Equivalence in f-stops Yes, NO or But...

[Tronhard]

I love a good debate, occasionally it becomes an argument, but as long as it stays clean and no one is harmed it can be fun. SO....

In discussing Equivalence, "at its most simple, is a way of comparing different formats (sensor sizes) on a common basis." from DP Review.com

In this article and other material I have read, the concept that to establish the equivalent performance of a lens on a non 35mm format, one would multiply the focal length(s) of a lens by the crop factor - the Focal Length Multiplier. This seems to be pretty much beaten ground and accepted.

Less so, it would appear, is the concept that one also multiples the valued for f-stop as listed on the lens. This seems to have evoked a fairly robust debate, so I am curious about what our esteemed members think of this. And no, I am not going to express an opinion on this one.

What I will say is that no-one seems to suggest that putting a 100mm f4 lens on a crop-body sensor will actually change the physical attributes of the lens or the light density it delivers to the sensor. The issue is about the Equivalence of the light across the sensor as a whole.

The sources I have seen listed on the subject are listed here, not so much to support any belief but to demonstrate the range of views on this subject that are available to anyone on the web

Northrop Photography on Youtube:
https://www.youtube.com/watch?v=DtDotqLx6nA

An article on DP Review on the specifics of Equivalence:
https://www.dpreview.com/articles/26...-should-i-care

Within the CiC materials:
https://www.cambridgeincolour.com/tu...ensor-size.htm

Other articles on the subject, but taking a different view:

http://www.discoverdigitalphotograph...aperture-mean/

http://admiringlight.com/blog/full-f...doesnt-matter/

I also noted that while he did not directly reference this issue, Prof Levoy in his Stanford lectures also used f-stop equivalence at several points in his lecture.

All that is bad enough, but at least it refers to numbers printed on the lens. Now consider those who want to do away with f-numbers and use T-stops...

SO, I am looking for definitive information, NOT just based on your experience, beliefs or whatever (with all due respect). I am looking for research or scientifically supported and referenced materials to answer this one way or another.

[Shadowman]

Trev,

At one time I was considering testing FX/DX, for noise specifically; and then read that the equivalent focal length would be different based on the sensor size/lens. Then I had to decide if the lens I currently owned would have equivalent focal lengths or would I have to use different lenses on each camera, then the issue of how equivalent in performance were the two lenses I'd use, would a wide open 1.8 be equivalent to a 3.5 aperture. Then I decided to just enjoy the cameras and use them when I felt they met my personal expectations.

[Tronhard]

Hi John:

On a PERSONAL basis I agree with you... I get the best gear I can afford and spend a lot of my time and effort getting the shots right - often with dubious results!

This post is in reaction to discussions I have come across in other forums where people who are supposed to be highly technical, experienced and highly skilled take grossly opposing positions on this. I have, out of curiosity, done my own research, but I want to see what others have got on the subject - but having heard a lot of unsubstantiated opinions I am looking for real hard facts.

Yes, I know I need a life - or in this case a working knee so I can go out and take more photos, but I can't right now so...

[dje]

Hi Trev

Personally I think the article by Richard Butler (your link 2) puts forward a good explanation. (Sorry to provide a personal view but I have to start somewhere!) The concepts of equivalence in focal length and equivalence in aperture in relation to depth of field are fairly convincing, and from what I've observed in a variety of articles on this topic, are not terribly contentious (as long as bokeh is not confused with DOF). There's not a lot of high level science involved really and practical demonstrations can be readily set up.

The idea of taking this one step further to include the equivalence of aperture in relation to sensitivity/ISO and noise is more complicated but the general concept seems valid as an approximate indication. Basically the concept is that with a larger sensor, you need a smaller aperture to get the same DOF and then with the smaller aperture you need higher ISO to get the same shutter speed. The higher noise associated with the higher ISO offsets the inherently lower noise characteristics of the larger sensor.

Sensor technology comes into noise performance too of course and pixel size to some extent, mainly if you are pixel peeping at 100%. If you want a reference on the effect of pixel size, please see here. There you go, a scientific reference.

And as Richard Butler mentions, variations in TStop (transmission loss)from one lens to another comes into the sensitivity equation too.

Dave

[george013]

I don't know exactly what you want. You're referring to discussions we didn't have. But if "The issue is about the Equivalence of the light across the sensor as a whole." is your question, than I would say the light needed to hit the sensor is determined by the angel of view and the aperture diameter. The f-number is a ratio between aperture diameter and the nominal focal length. As long as the angle of view and the f-number are the same, the amount of light hitting the sensor is the same. I'm not familiar with light, still struggling, but the explanation is in the units used for the light. I think

George

[Tronhard]

George,
If I may explain... I have been privy to some robust debates on Equivalence and in particular its association with F-stops. I am not referring to discussions on THIS forum that I am aware of, but I am starting the discussion using as much scientific methodology and integrity as possible.

Equivalence is a way of comparing the performance of camera components between different formats. Digital lenses have their focal lengths and f-values defined for a full-frame 35mm format. However there are now other common formats such as APS-C (1.5 and 1.6 crop) and micro 4/3 (which is a 2.0 crop).

As I said it is generally accepted that when considering the performance of a lens on a crop sensor, one uses a Focal Length Multiplier. Thus a lens identified as 100mm would actually provide a FoV equivalent to 150 or 160mm on an APS-C and 200 on a micro 4/3.

But what about the f-value, the other metric identified on the lens in terms of a 35mm format. What happens to that?
I have seen impassioned but unsubstantiated arguments for and against the use of the same multiplier for that. So to follow that logic, an f4 lens of a FF would have the effective performance of f4 x the crop factor. NOTE that no-one is suggesting that there is a physical change to the lenses, but rather on the effect of the crop on the delivery across the whole sensor. ALSO NOTE that equivalence does not deal in measures per mm, but rather on the whole surface of the sensor.

This is an interesting question when one looks at lenses intended for non FF cameras claiming low f-values for their lenses - some talk about 35mm equivalence for their focal lengths but not the f-values. Is that correct, or not?

SO dear readers... I have seen lots of opinions, but have seen VERY LITTLE HARD EVIDENCE backed up by scientific facts or process So the deal is PLEASE NO COMMENTS based only on:
Your opinion or Individual experience
Heresay
Vague references.

I am looking for hard, definitive facts or authoritative sources, e.g. academic material.

Dave has made some comments and bless his heart offered some referential material. That's heading in the right direction.

I will offer AS AN EXAMPLE some videos on Youtube associated with this issue as they not only address the issue but offer demonstrations and use identified sources for their conclusions. Are those conclusions correct? We'll see...
I would equally welcome similarly supported material to counter his conclusions.
Northrop:

Crop factor with ISO and Aperture
https://www.youtube.com/watch?v=DtDotqLx6nA
Crop Factor: Why you multiply the aperture by the crop factor when comparing lenses
https://www.youtube.com/watch?v=f5zN6NVx-hY
Crop Factor part 3: Crop Factor Part 3: Responding to Critics, Corrections
https://www.youtube.com/watch?v=6Im4W_9blhY
Crop Factor 4: Debate with a Critic & New Sensor Performance Data
https://www.youtube.com/watch?v=0OtIiwbAZi8

If nothing else we may all learn something!

[george013]

That will be a chaotic discussion. I think you've to narrow the question to start with. Do you want to compare the optical issues or the photographical issues, when you can make a distinction between them. And even then with the same framing, the same subject distance, the same aov, the same viewing enlargement, viewing size or just at 100% and of course the dof.

I must watch the movie later.

George

[dem]

I think that people who bang on about an F/2.8 being always an F/2.8 no matter what sensor is put behind it (1) do not understand how field of view and DoF work (2) insist on comparing all camera formats at the same ISO (3) pick the definition of exposure that covers only the amount of light falling on the sensor but has nothing to do with image brightness (the ISO is considered to be NOT a part of exposure triangle).

I am not saying they are wrong, but the only conclusion that comes out from this line of argument is that ISO sensitivity should be independent of sensor size. Nothing about framing, DoF, diffraction softening or noise in the final image printed on an A4 sheet.

If one wants to replicate ALL OF THE ABOVE using different camera formats, this is what equivalence is, one needs to adjust 3 things at the same time: focal length, aperture and ISO.

IMHO, the most comprehensive and rigorous source on this subject is
http://www.josephjamesphotography.com/equivalence/

[george013]

Looks good. Another 97 pages to print.

George

[Tronhard]

Thanks Dem, I shall read and digest

I would be interested in any comments on the Youtube links I gave at the bottom of my last post.

[pnodrog]

I watched the first youtube clip until the end of the ISO and got fed up with the inaccurate and over simplified explanation of ISO and noise in regard to sensor size. To completely ignore different sensor technologies make his conclusions meaningless.

Good luck with your research.

[Tronhard]

Hi Paul: Thank you for your comments!

Ha I saw the same issue, as did some others, but if you persevere you will find that they are addressed in later videos, along with a lot more detail - yep, I had to be patient.

I watched the first youtube clip until the end of the ISO and got fed up with the inaccurate and over simplified explanation of ISO and noise in regard to sensor size. To completely ignore different sensor technologies make his conclusions meaningless.

Good luck with your research.

[Dave Humphries]

Hi Trev,

I had this discussion with someone much cleverer than I when I was fairly new here (so 2008-9), but the CiC Search won't go back far enough to enable me to locate any thing that old. It proved beyond doubt to me that the cf multiplier should apply to aperture as far as DoF is concerned - because that varies with the difference in magnification of the (whole sensor) image between differing sensor sizes.

I haven't looked at any of your links Trev (sorry) because I am "a believer" as a consequence of my earlier discussions and don't need to revisit it again. Life's too short - or my brain's too full - take your pick

IOW for say; portraiture, if I shoot at f/2 on a 50mm lens on my DX camera (D7100), that'll be roughly the same DoF and angle of view as someone shooting with a FF sensor (e.g. D810) at f/2.8 with a 75mm lens (i.e. it's likely not the same lens, unless it is a fast zoom with sufficient range).

Cheers, Dave

[William W]

Trev,

Dem beat me.

I use Joseph James' paper as reference material.

WW

[Manfred M]

Joseph James paper is very thorough and in my view very accurate. On the opposite side of the spectrum is the Tony Northrup video which is really not very good at all, and like much of the stuff that Tony turns out, he has a lot of inaccuracies and errors in his tutorials. I try to steer people away from him.

[MrB]

Dem, thank you for posting the link to the Joseph James article - very interesting and helpful.

Cheers.
Philip

[Tronhard]

Hi Manfred!

It is an example of the range of materials on this subject and helps you understand why I am looking for authoritative sources. I have actually written to Prof Levoy to see what he has got on the specifics of this subject, but I suspect a response will take a while!

On the other forum I got responses that said in equivalence the F-stop effect is in no way alerted based on their "years of photographic experience", which is nice but... on either side of the argument, in seeking documentation I am taking a "put up or shut up" stance, everything else is heresay...

Joseph James paper is very thorough and in my view very accurate. On the opposite side of the spectrum is the Tony Northrup video which is really not very good at all, and like much of the stuff that Tony turns out, he has a lot of inaccuracies and errors in his tutorials. I try to steer people away from him.

[William W]

. . . On the other forum I got responses that said in equivalence the F-stop effect is in no way alerted based on their "years of photographic experience", which is nice but... on either side of the argument, in seeking documentation I am taking a "put up or shut up" stance, everything else is heresay...

I used to do a practical exercise with general Students and sometimes Portrait Classes to show the effect (or lack of) F/Stop and its relationship to DoF. The exercise had a few aspects, one was to dispel the oft held mantra that one "needs" extremely fast lenses (e.g. F/1.4~1.8 ish) to attain extremely shallow DoF for Portraiture . . .

The part of the exercise that you might be interested in, is what I termed the "tight head shoot out":

> I'd get two cameras, one APS-C Format (Canon) and one 135 Format;
> put a 50mm Lens on the APC-C Format and and 85mm lens on the 135 Format;
> get the two photographers opposite each other and each to frame a Tight Head Shot of the other
> get them each to make the shot, using the same F/Stop (typically F/4)
> get the photographer using the APS-C Format to open up to F/2.5 (one and one third Stops)

> visually evaluate and compare the DoF for each PAIR of shots (i.e the "F/4 - F/4 Pair" and the "F/2.5 - F/4 Pair").

WW

[pnodrog]

Is it really worthwhile pursuing?

Each camera format and even model has it's own characteristics. Those that know and understand photography are going to be aware of the differences and those that don't are just going to get confused about the relevance of all the cross comparisons. Much of the requirement and arguments seem to stem from trying to protect people from claims in advertisements. Good luck with that.

When buying a vehicle we do not look for equivalence values between types. Is a 5 ton truck equivalent to 2 x SUV or 15 x motorbike? Unless you understand the different options equivalence is meaningless and if you do understand the differences of very little benefit. For arguments sake I am taking an extreme view and using an extreme example.

The mere fact that f-stop equivalence generates so much argument among experts makes it totally useless as a comparison tool for novices.

[Tronhard]

Hi Bill!

Thank you for your contribution. Yes, I am a great believer in practical demonstrations (seeing is usually believing) and one of the reasons I offered the much-maligned Northrop videos was because, flawed as they may be, they make some kind of effort to show the effects he is trying to express in practical terms.

I think a lot of the debate about f-stops in equivalence comes from a fundamental difference of understanding of what an f-stop is in terms of delivering light from the lens and what equivalence is trying to say regarding light gathering by the sensor. Much of that seems to boil down to an argument between light intensity per-unit-area, vs. light gathered by the whole sensor's area.

In that case people will talk past one another and the debate will rage forever!