[vonBaloney]

So if you are claiming that you can FF-f/2.8 (i.e., the amount of background blur you can get from an f/2.8 lens on an FF-camera) from a f/2.8 lens on APS-C camera, you are wrong. You can only get APS-C-f/2.8 blur from an f/2.8 lens on an APS-C camera (which corresponds to ~FF-f/4.3). Note that I never said that any lens changes its physical properties or specifications. I only point out that one can translate specifications into equivalent specifications on a different format.

Surely you will accept that a 50mm lens produces a different AOV on a FF-camera compared to an APS-C camera. The same holds true for the maximum background blur that can be achieved (when keeping everything else equal). This change in effective properties can be expressed in terms of equivalent specifications.

So, it sounds like you are saying that if I have a FF camera and I shoot something with a 50mm at 2.8 and the background blur is whatever it is, then if I stay in the exact same spot and pop that exact same lens on my APS-C body and shoot the same thing (with a narrower AOV, naturally) at 2.8, then the background that is common to both images will be less blurry on my APS-C camera? If there is text in that background that I couldn't read on the FF due to blur, now it might be clear enough to read? (To make it a fair comparison, assume the FF has equivalently more megapixels overall so the APS-C areas on each will have the same resolution.)

[lytrytyr]

There is no difference between "light gathering" and "blurring ability" (bokeh). That 8.5/1.9 lens on a Q gathers as much light as a 46.75/10.45 on a full-frame camera. Both lenses have the same aperture diameter wide open. The latter determines "light gathering" ability.

See

Pentax-01 Standard Prime 8.5mm f/1.9 (Pentax Q) - Review / Lens Test

"The max. aperture of f/1.9 sound[s] fast and it is [in] terms of light gathering."

[alohadave]

Not true.

There is no difference between "light gathering" and "blurring ability" (bokeh). That 8.5/1.9 lens on a Q gathers as much light as a 46.75/10.45 on a full-frame camera. Both lenses have the same aperture diameter wide open. The latter determines "light gathering" ability.

The physical size of the aperture doesn't matter when comparing two different lenses.

The f/stop number is a ratio of the focal length/diameter of the aperture opening.

f/1.8 on an 8.5mm lens is the same ratio as f1.8 on a 50mm lens. The physical openings are different sizes because the focal lengths are different.

They both have the same light gathering ability at the same f/stops (at the aperture, the size of the lens elements has some impact).

[nater]

So, it sounds like you are saying that if I have a FF camera and I shoot something with a 50mm at 2.8 and the background blur is whatever it is, then if I stay in the exact same spot and pop that exact same lens on my APS-C body and shoot the same thing (with a narrower AOV, naturally) at 2.8, then the background that is common to both images will be less blurry on my APS-C camera? If there is text in that background that I couldn't read on the FF due to blur, now it might be clear enough to read? (To make it a fair comparison, assume the FF has equivalently more megapixels overall so the APS-C areas on each will have the same resolution.)

It might be useful to introduce the topic of Circle of Confusion at this point, and in particular this equation:

CoC (mm) = viewing distance (cm) / desired final-image resolution (lp/mm) for a 25 cm viewing distance / enlargement / 25

[m42man]

So, it sounds like you are saying that if I have a FF camera and I shoot something with a 50mm at 2.8 and the background blur is whatever it is, then if I stay in the exact same spot and pop that exact same lens on my APS-C body and shoot the same thing (with a narrower AOV, naturally) at 2.8, then the background that is common to both images will be less blurry on my APS-C camera? If there is text in that background that I couldn't read on the FF due to blur, now it might be clear enough to read? (To make it a fair comparison, assume the FF has equivalently more megapixels overall so the APS-C areas on each will have the same resolution.)

Sorry to butt in here, but there is a difference between lenses which are physically identical and lenses which have the same AOV and f-number. The latter will have different focal lengths for different sensor sizes, of course.

If you take the case of the lenses with same AOV, the one mounted of the FF camera will have more background blur (and therefore less DOF) than the one on the APS-C body (assuming the same f-number).

The case which you've described, whereby you use the same lens on the two bodies, results in a different AOV, so the APS-C version of the image will be a crop of the FF image. Measured in microns across the surface of each sensor, the blur will of course be the same for both cameras. HOWEVER, assuming you view the two images printed to the same size, the visible blur will be greater for the APS-C case (because it is, in effect, a slightly zoomed-in version of the scene, so everything is magnified, blur included). So, the DOF will be reduced, compared to the FF case.

To sum up: same AOV results in greater DOF for FF, same lens results in greater DOF for APS-C.

[Class A]

So, it sounds like you are saying that if I have a FF camera and I shoot something with a 50mm at 2.8 and the background blur is whatever it is, then if I stay in the exact same spot and pop that exact same lens on my APS-C body and shoot the same thing (with a narrower AOV, naturally) at 2.8, then the background that is common to both images will be less blurry on my APS-C camera?

For equivalent images -- it only makes sense to compare these -- you have to change focal length when changing sensor size.

If you find this article by Falk Lumo too technical, you may prefer this one by Bob Atkins.

"The max. aperture of f/1.9 sound[s] fast and it is [in] terms of light gathering."

That's not the only bit of nonsense that you can find on photozone.de pages.
Please see the article by Falk, or the "Equivalence" article by Joseph James.

The physical size of the aperture doesn't matter when comparing two different lenses.

It is the only thing that matters when it comes to light gathering and amount of background blur (EDIT: Assuming the same effective AOV, of course). Please see the articles I provided pointers to.

[zapp]

I have seen it said many times that 'if you have a 1.5 crop factor camera your lens will in effect have more zoom'... specifically a 100mm lens would be '150mm equivalent'...

Well here is my question...I assume the above type of statement was made using lenses built for full frame 35mm cameras. You in essence have the geometry of the lens ie the focal length, the diameter, and so forth where the lens allows a circle of light back to your sensor...

What a crop sensor effectively does is crop your photos at the time of taking the photo. A 100mm lens is still 100mm and the zoom is the zoom, but you are merely taking a recorded crop of that image.

So here is the crux of the question... if you have a lens built specifically FOR a 1.5 crop camera, ie a smaller barrel, a different angle of view than a full size counterpart, in short geometry and physics built around the smaller sensor...wouldn't a 100mm lens be considered '100mm' even in 1.5 crop sensorland?

Just to be clear with a full size normal lens for 35mm cameras they project a certain size light circle back to your sensor where it takes a crop of it...and thus leading to the 'appearance' that it has 'more zoom'... (whereas in reality if you had a full size sensor camera, took the same picture and then cropped it with software to the same dimensions it would be almost effectively the same thing)

With lenses 'build for the 1.5 crop sensor' do these dynamics change at all? Or does it even matter?

You start out telling us that you understand the concept of crop factor - it has no physical meaning. Later you ask us whether the crop factor has a physical meaning: no it does not. Focal length does not change only the angle of view.
The only thing that could change is the design of wide angle lenses as the smaller mirror requires less clearance to flip up and down. So a 35 mm APS-C lens can be build shorter than a fuill frame 35mm lens - image quality will be better and distortion will be much less. Check out what retrofocus design means - the major flaw of any SLR wide angle lens. Therefore a 31 mm lens designed for a full frame camera behaves different from a 31 mm lens designed for a smaller sensor.

[Blue]

You start out telling us that you understand the concept of crop factor - it has no physical meaning. Later you ask us whether the crop factor has a physical meaning: no it does not. Focal length does not change only the angle of view.
The only thing that could change is the design of wide angle lenses as the smaller mirror requires less clearance to flip up and down. So a 35 mm APS-C lens can be build shorter than a fuill frame 35mm lens - image quality will be better and distortion will be much less. Check out what retrofocus design means - the major flaw of any SLR wide angle lens. Therefore a 31 mm lens designed for a full frame camera behaves different from a 31 mm lens designed for a smaller sensor.

Regarding physical sizes of lenses, my DA 35/2.8 LTD isn't smaller than my F 28/2.8. The aperture size also has a major roll in the physical size of lenses. Plus, the *istD has a similar size mirror to my MZ-3. The DA 21/3.2 LTD is a pancake wide angle, but part of that size is due to it being f3.2. The DA 15/4 is relatively compact for a 15, but the DA 14/2.8 is a relatively large lens. Lastly, the angle of view is a big deal. That is why some want a full frame dSLR from Pentax. That way, the K 15 gets them a 15mm fov on a 135 size sensor. At the end of the day, the retrofocus lenses have the same registration distance as regular lenses. Mirror cages can be altered. I have an auto-tak 35/2.3 which was a trend setter in the 1950s and was considered a fast 35mm lens. It clears the mirror just fine on the *istD and K20d as well as my Asahi Pentax K. The DA 40/2.8 LTD will actually cover the 135 film on bodies than can control aperture. What I am getting at in a round about way is that while aps-c may make it possible to make some smaller lenses, there is a law of diminishing returns, pun intended.

Compare the size of the FA* 300/4.5 and DA* 300/4 and you will see that the DA* is larger and heavier although the latter is a bit faster. While the FA* 200/2.8 and DA* 200/2.8 are roughly the same size, the DA* is 40 grams heavier. The DA* 55/1.4 is not a small lens compared to the FA 50/1.4.

[vonBaloney]

For equivalent images -- it only makes sense to compare these -- you have to change focal length when changing sensor size.

Exactly, but no one ever says that when explaining these things. Because it does NOT "only make sense to compare these" (equivalent images). It makes much more sense to me to first explain the compare the differences between the formats ALL ELSE BEING EQUAL (same lens, same aperture, same distance to subject), which basically comes down to images from the "the smaller sensor are just crops of what you'd get from the bigger one". Blanket general statements like "smaller sensors have larger depth of field" (which I see all the time) only make sense in the context of "if you are trying to match the framing of a lens of longer focal length on a camera with a larger sensor with your lens of shorter focal length but that is relatively the same AOV on your smaller sensor" which is obviously a complicated thought and is vitally important, yet no one seems to want to provide this context, but merely stating "Small sensor. Big DOF. Ugh." which again, is very confusing because it makes it sound like I just *can't* get a narrow DOF with my smaller sensor, but of course I get exactly the same DOF as I'd get on a bigger sensor with the same lens, I just get less stuff in the image.

So then we get into ultimate magnification of the image, and how images from the smaller sensor will most likely be more magnified for viewing than those from the big sensor. This is a real issue in daily practice. And of course it has the opposite effect of shrinking the DOF on the smaller sensor because when we blow things up, stuff that looked sharp smaller won't look as sharp bigger (see the circle of confusion links), which means hammering the point that small sensor = larger DOF makes even less sense to stress all the time. It is also more confusing because it doesn't just depend on the sensor size, but the pixel density, which is quite variable for the same sized sensors and changing with every new model.

So while the equivalence stuff is good to know when considering a new camera of different format than you are used to, for most it is just total confusion and generally irrelevant, especially given the way people seem to insist on explaining it, which is to skip all the important foundational concepts you need to know to really get it and just make generalized conclusions that make no sense without understanding the context of the conclusion. It might also be nice to point out that "equivalent" images are impossible, and if you one lens on one format and an "equivalent" lens on another format and attempt to take the "same" picture, they will look utterly different even putting aside the DOF differences.

[Ben_Edict]

Not true.

There is no difference between "light gathering" and "blurring ability" (bokeh). That 8.5/1.9 lens on a Q gathers as much light as a 46.75/10.45 on a full-frame camera. Both lenses have the same aperture diameter wide open. The latter determines "light gathering" ability.

I don't think, this is quite that easy: the light gathering ability would be the same for equal diameter lenses (as in your example) for point sources. In photography we usually deal not with point sources (in astronomy, this would be meaningful, because stars are point-like), but with subjects which have an area, an apparent angle. In that case, the ratio between focal length and diameter, aka the chosen aperture (or the max. aperture) is the decisive factor.

Ben

[alohadave]

It is the only thing that matters when it comes to light gathering and amount of background blur. Please see the articles I provided pointers to.

Two lenses.

500mm at f/8

25mm at f/8

The 500mm lens at f/8 would have an aperture opening of 62.5mm.

The 25mm lens at f/8 would have an aperture opening of 3.125mm.

They both have the same light gathering ability, even though the aperture sizes are completely different. That is why you don't compare the physical aperture measurement. You reference the ratio, and the ratio tells you the light gathering.

[Class A]

It might also be nice to point out that "equivalent" images are impossible, and if you one lens on one format and an "equivalent" lens on another format and attempt to take the "same" picture, they will look utterly different even putting aside the DOF differences.

What makes you think the images would look different?

They both have the same light gathering ability, even though the aperture sizes are completely different.

But they don't have the same AOV.

When I previously wrote that they aperture size is the only thing that matters, I was assuming that we were talking about the same AOV (my underlying assumption is always that of equivalent images because it is nigh on impossible to compare images made from different viewpoints or with a different AOV, etc).

So while you are correct that keeping the f-ratio constant when changing focal length ensures the same exposure for one and the same sensor format, it is equally true that when you change the sensor format and hence have to adapt the focal length, it is then important to keep the same aperture size (not the same f-ratio). This is why when one calculates equivalent lens parameters, one not only multiplies/divides focal length by the crop factor, but also the f-ratio.

I don't think, this is quite that easy: the light gathering ability would be the same for equal diameter lenses (as in your example) for point sources. In photography we usually deal not with point sources (in astronomy, this would be meaningful, because stars are point-like), but with subjects which have an area, an apparent angle. In that case, the ratio between focal length and diameter, aka the chosen aperture (or the max. aperture) is the decisive factor.

When you assume equivalent images, i.e., same AOV, etc. the same (not in the sense of "identical") scene photons will attempt to pass through the aperture. The scene contents do not matter as both lenses "see" exactly the same photons (again not identical ones because two sensors cannot be in the same place at any one time, but the idea is that the same images are formed on the sensors).

As a result, it is necessary that the same scene photons (in particular the same amount) pass through the aperture and hence the aperture size must be the same for both lenses.

At the sensor, the photons will be spread out over a larger area in case of the larger sensor camera. This leads to a reduction in local light intensity (which is why one has to increase the ISO setting for the larger sensor camera if one intends to achieve the same exposure). However, the total amount of light gathered by each sensor (i.e., local light intensity times surface area) is exactly the same and this reduction in light intensity on the larger sensor surface is necessary for "fairness". Otherwise the larger sensor would receive a higher amount of total light, hence lower shot noise.

I kindly ask anyone questioning me to read the articles I linked to. I may or may not be around later to further discuss this topic.

[hjb981]

Dear all,

I will butt in and try to clarify some of the concepts. I hope it will be helpful, and I have highlighted what I believe to be the most important concepts.

It is true that f/2.8 always means the same light gathering ability as measured per area unit of the sensor. That means that 1 square centimeter of the sensor will always see the same amount of light from a 2.8 lens (if it is pointed towards the same even light source, say a clear sky or evenly illuminated wall for example). So while the amount of light per area unit of the sensor (or film) stays the same, the total amount of light that the sensor gathers will be proportional to the sensor size.

For those less into theory, and more into trying: just take two cameras, one with the biggest sensor you can find, and one with the smallest sensor you can find. Take a shallow depth-of-field picture with the big-sensor camera, then try to make the same picture with the small-sensor camera. Same picture means same distance from camera to subject, and same field of view. It will not be possible to for example reproduce the shallow depth of field of a FF f/2 70 mm on a compact camera that has a sensor 1/20th the size of the FF. This is quite noticeable if you regularly use both kinds of camera.

The below discussion is rather theoretical. I have tried to make it clearer by highlighting the most important parts, and I would urge anyone reading it to do his or her own thinking. In other words: consider the arguments, make drawings of different sensor sizes (that is what I did) or whatever you do to grasp a new concept.

The ISO number also measures the amount of light per surface area unit (just like the F number). That means that a large sensor gathers much more light in total at the same ISO setting as a small sensor, because the amount of light per surface area unit is the same, and it has a bigger surface area. If you look at for example the DXO ratings and compensate for this, given the same amount of light in total, big sensor do not perform better than small sensors. However, with a big sensor, you have the option of opening up the aperture and sacrifice depth of field and gather the same amount of light in total with a shorter shutter time. If I take a picture with my Panasonic LX3, at its maximum aperture F/2 at ISO800 and at its shortest focal length, 5.1 mm in a dark place, I will get a crappy-looking (noisy) picture, with the equivalent field of view as 16 mm on an APS-C camera (24 mm on FF). If I wanted to reproduce that with my K-5, and get the same depth of field and shutter time, I would have to use a 16 mm lens at F/6.3. Since I want to keep the shutter time the same, and F/6.3 allows less light per surface area unit of the sensor, the sensor's sensitivity to light per surface area unit will have to be increased, which means that I will use a higher ISO setting -- in this case about 8000 (which means that the sensor in the K-5 has an area about 10 times as big as the one in the LX3). In doing this, the total amount of light hitting the sensor in the K-5 would be the same as in the LX3, only spread out over a larger sensor. Of course, I could use the 16 mm lens on the K-5 at F/2 and also keep the K-5 at ISO800 (shutter time still the same). That would render a much cleaner image with less noise than what would be possible to achieve with the LX3, but that image would also have a shallower depth of field than the one from the LX3, and hence they are not equivalent.

Yes, a 5.1 F/2 is a 5.1 F/2 on all sensors, but I think it should be quite clear to everyone that putting a 5.1 mm lens on the K-5 would render an image that is very different in terms of field of view compared to using the 5.1 on the small-sensor LX3. It makes much more sense to compare the equivalent focal lengths (5.1 mm for the small sensor and 16 mm for the large sensor).

Dear all, please consider what Class A wrote and give it some thought. Read the articles (s)he is linking to! (S)he is right.

[Ben_Edict]

At the sensor, the photons will be spread out over a larger area in case of the larger sensor camera. This leads to a reduction in local light intensity (which is why one has to increase the ISO setting for the larger sensor camera if one intends to achieve the same exposure). However, the total amount of light gathered by each sensor (i.e., local light intensity times surface area) is exactly the same and this reduction in light intensity on the larger sensor surface is necessary for "fairness". Otherwise the larger sensor would receive a higher amount of total light, hence lower shot noise.

.

The AOV will only be the same, if the lens for the larger film/Sensor format has a longer focal length, thus the f-ratio will go down, if you keep the absolute aperture diameter constant. That is the reason for a dimming of the image on the sensor. If you forget the absolute diameter and keep instead the f-ratio constant, the illumination on the sensor plane will be the same, whatever be the sensor size. <-- Simply, because the longer fl lens will have bigger absolute diameters of the aperture iris at any given setting.

Have you ever seen a lightmeter that gives different reading, dependent on film format?

It is, I think, quite easy to comprehend and does not need pages of articles for explanation - which by the way are not bad.

Ben

[lytrytyr]

If you forget the absolute diameter and keep instead the f-ratio constant, the illumination on the sensor plane will be the same, whatever be the sensor size.

This is what is meant by "light-gathering" when that term was used earlier, e.g. in
Pentax-01 Standard Prime 8.5mm f/1.9 (Pentax Q) - Review / Lens Test