[aleonx3]

Interesting discussion here... enjoy all the views of different posters (mostly not fully understand but slowly getting to it). I always have a burning question to ask since I enjoy using a FF lens (Topcor 58f1.4 converted to m42 mount) on my k-3. My understanding is very simple, because of the APS-C sensor, with this lens the FOV will be 87 on my viewfinder as if it is cropped from FF to APS-C size . And the resulting image taken with the k-3 would be same (?) as if it is taken with a FF except that it is already cropped in the camera standing from the same position (if I have a FF to compare).

Now, assuming that the pixel density and photosite characteristics on k-3 and the future FF are the same (this FF camera would have to be somewhere around 48MP), would I expect the DOF and even exposure (shutter, ISO) be the same in both cases if I keep the aperture the same at say f1.4? Thanks...

[Class A]

Now, assuming that the pixel density and photosite characteristics on k-3 and the future FF are the same (this FF camera would have to be somewhere around 48MP), would I expect the DOF and even exposure (shutter, ISO) be the same in both cases if I keep the aperture the same at say f1.4?

An FF camera with the same pixel-pitch as the K-3 would have to have almost 55 MP.

If you use the FF camera in APS-C crop mode on such an FF sensor you'd get exactly the same results as on the K-3.

EDIT: If you use all of the FF sensor for the Topcor lens then leaving the f-stop at f/1.4 would yield shallower DOF than on the K-3, i.e., shallower than using a 39/1.4 lens on the K-3. I'm using a "39mm" lens on the K-3 now, because this is the focal length you'd have to use to replicate the look of the 58mm on an FF camera. To achieve the same DOF wide open as the 58/1.4 on FF, the lens on the K-3 would have to be 39/0.93 lens.

[Class A]

The way this thread is still going, as I see it, that you guys are discussing cards stacked on top of a pile, but nobody knows about the underlying pile.

Nobody? Really?

You see, insinuations like these make your suggestions less inviting than they could be.

If you are interested in a more formal treatment, did you check out the “Camera equivalence" article by falconeye?

What makes you think this is not solid enough?

And, no, I do not believe that equivalence deniers could be convinced by further such treatments. I can already envision the responses ("...mambo jumbo...formula schwormula...go out and shoot more and see for yourself...").

[aleonx3]

An FF camera with the same pixel-pitch as the K-3 would have to have almost 55 MP.

If you use the FF camera in APS-C crop mode on such an FF sensor you'd get exactly the same results as on the K-3.

If you use all of the FF sensor for the Topcor lens then leaving the f-stop at f/1.4 would yield shallower DOF than on the K-3. It would look like using an f/0.93 lens on the K-3. Shutter speed would be different accordingly if you choose the FF equivalent ISO setting on the FF camera.

Thanks, this "the above underlined" is where I fail to understand, I thought 58mm focal length is the same no matter what (that 58mm comes with a native perspective - which is really what I am aiming at, not 'compressed' as in 87mm in crop mode (true or not true?). I can understand the exposure (shutter) would be different since it will be metered on a larger view. Note: I am not referring to maintaining the same FOV comparing FF to APS-C by getting closer subject distance; I am talking about 'cropping' the FF image to get the same FOV as the APS-C image.

---

Last edited by aleonx3; 07-20-2015 at 09:51 AM.

[osv]

What you might find interesting with in those discussions is that the final individual that filed a complaint with DP about dtmateojr was indeed a person well versed and many years in senor design and also hold more patents in sensor technology and written more published papers then most people have taken photos.

That person gave dtmateojr all the time in the world and tried to educate him but at the end it was dtmateojr colorful charm that gave rise to the filed complaint

i wonder if there were any posters like clackers in those dpr threads, in other words, people who actually believed what dtmateojr had to say, and agreed with him.

---------- Post added 07-20-15 at 10:55 AM ----------

All of the light that was falling on the 100mmx100mm sensor is not now concentrated on the 50x50 sensor... Only half of it is.

clackers, you quoted his post: "the exposure (light / area) is different between the 50mmx50mm sensor and the 100mmx100mm sensor"... so what part of that did you not understand?

clackers, stop repeating what people say, and then pretending that your parroted words make what they said wrong... it just trashes the discussion with nonsense.

[bossa]

I'm pretty sure the original intent of the concept of equivalence was to allow film based FF photographers to get a handle on crop mode digital cameras so they could maintain their traditional relationship with FOV and DOF. This would have been used as a simple guide to allow transition across formats but has mutated into a ridiculous online p*ssing contest.

Some of the posts here are quite personal and denigrating in nature. I'd like to say that if your true intent is to win an argument then you don't win any arguments by laughing at people and putting them into a competitive orientation.

[everydaylife]

Interesting discussion here... enjoy all the views of different posters (mostly not fully understand but slowly getting to it). I always have a burning question to ask since I enjoy using a FF lens (Topcor 58f1.4 converted to m42 mount) on my k-3. My understanding is very simple, because of the APS-C sensor, with this lens the FOV will be 87 on my viewfinder as if it is cropped from FF to APS-C size . And the resulting image taken with the k-3 would be same (?) as if it is taken with a FF except that it is already cropped in the camera standing from the same position (if I have a FF to compare).

Now, assuming that the pixel density and photosite characteristics on k-3 and the future FF are the same (this FF camera would have to be somewhere around 48MP), would I expect the DOF and even exposure (shutter, ISO) be the same in both cases if I keep the aperture the same at say f1.4? Thanks...

And the resulting image taken with the k-3 would be same (?) as if it is taken with a FF except that it is already cropped in the camera standing from the same position (if I have a FF to compare).

If you use all of the FF sensor for the Topcor lens then leaving the f-stop at f/1.4 would yield shallower DOF than on the K-3. It would look like using an f/0.93 lens on the K-3. Shutter speed would be different accordingly if you choose the FF equivalent ISO setting on the FF camera.

This is where I get confused. The difference in dof. I can understand that if I put my fa50 1.4 on a 1.5 apsc camera the field of view will be the same as a 75mm lens on an FF.
I cannot understand that if I put the 50mm set at f1.4 on an ff camera with the same camera to subject distance why the dof is not the same as in the image taken with the 1.5 apsc camera.

---

Last edited by everydaylife; 07-20-2015 at 07:21 PM.

[Class A]

Gee, you don't understand shot noise either, but you talk about it anyway!

May I ask what the purpose of the above personal comment is?

Why don't you just present your (incorrect) alternative understanding of "shot noise" and then continue to have a civil discussion?

The shot noise is a property not of the sensor size at all, but the electron holding capacity of the individual pixels on it (how 'deep' they are, for an unsatisfactory analogy, but you get the idea).

Incorrect.

Perhaps we can start making progress if you a) look up what "shot noise" (aka "photon noise" in this context) means, and b) admit that the full-well capacity has nothing to do with photon noise.

While the term "shot noise" is also used in the context of discussing electronic devices, I clearly used it to refer to "photon noise". See the second paragraph of the "Origin" section on "shot noise" at Wikipedia for an example explaining shot noise that does not involve any sensor or electronics at all. In short, light itself is noisy.

Your claim that shot noise depends on the full-well capacity (FWC) of pixels is incorrect because the SNR of a series of measurements is independent of the dynamic range of the measurement device as long as signal levels do not exceed the latter. Consider the following analogy:

Say you want to measure rain intensity with a bucket by exposing it to rain for 10min. Provided that the bucket does not overflow in the 10min, do you think the height of the bucket makes any difference to the measurement? If you have a 2l bucket and a 10l bucket with the same diameter opening -- i.e., they only differ in height -- clearly the 10l bucket has a higher capacity but will you measure a different amounts of rain / 10min with it?

Of course you won't, so why should the FWC of a pixel influence the SNR of a light signal?

Can you please admit that you were wrong, so that we can make progress with further misconceptions?

All this talk of Poisson ... Einstein was suspicious of people covering up lack of knowledge by parroting formulas and definitions ... "If you can't explain it simply, you don't understand it well enough."

What do you mean by "All this talk of Poisson"? I only provided the name of the type of noise distribution we are dealing with. I never tried to convince by referencing authority.

May I point out to you that it is you who is associating the weight of an argument with from whom it is coming from? You do not tire to repeat that Richard Butler and Joseph James are not scientists, as if that had any relevance. As a former Physics teacher, you should know that Physics was revolutionised by a patent officer. The respective genius did not hold a scientific position, he was merely evaluating patents. What if the scientific community had rejected his ideas on the grounds that he is "no scientist"? Well, scientists do not care who submits an argument, they evaluate the argument.

[Class A]

All of the light that was falling on the 100mmx100mm sensor is not now concentrated on the 50x50 sensor.

Incorrect.

When I asked you "Do you agree that the exposure (light / area) is different between the 50mmx50mm sensor and the 100mmx100mm sensor?, I explained that both sensors receive the same amount of light.

Do you accept that "focal length" and "registration distance" are not the same?

If you do then can you imagine the same scene being captured with two cameras that use different registration distances?

Can you now imagine that the lens on the camera with the larger registration distance will produce an image circle that is larger and further away from the mount?

Can you finally imagine that both image circles contain the same amount of light but that of course the intensity of light in the larger image circle will be lower?

If you followed me so far, how can you not agree that the exposure (light / mm^2) of the larger image circle is lower?

Only half of it is.

You are obviously thinking about a "cropping" situation, i.e., only changing the sensor size, but not the size of the image circle at the same time.
My example was different, but let's assume your cropping scenario.

Do you realise that the surface area of a 100mm x 100m sensor is four times larger than that of a 50mm x 50mm sensor?
Why would still "half" of the light be captured by the 50mm x 50mm sensor?

If we had a cropping situation -- which we had not in the thought experiment I described -- then only one quarter of the light (not one half) would be captured by the smaller sensor.

Richard Butler ... clearly cannot control variables, since he uses four different cameras and lenses for his DPR 'experiment' you obsess over)...

Can you please explain to me what exact variables he fails to control?

He has to use different lenses, not only because different mounts are involved but because he needs different focal lengths for the different sensor sizes. While that may influence bokeh, CA, etc. these are image aspects we are not interested in here, as the focus is on image quality as determined by noise levels.

Obviously he has to use different cameras to compare different sensor sizes.

Can you please furthermore explain why the variables he did not control (e.g., camera brand) would invalidate the experiment?

Note that Richard Butler explicitly mentions "sensor performance" and "differences in transmission between lenses" as important factors that influence results, so he is aware of these threads to validity. So what did he do wrong, exactly?

[Class A]

Thanks, this "the above underlined" is where I fail to understand, ...

I revised my post to be more explicit about the respective change in focal length that would have to occur. I hope it is easier to understand now.

I thought 58mm focal length is the same no matter what (that 58mm comes with a native perspective - which is really what I am aiming at, not 'compressed' as in 87mm in crop mode (true or not true?).

Please note that "perspective" is only determined by the camera location, i.e., not focal length. The latter determines how much you see of the scene (i.e., the FOV) but not the spatial relationships between objects in the scene.

However, how much of the scene is included has an impact on the viewing distance for the final output. If, for instance, only a small portion of the scene is visible, but viewed from a small distance, spatial relationships between objects will appear to be much more flattened than they appeared to the photographer from the camera position. If the same image is viewed from a larger distance (image diagonal * focal-length / normal-focal-length), the spatial relationships between the objects will appear to be the same as they presented themselves to the photographer at the camera position in the real scene.

So focal lengths do not define a perspective, but they imply viewing distances if you want to recreate the original spatial relationships. This is, of course, why focal length can be used for effect, as typically the viewing distance is roughly the same and hence angles in the image will look exaggerated (wide-angle) or diminished (longer than normal focal length).

For the purposes of predicting angle exaggeration or flattening, you need to treat the 58mm lens on the K-3 the same as a 87mm on an FF, as it is the effective FOV that is relevant here, not the optical power of the lens.

[Class A]

This is where I get confused.

I revised my post. Hopefully, it is clearer now.

I cannot understand that if I put the 50mm set at f1.4 on an ff camera with the same camera to subject distance why the dof is not the same as in the image taken with the 1.5 apsc camera.

What "image taken with the 1.5 apsc camera" do you mean?

Do you mean an image taken with the same 50mm at f/1.4 on the APS-C camera?
Obviously the resulting image would look very different (include less of the scene) and hence a DOF comparison makes little sense.

You can of course calculate and compare DOF even for images that do not have the same FOV but it becomes an apples to oranges comparison. Note that in the linked to article, the crop factor is assumed to be 1.6 rather than 1.5 because he assumes the (slightly smaller) Canon APS-C format.

[kh1234567890]

Say you want to measure rain intensity with a bucket by exposing it to rain for 10min. Provided that the bucket does not overflow in the 10min, do you think the height of the bucket makes any difference to the measurement? If you have a 2l bucket and a 10l bucket with the same diameter opening -- i.e., they only differ in height -- clearly the 10l bucket has a higher capacity but will you measure a different amounts of rain / 10min with it?

Except that if you increase the sensel size then you increase its capacitance and hence the same charge (number of incident photons) will give you a smaller voltage, so you need more gain in the ADC etc. stages resulting in more readout noise.

[BrianR]

I cannot understand that if I put the 50mm set at f1.4 on an ff camera with the same camera to subject distance why the dof is not the same as in the image taken with the 1.5 apsc camera.

The effect on DoF when you keep the same lens, f/stop, subject distance, and print size, but move to a smaller sensor size (and of course different field of view) is exactly the same as you'd get if you crop in post. You can try it at home, but here's an example:

https://www.pentaxforums.com/forums/169-pentax-full-frame/240468-dof-ffs-aps-...ml#post2553553

[Rondec]

Equivalence is mostly useful if you happen to shoot more than one format and can't figure out which lens to use on one camera or another. If you happen to shoot only one format -- APS-C, 645 digital, or four thirds, then it is meaningless to you, particularly if you are not someone who ever shot film much. Relating all focal lengths and apertures to a magical format that few young photographers have actually used, is a little silly. Personally, I do equivalence backwards, relating full frame to APS-C since I have shot that by far the most over the last 10 years.

There is a somewhat magical statement that is required as well, which is that the sensors involved need to have the same technology and performance. The thing is that the statement is made that when you are comparing, 135mm f2.8 iso 100 on APS-C is roughly equivalent to 200mm f4 iso 200 on full frame. This is true because SNR and dynamic range on full frame are generally one stop better so you don't pay a penalty for bumping your iso up one stop. Unfortunately, that is not necessarily true. The best APS-C cameras now (D7200) do a lot better with regard to dynamic range than older APS-C cameras and match or exceed many of the full frame cameras on the market. They still trail significantly in the SNR department, but you can't really say that you can take an image on a 6D, stop it down a stop and have the equivalent image with regard to all image characteristics as the D7200. If DXO Mark is right, the D7200 image will be a better image than the 6D image with regard to dynamic range.

Unfortunately, equivalence often comes across as folks who shoot full frame beating others over the head with the fact that the format they shoot is better in all respects. The reality is that it is just different. It is capable of great images and great tonality, but the fact that it has faster lenses available for it does not mean that it actually is "better" than digital medium format, for example, although equivalence would say that it is.

[Class A]

Except that if you increase the sensel size then you increase its capacitance and hence the same charge (number of incident photons) will give you a smaller voltage, so you need more gain in the ADC etc. stages resulting in more readout noise.

Higher-capacity sensels indeed can have a higher read-out noise (*), but this is a small effect, compared to the large shot noise effects we are discussing. Also, this small read-out noise is not "shot noise" as clackers claimed.

In other words, extreme situations excluded (extreme low-light and high-capacity sensels), shot noise (photon noise) will dominate the image noise.

For all purposes and intents, we can ignore pixel size for image quality discussions.

(*) N.B., that's why the often purported "larger pixels are better" is a myth.