will be considered faster than one with an aperture of f/5.6, though in general both lenses would be considered slow. |snip|

Not because I really think that it needs a citation, but because in keeping with the opening para, and from the cited sources, and leaving out the math, lens speed is always relative, not absolute, and must be referenced to another lens at the same focal length.

My friendly admiration and kudos to the original author for getting this article off to a great start. This mutual collaboration is what makes WP so powerful and I appreciate your good will and flexibility. Caltrop 15:13, 9 October 2006 (UTC)

I plan to work on this article in the near future
As of right now I am swamped with work (the glories of Monday). look for more contribs from me towards the end of the week.

However, I will say that I have an idea to structure this article in terms of format. I agree with you that lens speed is a term relative to format size: while a 1.4 lens is perfectly reasonable for 35mm, it would be unheard of in 4x5, where 5.6 is considered "normal" and press lenses would be considered "fast". My original article was oriented more towards 35mm, and I think with regards to that it was quite good.

I think it would be a good idea to have seperate sections for 35mm/digital SLR, MF and LF. Right now I feel the article relies to heavily on the use of the word "relative" and that it would be confusing to a novice reader who isn't familiar with the full range of photographic formats available. However, I think using the wider-scope "relative" definition would be good as a transition into discussions of larger formats.

The other thing we have to keep in mind is that the article shouldn't be too long. After all, lens speed is not a particularly hard subject to grasp, and there's no need to write a book about it (not that I think it's too long right now, but with our combined contributions it could grow to an obscene size unless we restrain ourselves! :). So perhaps we could write a brief summary geared towards small format and then expand it into a general discussion of the theory behind it, moving into larger formats as that progresses.

And truth be told I would very much like to see a separate "lens speed (optics)" article, as I find the theory fascinating, though I feel ill-equipped to write it myself.

--Johnnydc 03:53, 10 October 2006 (UTC)


 * I agree. Yes you are right on all counts. We do not want a book here, and a maths (optics) article would be brilliant. And of course your point about formats, esp. 35mm is well stated. I still think mostly in 35mm terms myself. As you point out, digital, medium format etc. may need separate sections while still keeping a concise article. Yes, lens speed should be an easy idea to grasp, but in the toils of lens lust whilst reading semi-er0tic sales brochures, it is also easy to become lost in the rhetoric :) -- this is why this excellent article caught my attention. Thanks for starting it so beautifully; one of the clearest photography articles on WP. Caltrop 11:55, 12 October 2006 (UTC)


 * What would you want in such an article that is not already covered in the articles on aperture, f-number, and depth of field? Dicklyon 14:47, 12 October 2006 (UTC)


 * Saw your edit, thanks for the catch! BTW which article are you referring to in your query above, JD's proposed Lens Speed (optics) or the existing article? I think JD's idea might be to show the maths and theory separately to keep the existing article short and less technical. But I will let him respond for himself. Just dropping in my thoughts :) Caltrop 16:21, 12 October 2006 (UTC)


 * I was referring to the new proposed article. I don't see the need for it, since I don't know what is thought to be missing. Dicklyon 16:46, 12 October 2006 (UTC)

Dear Dick
I can tell from your rather prickly style and controlling attitude that I will be adding you shortly to my Little List. Caltrop 17:09, 12 October 2006 (UTC)


 * Is that because I reformatted replies to not be new sections? Or something I did to the article content?  Not to disagree with your observations, just could use more specifics. Dicklyon 20:21, 12 October 2006 (UTC)

Merge (and pet peeve)
Why do people insert merge tags with a link to a discussion page, and then not say something to start the discussion? Weird. I see this a lot. Dicklyon 03:35, 13 October 2006 (UTC)

Anyway, someone proposed to merge fast lens into lens speed.


 * Support — a no-brainer; this is one guy's unique contribution, not worthy of a separate article just because he didn't notice lens speed Dicklyon 03:35, 13 October 2006 (UTC)


 * Support - I support this merger too. It is more of a stub than a full article, and the lens speed article makes it redundant.--Johnnydc 04:40, 13 October 2006 (UTC)

Thoughts on this discussion so far
re: "What would you want in such an article that is not already covered in the articles on aperture, f-number, and depth of field?"

The reason I wrote a seperate article is because alot of people new to photography do not understand what "lens speed" is at all. They are not aware that it is related to aperture, f-number or depth-of-field. Furthermore, while those concepts are basically scientific in nature, "lens speed" is more of a human concept, and deserves to be explained on it's own terms.--Johnnydc 04:45, 13 October 2006 (UTC)


 * Did I misunderstand? I thought you were proposing yet another new article beyond this one. Dicklyon 04:38, 13 October 2006 (UTC)

I was proposing a new article beyond the basic "lens speed" article to discuss the optical concept of lens speed, which is different than the photographic concept of lens speed, which I confess I know nothing about. Cal seemed to think this concept was worthy of mention, and I defer to him with regards to the creation of the new article.--Johnnydc 04:45, 13 October 2006 (UTC)


 * Since he deferred to you above, I'll just forget about it until/unless it comes up again. I think most of what you're interested in is already covered in the articles I mentioned. Dicklyon 04:50, 13 October 2006 (UTC)


 * I agree. I think that the technical aspects of lens speed (from an Optics perspective) are already well covered elsewhere. It may be beneficial to move some photography-specific material from those articles to this one but at all costs keep this article low on technical and mathematical detail. I have moved the list of fast lenses from f-number to this article. It is a much better fit here, both in tone and in intended readership.--Srleffler 05:39, 10 November 2006 (UTC)

Removal/change of passage (1911 britannica)
I propose eliminating or changing the section that makes references to 1911 Britannica article, as I fail to see how it illustrates advances in optical technology. While it's true that f/6 would be considered slow in comparison to today's 35mm cameras, most cameras in 1911 were large plate cameras. Even today, an aperture of f/5.6 (less than 1/3-stop brighter than f/6.0) is considered the "borderline" between fast and slow lenses in large format.

It is my understanding however that early "rapid" lenses suffered from severe distortion, vignetting and aberrations. The introduction of the "rapid rectilinear" lens, free of distortion, at the turn of the century was something of a photographic coup. So perhaps this should be included in the discussion.

I have declined to delete the section myself because I think the author should be able to respond first, as I appreciate the effort required to research and cite such a passage.

--Johnnydc 04:55, 13 October 2006 (UTC)


 * Well, I think the author is long gone, but since I added that bit to the article I guess I'm the editor you mean. I sort of like it, because I always to see historical context to aid people in interpreting current practice.  But I agree with your point that recasting it as at least partially a format dependence would make sense.  You are right that it was very hard to make lenses as fast as f/5.6 before the turn of the century.  See for example: Sir William de Wiveleslie Abney, A Treatise on Photography, 10th edition, Longmans, Green, and Co., London, 1905, which says "...with the introduction of Jena glass, some lenses of good definition can be made with apertures as large as f/5.6."  I do not find this in the 1st or 5th editions of some years earlier.  Dallmeyer's 1874 lens booklet has rapid rectilinears already, and I find a reference to one he made in 1870 with 30-inch focus and 4-inch diameter (f/7.5); that was rapid for it's time (and format, as you point out).  He mentions a portrait lens of f/3, but I don't think that qualifies for what Abney calls "of good definition".  Anyway, the history is complicated and interesting.  Take a whack at it if you like, or maybe I will. Rudolf Kingslake's History of the Photographic Lens is the real authority on this stuff, but my copy walked away. Dicklyon 06:17, 13 October 2006 (UTC)

Actually, the rapid rectillinear followed quick on the heels of the Cooke triplet. A typical rapid rectillinear lens is basically two triplets placed symmetrically on either side of the diaphragm, similar to a convertible lens. I think it might also be important to note that there were quite a few "fast" lens designs that have existed for nearly 100 years, but were impractical until the advent of anti-reflex coating because of the high number of air-glass surfaces. The Zeiss Planar is the best example I can think of off the top of my head. IIRC, Paul Rudolph actually designed the Planar in the late 1890's. I have to find the source for this before I write it in the main article though. I seem to remember it was in a book or article about Mr. Rudolph himself.

--Johnnydc 07:13, 13 October 2006 (UTC)

Extreme lens
Some more extreme lens:


 * GOI CV-catadioptric lens 20mm f/0,5
 * Zeiss Super-Q Gigantar f/0.33 40mm

It would be helpful to rearrange the fast lens list in order of appearance, with more comments on the particular qualities of it. For example, the 50mm f/0.95 in the Canon 7 Rangefinder 35 mm camera was considered an advanced achievement because it was available from the early 1970's and was widely distributed, although it was not particularly sharp or well corrected lens, but it is considered a milestone lens. amclaussen. —Preceding unsigned comment added by 192.100.180.19 (talk) 20:09, 10 May 2011 (UTC)

I've read that "gigantar" lens is a hoax. —Preceding unsigned comment added by 201.210.191.163 (talk) 20:30, 23 May 2011 (UTC)

Question about lens speed marking inscribed on lens
I have a question about the Canon lens in this article. The lens speed is given as 1:1.4. I know that 1.4 refers to the maximum aperture of the lens, but I don't know what the 1: before the lens speed refers to. The lens speed is always preceded by a 1:. What does the 1: refer to?

Anthony22 (talk) 18:44, 1 February 2013 (UTC)


 * The colon ":" is a standard mathematical notation for a ratio. In this case if you have a lens of 50mm focal length, and an aperture of 25mm diameter, the ratio of aperture to focal length is 25mm:50mm or 1:2.
 * --195.137.93.171 (talk) 21:12, 12 October 2013 (UTC)

It's also worth noting that in German, : is a symbol for division, so you could also read 1:1.4 as 1/1.4. Groogle (talk) 00:44, 8 December 2020 (UTC)
 * Fractions (division) and ratio are closely related. Note that rational comes from the ability to be a ratio (of integers). The use, at least in many countries, though, is different from a division operator. Gah4 (talk) 02:30, 8 December 2020 (UTC)
 * Fractions (division) and ratio are closely related. Note that rational comes from the ability to be a ratio (of integers). The use, at least in many countries, though, is different from a division operator. Gah4 (talk) 02:30, 8 December 2020 (UTC)

T-stops
I find it odd that the marking of cine lenses for transmission values or T-stop isn't mentioned here. It is more realistic to measure the brightness of a lens, than just its shape - including colouration of the glass, and light lost to reflections at the surfaces. --195.137.93.171 (talk) 21:12, 12 October 2013 (UTC):

For still photography, it is usual to use f/stops and not t/stops. With today's through-the-lens metering, the exposure will be right, independent of the transmission factor. For comparing lenses of similar construction, (and quality of anti-reflection coating), the difference factors out. As I understand it, it is necessary in cinematography so that there isn't a noticeable change in brightness with scene and lens change. Gah4 (talk) 17:31, 12 August 2016 (UTC)

Baker-Nunn
Unusual but f0.75 lens: Schmidt_camera

©Geni (talk) 20:19, 19 May 2015 (UTC)

Maximum possible speed
The theoretical maximum possible speed for a lens surrounded by air is f/0.5. I have no reference handy, but it can be easily derived from the conservation of etendue (see also the article Numerical aperture).

It is possible for an in-air lens to have an f-number smaller than 0.5 if the lens is macro-only. In this case, the 0.5 limit applies to the effective f-number, not to the plain f-number. But it just happens that, in the macro realm, only the effective f-number correctly assesses the lens speed (i.e. its light gathering power). I would guess that all the lenses listed here as “Reproduction lens” fall in this category, and are not nearly as fast as their (uncorrected) f-numbers suggest.

There is also an 81mm f/0.38 listed in the article. From the optical diagram in the provided reference, it appears that the image plane is coincident with the surface of the last element. This means that the lens has to be used with the film emulsion in direct contact with the glass. Which is expected, as any air gap between the glass and the emulsion would bring with it the f/0.5 limit and loose all the extra speed to total internal reflection.

I think the 0.5 limit should be mentioned in the article, otherwise it looks like you could have arbitrarily fast lenses. The f/0.38 example is really interesting, but most readers are probably unaware that such a lens could never be adapted to a “regular” camera, as its speed requires the film emulsion to be in direct contact the glass.

— Edgar.bonet (talk) 13:10, 10 September 2015 (UTC)


 * I added a new section with most of this material. — Edgar.bonet (talk) 13:14, 15 October 2015 (UTC)

Maybe this Document "The theory of diffraction-limited resolution in microparticle image velocimetry" by Carl D Meinhart and Steven T Wereley can give some asistence for solving this problem. — Preceding unsigned comment added by 92.193.47.151 (talk) 22:47, 5 May 2016 (UTC)

The argument is pretty simple (and the page has a lot of factual errors):

1. Lenses are characterized in terms of effective f-number: how much light they gather. 2. If lenses gathered infinite light, they would passively heat up, breaking thermodynamics. 3. Therefore, you can't go beyond a certain f-stop. I've always heard that was f/0.5, which sounds right, but I'd need to work through the math.

It's worth noting that at sane apertures (e.g. f/1.4 or more), you have equivalence:

1. Aperture-diameter-to-focal-length ratio 2. Light gathering 3. Size of bokeh balls / DoF

Note that all of these could have been chosen to define f-number. #2 was actually chosen. All relate in a simple way at sensible apertures. The approximation, which makes sense beyond f/1 or f/1.4 or so, is that the aperture is all approximately the same distance from the sensor. If that's the case, you have sensor equivalence too (an MFT f/2.8 25mm gives identical images to FF f/5.6 50mm). It's a really good approximation.

It breaks down for extremely wide apertures. If the sensor is 1mm, and the aperture is a mile wide, some light travels 1mm to get to the aperture, and some travels 1 mile. Enlarging the aperture has almost no further benefit for gathering more light, since it's far away from the sensor. Or it adds a small amount of angle to the sensor's field of view. An abstract 10-mile wide sensor wouldn't suffer from this issue.

So you're stuck with a limit. At best, the sensor can gather light from a 180 degree field of view, which translates to light gathering of (if I recall correctly) f/0.5.

73.17.150.215 (talk) 14:48, 18 September 2020 (UTC)

Zoom
The range of lenses considered "fast" has evolved to lower f-numbers over the years seems not to follow the trend toward the use of zoom lenses. With more sensitive digital sensors, there is more trend toward slower lenses, as it is much harder to make an equally fast zoom lens. Gah4 (talk) 14:26, 23 November 2015 (UTC)