Lenses focal length

[Bjørn Rørslett]

My reply wasn't directed towards you. I spoke on a general note. I also note that people rarely if ever speak of the reasons why this topic should be so important.

A horse can be beaten to death repeatedly, but the efforts are pretty much wasted after the first round.

[simsurace]

Ok, I do not take it personally but I felt addressed because I had used the word 'same' a couple of times.

My view is that the topic of format comparisons is important is because many people do compare formats for all kinds of reasons, and some of them strive to do it in a rational manner. Often, the comparison is done in an either/or way because the person doing it can afford to put money in one and only one system, and hence has to choose a format over every other. The person perhaps wants to know the potential implications of this choice. They want to know whether system A allows them to do the same or similar as system B, or whether there are limitations to system A that do not affect system B. I do not judge anyone for having these thoughts -- I simply observe that they occur. I think the best we can do is to provide a framework to aid in these decisions, rather than to assert that the questions are invalid/unimportant.

Even to those who only ever use one camera system with one sensor size, the topic is important to understand the implications of cropping.

People who understand the basics can use them to choose the right lenses within their system.

I am sure other people could list more motivations here.

Anyway, the assertions that we make regarding relationships between different formats (e.g. the one in my post above or the ones in the treatise of Joseph James) are either true or false regardless of the importance of their truth or falsehood for individual photographers (artistically speaking). Thus we should be able to reach a consensus within this educated community about what is true and what is false. I think we can only profit if everyone participating in the discussion explains in detail why he/she thinks that a given assertion is false.

[Anthony]

As far as I can see, this topic is not about photography specifically. It is about optics and digital sensors. Photography just so happens to make use of these tools to communicate a message. There are great artists who do everything intuitively and don't care about the science behind it. But that's beside the point.
There is no artistic aspect to this topic as far as I can see. It is purely technical/scientific.
It is only because I see that we are lacking a consensus on the technical matters that I am bothering to write anything at all.
If what I write is not helpful or does not stimulate any thoughts/discussions, it is quite easy for me to quit doing it. As you suggest, I have plenty of other useful stuff to do with my time.

Please continue.  I find your comments helpful.  As there is no definitive statement on this thread, I remain somewhat confused.

I think it is important to understand the effect of different sensors on the final image.

To give a real world example.  When I go out on the streets, I like the perspective and framing of a 50mm lens on FX.  If I wish to carry a DX camera, to keep roughly the same perspective and framing then I think I need to use a 35mm lens.  I cannot use a 50mm and step back to preserve the framing, because perspective will change, and also there is typically no room to step back sufficiently.  I think that, assuming an evenly illuminated scene, the aperture  and shutter speed should be the same for both formats.  But if I want the same DoF with DX as with my FX equipment, I will have to increase the aperture by one stop, and therefore have to increase the shutter speed by one stop.  I also think that there will be more noise with the DX image, which will become more noticeable in low light conditions.  For this purpose I assume same generation up to date Nikon technology.

Is my analysis correct?

[Bjørn Rørslett]

Simsurace: That assumes a person consider the question worthy of attention in the first place .... Personally, I couldn't care less. One does what the camera does best, or if one only has a given gear at disposal, makes the most out of the situation. Period.

These "theoretical" considerations really boil down to idle speculation and in most cases, circular reasoning.

I have done my photography with camera formats from 1/3" to 8x10". The current frenzy of equivalence never entered the playing field, not once.

[simsurace]

Please continue.  I find your comments helpful.  As there is no definitive statement on this thread, I remain somewhat confused.

I think it is important to understand the effect of different sensors on the final image.

To give a real world example.  When I go out on the streets, I like the perspective and framing of a 50mm lens on FX.  If I wish to carry a DX camera, to keep roughly the same perspective and framing then I think I need to use a 35mm lens.  I cannot use a 50mm and step back to preserve the framing, because perspective will change, and also there is typically no room to step back sufficiently.  I think that, assuming an evenly illuminated scene, the aperture  and shutter speed should be the same for both formats.  But if I want the same DoF with DX as with my FX equipment, I will have to increase the aperture by one stop, and therefore have to increase the shutter speed by one stop.  I also think that there will be more noise with the DX image, which will become more noticeable in low light conditions.  For this purpose I assume same generation up to date Nikon technology.

Is my analysis correct?

Your analysis is correct in my opinion, and one could add the following two points:
- You changed your aperture to match DOF but you also changed your shutter speed, which is ok for a static scene but might have unwanted consequences for moving subjects. To get the same amount of motion blur you are not allowed to change your shutter speed. You therefore get double the exposure than you wanted, but unless you are at base ISO, you can use that extra stop to lower your ISO by one stop. You will then match the noise of the FX cam (assuming the same technology). Thus, you did your best to make your DX photograph look as similar as possible as the FX photograph, and this is what you wanted. Great!
- If the scene does not move or indeed you like the shorter shutter speed to freeze the movement, you can do as you suggested, and have the same exposure for DX and FX. As you say, the noise is expected to be higher for DX than for FX. Inquiring minds want to know why...*

* There are various (equivalent) ways to explain this, one of them being that the bigger sensor registered more photon events, hence has more information to process than the smaller sensor.

[simsurace]

Simsurace: That assumes a person consider the question worthy of attention in the first place .... Personally, I couldn't care less. One does what the camera does best, or if one only has a given gear at disposal, makes the most out of the situation. Period.

These "theoretical" considerations really boil down to idle speculation and in most cases, circular reasoning.

I have done my photography with camera formats from 1/3" to 8x10". The current frenzy of equivalence never entered the playing field, not once.

Still, the same physics must have affected the outcome of your photographs that affects all of our photographs, whether we are thinking about it or not. I do not see the conflict

[Bjørn Rørslett]

There is no conflict. I'm aware of what I'm doing.

[JohnMM]

The noise amplitude is lower for the bigger format and the information theoretic reason for that is that more photons were collected by the bigger sensor if both sensors were exposed the same (same energy per unit area).

Actually the noise itself is higher for the larger format. The noise varies as the square root of number of photons which means that the signal-to-noise ratio (SNR) varies in the same way. So the larger format has a higher SNR and it is this which manifests itself in a "less noisy" picture.

I mention this not in the cause of pedantry but because Eric Fossum made the same point to Great Bustard in a DPRev thread ( http://www.dpreview.com/forums/post/57949779 ) recently. Great Bustard is the user name of Joseph James, one of the promoters of "Equivalence". Eric Fossum is the designer of some of the CMOS architecture which is found in most of our sensors. So, if I'm viewed as a snake oil salesman, I find myself in good company.

I am of course in agreement with your ideas and and have argued for them previously on the earlier Nikon Gear site - now Fotozones. I have found the ideas expressed by James and others to be very helpful in correcting many misunderstandings which appear to have infilltrated the field of photography. I could give a detailed account but I fear that there is little interest here.

[simsurace]

Yes, thanks!
SNR is the proper term to talk about.

[simsurace]

There is no conflict. I'm aware of what I'm doing.

There is no question that you are. But I'm wondering then why you disagree so strongly with some basic assertions (like e.g. that the absolute number of photons collected is a relevant parameter in some arguments, vs. just exposure). Since I believe that you understand the argument, it must be that it strongly contradicts your observations. Even if you don't normally think in these terms, it would be of great interest to many of us if you would explain the contradiction such that we can discuss it -- hopefully learning something along the way.

E.g., you said that the difference in noise characteristics between different sensors is to be found elsewhere (than the total light captured). Please explain where and why. Or you said that Frank should find the contradiction in his argument. Please point it out, I think the discussion would greatly benefit.

[Bjørn Rørslett]

Frank has to answer for himself.  His example had logical flaws in it.

As I said, now several  times and phrased differently, yet obviously to no avail, I find these discussions so utterly worthless. We commenced with the false assertions of the "crop factor" thinkers and now obviously have entrenched ourselves in the borrows of the "equivalence" creationist. To what use?

Photography is not about "equivalence" and "sameness", it is about visual expression and being able to previsualise ideas. I have no clues as to why this is so difficult to understand unless photography as an art is dead and technical exercises replace it.

Added for clarity:

Just to make sure about the basic facts: formats are different. They always have been in their behaviour and technical solutions from the film days. Larger formats have less need of secondary magnification and hence show less grain (on a relative scale). Lenses for these formats resolve less because they must be designed to cover a larger area. The danger of empty magnification is small. They operate slower. A smaller format needs more secondary magnification, hence is prone to show more graininess. On the other hand, their lenses can be made faster and with higher resolving power thanks to their required image circles being smaller. The issue of empty magnification becomes more real. The systems operate faster.

For every different photographic application there is a system that optimises the specific requirements. This optimum would not be constantly placed on the "format scale". Acceptance over time for smaller format grew over time without being universally accepted.

Digital technology does not change these facts. What we have seen is, however, that the superior technology of digital  can push the relative relationships to cover smaller physical formats than before, but sooner or later the physical laws and the requirements for secondary magnification, dynamic range, etc.  introduce limits that lead to less than desirable performance. However, still the balance point is up to the individual photographer and their assignment, and no point on the "format scale" can be considered permanently fixed as the demands do change.

Now, this is the *end point* as far as I am concerned.

[Frank Fremerey]

The only question that remains for me ... I generally am 100% consistent with Simone Surace ...
is: from the same perspective the D500 + 1.4/24 shot still looks different from the D600 + 1.4/35 shot
although I did everything to have same
perspective
format adapted time and aperture
field of view

The difference I see is in the BG rendering.

Does the focal length play in here?

Or is it the lens characteristics?

Have  to leave. More later.

[Les Olson]

Actually the noise itself is higher for the larger format. The noise varies as the square root of number of photons which means that the signal-to-noise ratio (SNR) varies in the same way. So the larger format has a higher SNR and it is this which manifests itself in a "less noisy" picture.

I mention this not in the cause of pedantry but because Eric Fossum made the same point [...]

Another way of putting Eric Fossum's point is that you can treat the whole sensor as a single pixel, and the bigger the pixel the lower the noise.  You can do that because it turns out that you can bin pixels into larger and larger units with no increase in noise.  The problem with that analysis is that it does not tell you anything helpful about images with significant amounts of information content. 

In any image there are two sources of variation in voltage between pixels: noise and true pictorial information (noise as false pictorial information caused by lens aberrations could be considered separately as well).  The only practical way to identify noise is to eliminate pictorial information by photographing a uniformly illuminated screen, so you can assume that all pixel to pixel variation is noise.  If you are not interested in pictures of uniformly illuminated screens the SNR is not the square root of the number of photons.  The signal includes information about the picture, and you can't bin pixels without losing information about the picture.  A single pixel has an information signal to noise ratio of almost zero: the only information you get is the average scene luminance.

Smaller pixels allow you to collect more content information, because the sampling rate is higher.  Smaller sensors have more noise than bigger sensors, but the pictorial information the smaller sensor collects may be less, the same or more, depending on the pixel density (and on the quality of the lens), so the signal to noise ratio including pictorial information in "signal" can be anything. 

In most photographs the pictorial information is overwhelmingly predominant as a component of "signal" compared to the average scene luminance.  So sensor size has no important effect on the signal-including-pictorial-information to noise ratio, except when there is little pictorial information so that average luminance is a large proportion of the "signal" (eg, pictures of blue sky or areas of still water reflecting the sky). 

Of course, there is an additional component of "signal": aesthetic or, as the case may be, political or scientific information.  If it is present, as it all too often is not, that is so much larger a component of "signal" than any other that sensor size becomes completely irrelevant.

[Frank Fremerey]

Well, what you now describe has absolutely nothing to do with the starting point of this thread.

If we speak equivalence we have to have equivalent lenses and sensors. That is my point.

I do not answer to the starting point but to the flow of the discussion.

[David H. Hartman]

The difference I see is in the BG rendering.

Does the focal length play in here?

Or is it the lens characteristics?

May I note that the size of the aperture (not the aperture ratio) is larger for the 35/1.4 lens than for the 24/1.4 lens. It's my understanding that when the background is well outside the DoF zone the lens with the larger aperture (not the aperture ratio) will give more background blurring.

How Much Blur: 105/2.8 FX v. 70/2.8 DX v. 70/2.0 DX

How Much Blur: 35/1.4 FX v. 24/1.4 DX

Bokeh and Background Blur - Bob Atkins

Dave

---

Edit 2016-07-18: didn't want to bump the thread.

[I guess I should have included this link but I didn't because the calculator offered does not work with Windows7 and later. While the calculator doesn't work some of the text is worth reading. Here is the link...

Background Blur and Bokeh - a multifunction optical blur calculator

I'll quote one paragraph...

"One thing that this calculator demonstrates, and it's a point that people often fail to realize, is that depth of field and background blur are not the same thing and indeed may not even be closely related. People think that a lens setting with a small depth of field will blur the background more, but that isn't necessarily true. The math is quite different in the two cases. Depth of field will give you an estimate of local blurring about a subject in focus, i.e. how blurred the image will be just outside the traditional "depth of field" limits. However that doesn't correlate with how blurred the image will be of objects at a significant distance behind the subject in focus." --Bob Atkins.]