Depth of field FF V Crop

[scottduffy]

I was reading this guys page and he is going on about 85mm lenses being the one lens that can change your portrait photography most and is a must have in any portrait photographers bag. Then he states that "While I think every photographer should have a 50mm lens in their bag somewhere the 85mm on a crop sensor camera will be even more awesome in giving that shallow DoF, with the longer focal length giving a more pleasant perspective for close-up portraits." Is the talking about the 50mm equaling 85mm on a crop or is he talking about sticking an 85mm on a crop body? If that is the case then why would he not be using 135mm prime for his portraits rather than an 85mm if he thinks the longer length gives a more pleasing perspective? Does that make sense?

Here is what he actually says:

85mm on a Full-frame camera vs. 50mm on a Crop sensor
Photographers with crop-sensor cameras will of course wonder if a 50mm f/1.4 will give similar results. Since the crop will force a different perspective for the same composition, effectively giving that longer focal length’s field of view, you’d get a similar effect with shallow depth-of-field.

• Full-frame vs Crop-sensor comparison : Depth-of-field & Perspective

While I think every photographer should have a 50mm lens in their bag somewhere, the 85mm on a crop sensor camera will be even more awesome in giving that shallow DoF, with the longer focal length giving a more pleasant perspective for close-up portraits.

 

[petersmart]

As far as I read it he means the actual 85mm which on a 1.6 crop would be equal to 136mm or on a 1.3 crop would be equal to 110mm.

I think a 135mm on a crop would be too long giving equivalents of 216mm or 175mm on a 1.6 or 1.3 respectively.

If he did actually mean the 50mm then the equivalent focal length would be 80mm or 65mm.
.

 

[petersmart]

OOPS double post!

 

[Kodiak Qc]

•


There are so many understanding mistakes in what you
report and I don't even think it is worth commenting.

Instead, I will try to explain and hope it will help.

Any lens has its given set of characteristics and features,
from the fisheye to the super tele. Any given lens will keep
these what ever sensor it is coupled to. I mean all its cha-
racteristics and features. The only thing that may change
is the angle of view (AoV).

If I exaggerate things, hoping to explain them better, say
a 16 mm fisheye on a FF sensor. The whole FF image will
display a strong barrel effect more accentuated toward the
edges of the frame and non existent in the very centre.

So far, so good?

Say you crop that full frame down to the central 15 x 10 mm
area. Looking at the resulting image that will displays far less
barrel effect does not mean it is the equivalent of a 35 or 50
mm in any other term then AoV.

I hope I could translate my thoughts in a clear way.

 

[chris malcolm]

All a crop sensor camera does is to use a smaller sensor, giving exactly the same image as you'd get by cropping down to the same central portion in a full frame camera. Obviously cropping doesn't change DoF. But to fill the frame with a portrait you'd have to stand further back with a crop sensor camera. Which will give you exactly the same DoF as you'd get if you moved back to the same distance with the full frame camera. It's simple if you think of the geometry. It gets complicated if you try to avoid the geometry and use words instead.

 

[ABTog]

Using an f1.8 aperture on a crop factor body gives a different depth of field to using the same aperture on a full frame body. It doesn't change the amount of light getting in. So you have to multiply your f-stop by the crop factor to get the equivalent full frame depth of field.
So f1.8 on a Canon crop sensor is 1.8 x 1.6 = 2.88. So using your 50mm at f1.8 on a crop sensor will give you the equivalent depth of field of a lens at f2.8 (and a field of view equivalent of an 80mm lens).

If you think this is wrong, watch this demonstration from Tony Northrup:

View: https://www.youtube.com/watch?v=f5zN6NVx-hY


Remember crop factor doesn't the focal length or the actual aperture, just the field of view and depth of field equivalency.

 

[petersmart]

Well the crop factor changing the angle of view is fairly obvious but when he says the crop factor increases noise due to less light falling on the crop factor sensor that doesn't really make sense - the noise level increase is due to the smaller photo diodes producing less electrical signal which requires greater amplification which produces more noise.
.

 

[scottduffy]

Sorry guys, thanks for your responses however i just meant if this guy is stating that you get a more pleasant perspective from an 85mm on a crop sensor then why would he be advocating the use of 85mm on full frame instead of 135 on full frame which would give him the same pleasant perspective he seems to prefer?

 

[scottduffy]

Using an f1.8 aperture on a crop factor body gives a different depth of field to using the same aperture on a full frame body. It doesn't change the amount of light getting in. So you have to multiply your f-stop by the crop factor to get the equivalent full frame depth of field.
So f1.8 on a Canon crop sensor is 1.8 x 1.6 = 2.88. So using your 50mm at f1.8 on a crop sensor will give you the equivalent depth of field of a lens at f2.8 (and a field of view equivalent of an 80mm lens).

If you think this is wrong, watch this demonstration from Tony Northrup:

View: https://www.youtube.com/watch?v=f5zN6NVx-hY


Remember crop factor doesn't the focal length or the actual aperture, just the field of view and depth of field equivalency.

Thanks mate i'll have a wee look at this when i get out the office.

 

[ABTog]

Oh and here's a nice animation of what lens length does
http://i.imgur.com/XBIOEvZ.mp4

 

[woof woof]

Sorry guys, thanks for your responses however i just meant if this guy is stating that you get a more pleasant perspective from an 85mm on a crop sensor then why would he be advocating the use of 85mm on full frame instead of 135 on full frame which would give him the same pleasant perspective he seems to prefer?

Quite. And very possibly the FF image may be sharper, if that's a desirable quality.

In general I think it's possibly wise to assume that stuff you read on line ain't necessarily so. I mean, anyone can post anything on line

 

[sk66]

Sorry guys, thanks for your responses however i just meant if this guy is stating that you get a more pleasant perspective from an 85mm on a crop sensor then why would he be advocating the use of 85mm on full frame instead of 135 on full frame which would give him the same pleasant perspective he seems to prefer?

Either you misunderstood what he wrote, or he's an idiot... (more the latter).
You are correct in that the perspective is due to the distance required caused by the FOV (same composition). I read it as comparing 85mm on FF vs 50mm on APS for relatively the same FOV/Distance (perspective). But the DOF will be less on the FF (same aperture). I didn't read it as advocating 85mm on FF.
And when comparing the 50mm and the 85mm on the same body/sensor... the *DOF will remain essentially the same when working with shallow DOF (again, same composition/different distances).
The comparison is also somewhat flawed in that the difference in FL is not the same as the difference in FOV/crop factor... there's no way to "equalize" the comparison...

*The "appearance" of the BG will be different if there is any notable distance/separation (less included/larger w/ the longer FL).

 

[scottduffy]

Either you misunderstood what he wrote, or he's an idiot... (more the latter).
You are correct in that the perspective is due to the distance required caused by the FOV (same composition). I read it as comparing 85mm on FF vs 50mm on APS for relatively the same FOV/Distance (perspective). But the DOF will be less on the FF (same aperture). I didn't read it as advocating 85mm on FF.
And when comparing the 50mm and the 85mm on the same body/sensor... the *DOF will remain essentially the same when working with shallow DOF (again, same composition/different distances).
The comparison is also somewhat flawed in that the difference in FL is not the same as the difference in FOV/crop factor... there's no way to "equalize" the comparison...

*The "appearance" of the BG will be different if there is any notable distance/separation (less included/larger w/ the longer FL).

Hi Steven,

Thanks for this mate. It's not actually in the part I posted but the guy on his page states that the 85mm lens is the one that will change your portraits for the better and he uses a Nikon d800 of Canon 5dmk3 so that's where I was getting the full frame bit from. I was reading some more of his page and lots of it seems really useful. First time I've came across him. I think you're right though about him meaning 50mm on a crop v 85mm full frame.

Regards

Scott

 

[HoppyUK]

Using an f1.8 aperture on a crop factor body gives a different depth of field to using the same aperture on a full frame body. It doesn't change the amount of light getting in. So you have to multiply your f-stop by the crop factor to get the equivalent full frame depth of field.
So f1.8 on a Canon crop sensor is 1.8 x 1.6 = 2.88. So using your 50mm at f1.8 on a crop sensor will give you the equivalent depth of field of a lens at f2.8 (and a field of view equivalent of an 80mm lens).

If you think this is wrong, watch this demonstration from Tony Northrup:

View: https://www.youtube.com/watch?v=f5zN6NVx-hY


Remember crop factor doesn't the focal length or the actual aperture, just the field of view and depth of field equivalency.

There are errors in many of the posts here, but this link from ABTog to Tony Northrup's video is a good and clear explanation. Recommended viewing

There are numerous key changes when you fit the same lens to a different format camera. Field of view is the big one of course, meaning that you have a completely different image that makes clear comparisons impossible, so the first thing is to adjust focal length by the crop factor, eg 80mm on full-frame vs 50mm on 1.6x crop format. With that done, the following applies:
- At same distance, framing remains the same
- At same distance, perspective remains the same
- Depth of field is increased on crop format, also by the crop factor, ie f/2.8 on full-frame delivers the same DoF as f/1.8 on 1.6x crop (2.8/1.6=1.8)
- Noise will increase in the cropper image, as less light/photons are collected by the smaller sensor. This can be restored by reducing ISO, by one stop is a very close approximation, ie ISO400 on FF should be the same as ISO200 on crop format (assuming same generation sensors).
- Dynamic range will be reduced in the crop format image, in the shadow areas, again because less light/photons are collected, and equally this can be restored by reducing ISO.
- Sharpness will be reduced on the crop format image. This is the bit that often gets forgotten/misunderstood and there is no simple equivalence conversion, but smaller sensors require greater magnification for same size output/print and therefore demand greater resolution from the lens. Fact of physics is that when resolution goes up, image contrast (sharpness) goes down. Therefore smaller format cameras need better quality lenses, but in practise no lens has yet been produced that can properly bridge that gap - bigger sensors always produce sharper pictures.

On depth of field, it is all about magnification and this applies to every aspect of the imaging chain - sensor size, focal length, distance, diameter of the aperture (f/number), print size, viewing distance. Change anything there, and DoF changes. Also note that cropping the image in post processing is also an effective change to sensor format.

Good DoF calculator here http://www.dofmaster.com/dofjs.html
Edit: note that background blur/bokeh etc is about more than depth of field and lens aperture. Subject distance, background distance, focal length and perspective are all at least as important.

 

[odd jim]

Using an f1.8 aperture on a crop factor body gives a different depth of field to using the same aperture on a full frame body. It doesn't change the amount of light getting in. So you have to multiply your f-stop by the crop factor to get the equivalent full frame depth of field.
So f1.8 on a Canon crop sensor is 1.8 x 1.6 = 2.88. So using your 50mm at f1.8 on a crop sensor will give you the equivalent depth of field of a lens at f2.8 (and a field of view equivalent of an 80mm lens).

If you think this is wrong, watch this demonstration from Tony Northrup:

View: https://www.youtube.com/watch?v=f5zN6NVx-hY


Remember crop factor doesn't the focal length or the actual aperture, just the field of view and depth of field equivalency.

This is the video that finally got my head around it! And Chelsea is quite nice too

 

[sk66]

I don't particularly like that video... I particularly pretty much disagree with everything related to "the amount of light."

The exposure remains constant... it's not about "light collected," it's about pixel size (light per pixel, DR/Noise) and more directly the greater enlargement required for the smaller physical size (which all go along with the sharpness/MTF issue).

And yes, I know that for any given sensor size smaller pixels do not *necessarily* equate to more noise reduced DR/Color... in fact the opposite may be true. Truth is, there are so many variables that you can't really get "equivalence" other than DOF.

But as a "generic" way of understanding how things "kind of" interrelate I guess it works ok.

 

[sk66]

I was reading some more of his page and lots of it seems really useful. First time I've came across him.

I just followed the link.
Yes, his stuff is generally very good.

 

[HoppyUK]

I don't particularly like that video... I particularly pretty much disagree with everything related to "the amount of light."

The exposure remains constant... it's not about "light collected," it's about pixel size (light per pixel, DR/Noise) and more directly the greater enlargement required for the smaller physical size (which all go along with the sharpness/MTF issue).

And yes, I know that for any given sensor size smaller pixels do not *necessarily* equate to more noise reduced DR/Color... in fact the opposite may be true. Truth is, there are so many variables that you can't really get "equivalence" other than DOF.

But as a "generic" way of understanding how things "kind of" interrelate I guess it works ok.

Not really Steven, you're just adding another layer of complication and not comparing like with like. You're changing the pixel size, which may or may not be the case in reality, and photon collection also depends on the light gathering efficiency of each pixel, the actual size and depth of the well, gap-less microlenses etc etc. But the overriding fact remains - there is more light/photons falling on a larger sensor's surface area, and at same ISO (and sensor generation) smaller sensors show more noise and have less dynamic range - and it's not hard to see.

 

[sk66]

But the overriding fact remains - there is more light/photons falling on a larger sensor's surface area

No, in terms of "equivalence" that is entirely incorrect. It works exactly the same as the ISL, the physics of light doesn't change and it doesn't care about the size of the sensor.

A scene emits/reflects a given amount of light over a given amount of time. If you record the same portion of that scene (equivalent FOV) for the same amount of time then the amount of light gathered is the same. If you spread it over a larger area it has a lower intensity per (the ISL relationship). But the total light gathered is the same... And that is why the "correct exposure"/meter reading *do not change due to sensor size.
How anyone ever came up with the idea that a larger sensor gathers more light overall baffles me... it simply does/can not (for the same recorded scene/exposure). The issue is primarily the smaller physical size, and that "equivalent resolution" also means it has smaller pixels which suffer much more in a lack of efficiency and read/shot noise.

Someone must have thought "if you record a scene and then take out the DX crop area it has less total light"... that is true, and entirely irrelevant.

* The exposure doesn't change assuming both cameras use the same method of determining ISO (REI/SOS) and have the same/similar manufacturer bias.

 

[HoppyUK]

No, in terms of "equivalence" that is entirely incorrect. It works exactly the same as the ISL, the physics of light doesn't change and it doesn't care about the size of the sensor.

A scene emits/reflects a given amount of light over a given amount of time. If you record the same portion of that scene (equivalent FOV) for the same amount of time then the amount of light gathered is the same. If you spread it over a larger area it has a lower intensity per (the ISL relationship). But the total light gathered is the same... And that is why the "correct exposure"/meter reading *do not change due to sensor size.
How anyone ever came up with the idea that a larger sensor gathers more light overall baffles me... it simply does/can not (for the same recorded scene/exposure). The issue is primarily the smaller physical size, and that "equivalent resolution" also means it has smaller pixels which suffer much more in a lack of efficiency and read/shot noise.

Someone must have thought "if you record a scene and then take out the DX crop area it has less total light"... that is true, and entirely irrelevant.

* The exposure doesn't change assuming both cameras use the same method of determining ISO (REI/SOS) and have the same/similar manufacturer bias.

Edit: my reply isn't right. Apologies. See correction post #22.

I'm not sure we're both making the same comparison here, but the way I'm reading you, that's not right. The comparison of eg a 50mm lens on 1.6x crop and an 80mm lens of full-frame, at same f/number the intensity is the same, but the 80mm lens has a larger diameter aperture at that same f/number. It passes more total light, more 'volume' if you like - a bit more than twice as much.

Put another way, at same f/number and ISO, smaller sensors show more noise. Unquestionable fact (assuming similar generation sensors with similar efficiency). Apologies if I have misunderstood you, and we're some way off-topic now

 

[petersmart]

The comparison of eg a 50mm lens on 1.6x crop and an 80mm lens of full-frame, at same f/number the intensity is the same, but the 80mm lens has a larger diameter aperture at that same f/number. It passes more total light, more 'volume' if you like - a bit more than twice as much.

Which means nothing - at the same f no. the light reaching the sensor (which is the important thing) is the same whether it's a 50mm lens on a crop or an 85mm lens on FF - because you also have to take into account the inverse square law which governs all lenses and is the reason for f nos in the first place.
.

 

[HoppyUK]

Which means nothing - at the same f no. the light reaching the sensor (which is the important thing) is the same whether it's a 50mm lens on a crop or an 85mm lens on FF - because you also have to take into account the inverse square law which governs all lenses and is the reason for f nos in the first place.
.

Yes Peter, you're right. Apologies for my hasty reply above, I was actually over-thinking it in response to Steven. It's the inverse square law that maintains the same brightness/intensity of light. But no... full-frame sensors do collect more light/photons.

The real difference is simply the size/area of the image circle projected by full-frame lenses vs crop, or at least the area captured by the sensor - which is more than double the size on FF. And accordingly, that's why full-frame lenses are much bigger and heavier than crop format lenses of same focal length and f/number.

To take an extreme comparison, look at the very small lenses of huge focal length range and fast apertures fitted to some compacts and megazoom bridge cameras. That is only possible because of the tiny sensor size, and equally, as a direct result nobody would claim they do not have serious noise issues at higher ISOs.

 

[sk66]

The real difference is simply the size/area of the image circle projected by full-frame lenses vs crop, or at least the area captured by the sensor - which is more than double the size on FF

You're forgetting/missing that "the image circle" is the same scene... it has the same source/amount of light.
The fact is, in terms of "equivalence" you are recording the same image (FOV) with the same exposure. And that means the same total amount of light reaches the sensor (the ISL).

Take a scene with a white door or other point of light in it... regardless of how large or small it is w/in the image the correct exposure (total amount of light) for it will be the same (the ISL again). Exactly the same thing happens with a scene/image circle overall, it doesn't matter what size (sensor) you make it. In truth, for any given scene/exposure a larger sensor is exposed to *less light* per area (same overall). But it is much more efficient/effective at collecting that light.

With low light it's like trying to catch random raindrops in a light sprinkle from a hose (fixed qtty/rate)... the smaller sensor with smaller pixels (equivalent MP) is using teacups and the larger sensor is using buckets. There is a much greater probability that some of those teacups will have very little, or even no water (photons) in them and that is the base issue. Combine that with numerous other technical differences (amplification requirement, noise floor, fill factor, etc, etc) and the smaller pixels (sensor) are at a distinct disadvantage. For the same size sensor smaller pixels have much less of a disadvantage, and even potential advantages... primarily due to "oversampling" (combining pixels/information).

You can say that a larger sensor (generally) collects more light due to larger pixels/greater efficiencies, that would be more correct... but it has nothing to do with "area" or the amount of total light it is exposed to.

Like I said, as a generic/simplified way of understanding things Tony's "equivalence" video works ok.
Truth is that I think there are several correlations made that "work," but not really/exactly for the reasons stated. And due to the numerous other variables involved you can never really get true "equivalence," but close enough for generalized comparisons/understanding I guess.

Edit to add:
*IF the larger sensor was actually getting more light then the exposure (settings/metering) would change, it would have to, but it doesn't. "Exposure" doesn't have to do w/ light per as such, it is only affected by total light collected.

 

[HoppyUK]

You're forgetting/missing that "the image circle" is the same scene... it has the same source/amount of light.
The fact is, in terms of "equivalence" you are recording the same image (FOV) with the same exposure. And that means the same total amount of light reaches the sensor (the ISL).

Take a scene with a white door or other point of light in it... regardless of how large or small it is w/in the image the correct exposure (total amount of light) for it will be the same (the ISL again). Exactly the same thing happens with a scene/image circle overall, it doesn't matter what size (sensor) you make it. In truth, for any given scene/exposure a larger sensor is exposed to *less light* per area (same overall). But it is much more efficient/effective at collecting that light.

With low light it's like trying to catch random raindrops in a light sprinkle from a hose (fixed qtty/rate)... the smaller sensor with smaller pixels (equivalent MP) is using teacups and the larger sensor is using buckets. There is a much greater probability that some of those teacups will have very little, or even no water (photons) in them and that is the base issue. Combine that with numerous other technical differences (amplification requirement, noise floor, fill factor, etc, etc) and the smaller pixels (sensor) are at a distinct disadvantage. For the same size sensor smaller pixels have much less of a disadvantage, and even potential advantages... primarily due to "oversampling" (combining pixels/information).

You can say that a larger sensor (generally) collects more light due to larger pixels/greater efficiencies, that would be correct... but it has nothing to do with "area" or the amount of total light it is exposed to.

Like I said, as a generic/simplified way of understanding things Tony's "equivalence" video works ok.
Truth is that I think there are several correlations made that "work," but not really/exactly for the reasons stated. And due to the numerous other variables involved you can never really get true "equivalence," but close enough for generalized comparisons/understanding I guess.

Steven, seriously, no.

"You're forgetting/missing that "the image circle" is the same scene..." No, I'm not. What you're forgetting is that the image projected on to the sensor has the same brightness on a per sq mm basis, but the larger sensor has a lot more sq mms. That's the key.

I can't explain it any more clearly than I have, other than to suggest that you think about what you are saying here, and how the real-world evidence of sensor size, lens dimensions and noise performance contradicts it with bigger/smaller sensors. These things are not even in debate

Have a look at this article on equivalence from DPReview http://www.dpreview.com/articles/2666934640/what-is-equivalence-and-why-should-i-care It's pretty thorough and the only thing missing is re lens sharpness/MTF changes.

 

[sk66]

And accordingly, that's why full-frame lenses are much bigger and heavier than crop format lenses of same focal length and f/number.

They are larger because they have to be optimized for a larger image circle... and that requires larger glass. It's also the same reason that "better lenses" tend to have larger/heavier elements...even for the same size sensor/image circle.

 

[HoppyUK]

They are larger because they have to be optimized for a larger image circle... and that requires larger glass. It's also the same reason that "better lenses" tend to have larger/heavier elements...even for the same size sensor/image circle.

No, you've got the wrong end of the stick

 

[petersmart]

They are larger because they have to be optimized for a larger image circle... and that requires larger glass. It's also the same reason that "better lenses" tend to have larger/heavier elements...even for the same size sensor/image circle.

Better lenses are better not because of their size (look at the Canon 50mm f1.8) but because they are constructed with many different elements in an attempt to cancel out the many distortions they are prey to, including, but not limited to, chromatic aberration, astigmatism, field curvature, barrel and pincushion distortion, and coma.

These problems become even worse when dealing with zoom lenses, which is why many professionals, and serious amateurs with deep pockets, prefer prime lenses.

And when it comes to sensor size it is the pixel density, not the sensor size which is the limitation on noise and resolution which is why higher MP cameras need better quality lenses and, all things being equal, have higher noise levels.
.

 

[sk66]

What you're forgetting is that the image projected on to the sensor has the same brightness on a per sq mm basis, but the larger sensor has a lot more sq mms.

Exposure as we are discussing it here (equivalence) is not "the same brightness per sq mm"... it is total luminance (photons) *divided by* the sq mm (area). I.e. total light over the total area; a larger sensor receives less light per sq mm for a given exposure.

The "potential" for more light with a larger sensor is due to the larger FOV. FF lenses do not gather "more light," they project a "larger area" of light. But if the *captured/source FOV* is the same, then it has the same total amount of light.

There's another step to this equation... The larger sensor with less light per sq mm is going to be enlarged less for the final image, and the smaller sensor with more light per sq mm is going to be enlarged more. The net result is *the same light per sq mm in the final image/output*... the resulting image has the same exposure for any given settings regardless of sensor size, that is the point. And that is how it actually works.

If any of that were not true, then sensor size (or FF/DX lens design) would affect exposure/ exposure settings... and it doesn't. It's all just the ISL going back and forth...

A point of confusion is that exposure *in stops of light* IS photons per square millimeter... i.e. photons per pixel for per pixel brightness (exposure level). But that is the overall exposure/exposure settings. And it is also the same regardless of sensor size.

 

[sk66]

Better lenses are better not because of their size (look at the Canon 50mm f1.8) but because they are constructed with many different elements in an attempt to cancel out the many distortions they are prey to

Yes and no...
The additional elements to correct aberrations are (more) necessary due to smaller size. It is "easier" to create a more perfect lens of larger size and lesser curvature (and for smaller image circles)... but that isn't entirely practical.

 

[HoppyUK]

Steven, no again, and you cannot have read the link to DPReview's article. See page 2, Equivalence of Total Light, and page 4, Real World Total Light Demonstration. It's all there

If I might hazard a guess, you seem to be stuck on the fact that the same subject 'radiates the same amount of light' if I can put it like that, and it doesn't change. What we're discussing is the total amount of light captured by the lens and projected on to the sensor. That does change. A lot.

 

[petersmart]

There's another step to this equation... The larger sensor with less light per sq mm is going to be enlarged less for the final image, and the smaller sensor with more light per sq mm is going to be enlarged more. The net result is *the same light per sq mm in the final image/output*... the resulting image has the same exposure for any given settings regardless of sensor size, that is the point. And that is how it actually works.

But what happens if you put an EF lens designed for a FF camera onto a crop camera?

According to your view the crop sensor will somehow get more light than the FF camera does simply because it is a smaller sensor which just doesn't make sense.

Regardless of the sensor size any lens will transmit the same amount of light for any given stop but the crop sensor with its smaller photodiodes will have to amplify it more to get an image of equal density to the larger photodiodes of the larger sensor (assuming that they are both the same MP).

The smaller sensor doesn't somehow magically get more light than the larger sensor - it just amplifies it more - which, as I have already said, increases the noise levels.
.

 

[Deleted member 49549]

...... a larger sensor receives less light per sq mm for a given exposure.

Snip.

Does it?

 

[woof woof]

I find these debates strangely interesting but ultimately somewhat pointless as people tend to end up talking about slightly different things.

So rather than get bogged down in all this I'd rather avoid it all and this is possible as just about anything you want to know about FF v crop can be settled by 5 minutes shooting them side by side and a then an hour (or less) looking at the results on your PC. If you're lucky enough to have access to FF and crop cameras and lenses.

One thing I will say about this debate is that generally speaking I think it's best to use lenses specifically designed for the format you are going to use them on but sadly APS-C is arguably (but probably) relatively quite poorly served by Canikon with only FF lens options available for some choices of lens type, focal length and or aperture.

Just for fun...

These very similar pictures were taken with FF and MFT and apart from the slightly different framing you have to look at them side by side at high magnification to see any real difference in technical quality.

 

[odd jim]

Does it?

No.

 

[odd jim]

I find these debates strangely interesting but ultimately somewhat pointless as people tend to end up talking about slightly different things.

So rather than get bogged down in all this I'd rather avoid it all and this is possible as just about anything you want to know about FF v crop can be settled by 5 minutes shooting them side by side and a then an hour (or less) looking at the results on your PC. If you're lucky enough to have access to FF and crop cameras and lenses.

One thing I will say about this debate is that generally speaking I think it's best to use lenses specifically designed for the format you are going to use them on but sadly APS-C is arguably (but probably) relatively quite poorly served by Canikon with only FF lens options available for some choices of lens type, focal length and or aperture.

Just for fun...

These very similar pictures were taken with FF and MFT and apart from the slightly different framing you have to look at them side by side at high magnification to see any real difference in technical quality.

In terms of technical quality it's very noticeable between my Canon FF and my Canon crops, the FF body gives more dynamic and more over, sharper more contrasty images, and I'm not talking 100% pixel peeping, just normal sizes. And the clarity and noise levels are far better as you'd expect (on a lens for lens basis), even at lower ISOs such as 400. Aside from the technical merits, the end result is far more keepers.

You can't honestly say the differences in a like for like MFT and FF DSLR are negligible unless you pixel peep because they simply aren't, as good as modern MFT cameras are (this is coming from someone who uses everything from 1" compacts to FF and almost everything in between).

The FF body you used in the above wasn't your old 5d was it? The 5d didn't perform very well for me on a technical level even compared to my cropped Canons, I found it very underwhelming.

 

[petersmart]

These very similar pictures were taken with FF and MFT and apart from the slightly different framing you have to look at them side by side at high magnification to see any real difference in technical quality.

And, as you say the differences are minimal - but that is on a very bright sunny day where the ISO, I would guess, was quite low - but what about when the ISO has to go up to about 800-1600 or higher for indoor shots without flash?

I think then that the differences between the two would be more obvious - even more if you had to crop.
.

 

[woof woof]

In terms of technical quality it's very noticeable between my Canon FF and my Canon crops, the FF body gives more dynamic and more over, sharper more contrasty images, and I'm not talking 100% pixel peeping, just normal sizes. And the clarity and noise levels are far better as you'd expect (on a lens for lens basis), even at lower ISOs such as 400. Aside from the technical merits, the end result is far more keepers.

You can't honestly say the differences in a like for like MFT and FF DSLR are negligible unless you pixel peep because they simply aren't, as good as modern MFT cameras are (this is coming from someone who uses everything from 1" compacts to FF and almost everything in between).

The FF body you used in the above wasn't your old 5d was it? The 5d didn't perform very well for me on a technical level even compared to my cropped Canons, I found it very underwhelming.

And, as you say the differences are minimal - but that is on a very bright sunny day where the ISO, I would guess, was quite low - but what about when the ISO has to go up to about 800-1600 or higher for indoor shots without flash?

I think then that the differences between the two would be more obvious - even more if you had to crop.
.

The pictures were taken with a Sony A7 and Panasonic GX7 (f8 and f4 respectively) and weren't meant to be taken all that seriously, but...

At low to mid or even quite high ISO's when viewing normal sized images normally (it all hangs on what's normal though ) I'd say that differences can be minimal when the pictures are processed for best effect. Process them all the same and differences may be more apparent but personally I process images for best effect, I don't apply the same settings to every camera and picture and I doubt others here do either.

Yes, when you go to extremes like huge prints, heavy crops, very high dynamic range scenes and higher ISO's differences will be more visible but unless you're shooting at ISO 6400 and above and want to print large or crop like crazy and get gallery quality the differences may not be too critical. The smaller formats may lack a little dynamic range and if pixel peeping the larger format may be sharper and in the case of those pictures the A7 picture is sharper when viewed at 100% but maybe that's not a critical measure. YMMV. These differences may or may not matter but the quality you can get out of almost any half way respectable camera these days IMO surpasses what I got from 35mm film and easily so ISO 1600 film.

Regardless of the format filling the frame and so avoiding heavy crops may help, using lenses designed for the format may help and another thing you can do is ETTR if possible and back it off post capture if you think that'll help.

Is all that cheating? Well, I did say something to the effect that shooting with the cameras and looking at the results should get you there and it avoids all this mental gymnastics.

 

[juggler]

Sorry guys, thanks for your responses however i just meant if this guy is stating that you get a more pleasant perspective from an 85mm on a crop sensor then why would he be advocating the use of 85mm on full frame instead of 135 on full frame which would give him the same pleasant perspective he seems to prefer?

I haven't read the article... no idea, tbh, 135 is a great length for portraits if you've got the room to use it.

 

[pjm1]

I haven't read the article... no idea, tbh, 135 is a great length for portraits if you've got the room to use it.

Or a 150 macro if you haven't

 

[HoppyUK]

The best focal length for (solo) portraits is the lens that gives enough distance for attractive perspective, say a minimum of 5ft, with the framing you want. There's a big difference between a tight headshot and waist up. Personally, I like to be around 5-6ft - close enough for good communication, without being intimidating. On full-frame, lens could be anywhere between 75-135mm-ish.