comparining micro four thirds to full frame

[Ray]

Ray, are you aware that DR or ISO (noise in general) differences, are becoming negligible with time?

Guillermo,
You're sounding like someone who has sacrificed DR capability for other features which you find more attractive or useful, and are therefore in a state of denial about the benefits of what you've sacrificed. 

I do realize that DR issues have become less of a problem as the DR capability of cameras in general has increased. When I was using the Canon 5D and Canon 50D on a regular basis, I used to bracket almost all shots automatically because I was so fed up with the sight of unpleasant noise in the shadows when I processed the images on the computer. Those were also the days, because of limited DR, when ETTR processes were a bit of an obsession. If DR is a problem, then underexposing by half a stop makes the problem worse.

Now that I use Nikon cameras, which usually have the best DR available, I feel free of any obsessions with ETTR and the need to bracket exposures in case there is noise in the shadows.

Do you understand that a difference of 2,5 stops in DR between cameras with 11 and 13,5 stops of DR has not the same real world effect as 2,5 stops between a couple cameras with 8 and 10,5 stops we used to have years ago?.

Do you understand that people often use higher-than-base ISOs, such as ISO 400, 800 and 1600 and that the DR at these ISOs is sometimes as bad as the DR at the base ISO in older cameras?

The attached DXOMark graph shows a difference of at least one full stop in DR at all ISOs from 200 to 3200, comparing the D850 with the Olympus OM-D E-M1 MkII. Do you remember the days when people got so excited when the first Nikon full-frame DSLR (the D3) had a 1/2 stop better DR than the best Canon models, at high ISOs? The Nikon D3 still has 1/2 a stop better DR than the OM-D E-M1, at the nominal ISO of 800. But not at other ISOs, so you can relax. 

Real world scenes are not increasing their dynamic range, typical poorly lit scenes are not becoming more poorly lit with time. This means that the advantages in noise of larger sensors are vanishing in practical terms.

I think you are underestimating the full dynamic range of many scenes, as the eye perceives them. The eye is continuously adjusting to the lighting conditions as it peruses a scene. A simple example would be a shot of one's living room which takes in the view of a sky and landscape, with bright clouds, through a large window.

Without additional lighting in the living room, or the bracketing of exposures, there would probably be lots of noise in the photographic shot of the living room, or a blown sky. However, the eye when perusing such a scene, sees no such noise because it is in effect doing its own bracketing. As the gaze shifts to the bright clouds, the eye's pupils narrow in diameter. When the gaze shifts to details in the living room, the pupil widens and the eye takes in more light. It's estimated this process, which is very quick, gives the human eye a dynamic range capability of about 20 stops.

[Guillermo Luijk]

I do realize that DR issues have become less of a problem as the DR capability of cameras in general has increased.

Indeed, this is the only thing you needed to say to confirm that you understand that the differences in DR/ISO FF vs M4/3 are becoming irrelevant in practice Ray, and will continue to do so. Thanks.

[hogloff]

Indeed, this is the only thing you needed to say to confirm that you understand that the differences in DR/ISO FF vs M4/3 are becoming irrelevant in practice Ray, and will continue to do so. Thanks.

Well if for your type of photography DR is becoming irrelevant...then that's great for you. Others might shoot different types of scenes where DR, even with the top performing cameras, is still a challenge. Your generalization is based on your experience, not others.

[Ray]

Indeed, this is the only thing you needed to say to confirm that you understand that the differences in DR/ISO FF vs M4/3 are becoming irrelevant in practice Ray, and will continue to do so. Thanks.

No. You misunderstand me. When one is aware of the DR limitations of one's equipment, one tends to avoid certain subjects, or perhaps blacken the shadows instead of revealing noisy detail. Better DR allows one to tackle new types of scenes and perhaps be more innovative.

I agree with Hogloff in this respect.

[NancyP]

All this is to say that you can't cheat Ma Physics. 

Larger format telephoto lenses (refractors) of equivalent field of view and equivalent f/ are....larger, and being made of glass, significantly heavier.
Larger format lenses have less depth of field at equivalent field of view and equivalent f/. This is either an advantage or disadvantage, depending on your subject. Non-photo-gear-obsessed insect and other macro-subject naturalists often like the depth of field advantage of point-and-shoot cameras, as well as their convenience in the field.
Larger pixels will accommodate more photons, therefore, noise can be lower, color accuracy potentially higher, dynamic range higher, other things being equal. (Fantasy wish: full pixel dump and re-zeroing in real time, at least for some lower shutter speeds).
Larger format cameras are....larger, and generally heavier. I like the handling of a "prosumer"-size DSLR, the Canon 6D being an example, it fits my hand, the dinky Rebels feel slightly "crowded". On the other hand, I grew up with a much smaller film SLR camera, and liked its handling at that time. Conspicuousness might be a concern with larger cameras (street and PJ, travel in areas with lots of thieves, etc).

[Guillermo Luijk]

Well if for your type of photography DR is becoming irrelevant...then that's great for you. Others might shoot different types of scenes where DR, even with the top performing cameras, is still a challenge. Your generalization is based on your experience, not others.

If for your type of photography DR is fundamental, but your M4/3 has 20 stops of DR vs 22,5 stops in the FF, both cameras would be identically capable of capturing ANY real world scene. If sensors have more and more DR on every yearly iteration, there will be less and less real world scenes where FF will make any difference vs M4/3, so the "FF has more DR" story is becoming more and more irrelevant year after year. That is what I’m trying to expose.

According to DxOMark, two shots on my 2005 Canon 350D 4EV apart capture more DR than a single shot on a Nikon D850 at ISO64.

Regards

[hogloff]

If for your type of photography DR is fundamental, but your M4/3 has 20 stops of DR vs 22,5 stops in the FF, both cameras would be identically capable of capturing ANY real world scene. If sensors have more and more DR on every yearly iteration, there will be less and less real world scenes where FF will make any difference vs M4/3, so the "FF has more DR" story is becoming more and more irrelevant year after year. That is what I’m trying to expose.

According to DxOMark, two shots on my 2005 Canon 350D 4EV apart capture more DR than a single shot on a Nikon D850 at ISO64.

Regards

Sure...when we reach that 20 stops of DR...yes it will become irrelevant...but until then...it is relevant. Very much like high ISO noise reaching levels where it becomes irrelevant...but until that time...it too is still relevant.

[Telecaster]

My general observation is that the less real-world relevant a particular spec or feature becomes, the more emphatic people become in their insistence upon its continued relevance.     Just wait 'til we're comparing Camera System A with 24.9 stops of usable DR at ISO 102800 to Camera System B with 24.8 stops of same at same. There will be wars fought over this.

-Dave-

[hogloff]

My general observation is that the less real-world relevant a particular spec or feature becomes, the more emphatic people become in their insistence upon its continued relevance.     Just wait 'til we're comparing Camera System A with 24.9 stops of usable DR at ISO 102800 to Camera System B with 24.8 stops of same at same. There will be wars fought over this.

-Dave-

What are these “less real world relevant” specs that you are talking about?

[BJL]

BJL,
You are still not explaining yourself clearly enough. If one is using equivalent prime lenses with two different formats which have a different pixel pitch, the pixel count will inevitably be different. Adjusting the crop factor in order to equalize the pixel count causes the 'field of view' to be different. Downsizing or upsizing doesn't change the field of view.

Ray, I am not talking about equivalent prime lenses: I am not sure where you get that idea; the lenses that I have mentioned are a 12-50, 75-300, 12-100 and 100-400 zooms. My point is the simple fact (of which you are aware, as you have referred to it in other posts) that when a larger sensor with more pixels and pixel pitch 1.4x larger, getting the same pixel count on the same subject [the same angular resolution] require forming an image at the focal plane that is about 1.4x larger (linear), and so using a focal length 1.4x longer, and then cropping as needed (or using a longer focal length and so forming an even larger image, and then downsampling).

For example to match the telephoto reach of the long end of the Olympus 12-100mm on a 20MP MFT body like the EM1 Mk II would, on a D850, need at least 140mm, cropped to match FOV, and to match the long end of the Panasonic 100-400mm would require at least 560mm.

That said, your proposed matching of that 12-100 and 10-400 by a Nikon 24-120 plus [discontinued] Sigma 150-500 is close but falls a bit short: some combination like 24-140 and 200-560 would be needed. At the long end, one option I see is the Sigma 150-600/5-6.3 DG OS HSM | C.

Some weight comparisons:

Olympus 12-100/4   561 g
vs
Nikon 24-120/4      710 g
    Just a bit heaver and "shorter" in reach.

Panasonic 100-400/4-6.3            985 g, $1800. C.f. 200-560
vs
Sigma 150-500/5-6.3               1780 g
    A lot heaver and a bit shorter in reach.
or
Sigma 150-600/5-6.3 DG OS HSM | C      1930 g
    Even heaver, with a bit more reach.

As Nancy points out, Physics allows no free lunch in either direction: a larger sensor with more and larger photosites, paired with bigger, heaver (and usually more expensive) lenses can given results that look better under some circumstances.


P. S. Note that the EM1 Mk II has a pixel shifting high resolution mode matching about 40MP [Correction: about 50MP; I was thinking the EM5 Mk II], so if and when that can be used (tripod, good enough lens resolution, etc.) we are back fairly close to straight focal length equivalents, so needing a pair of F-mount lenses covering 24mm to about 800mm.

[armand]

This feels like deja-vu.
Here are more weights and thoughts:
http://forum.luminous-landscape.com/index.php?topic=110045.0
http://forum.luminous-landscape.com/index.php?topic=102175.0

[kers]

According to DxOMark, two shots on my 2005 Canon 350D 4EV apart capture more DR than a single shot on a Nikon D850 at ISO64.
Regards

I often need 3 shots of a d850 2EV apart...  ( there were times in the low-DR-past i needed to blend 5 to even 7 exposures)
2 ev apart because the blending is done more easy...

+ would not want to miss the large prism viewfinder- one reason i would not go to dx format.

[Ray]

Ray, I am not talking about equivalent prime lenses: I am not sure where you get that idea; the lenses that I have mentioned are a 12-50, 75-300, 12-100 and 100-400 zooms. My point is the simple fact (of which you are aware, as you have referred to it in other posts) that when a larger sensor with more pixels and pixel pitch 1.4x larger, getting the same pixel count on the same subject [the same angular resolution] require forming an image at the focal plane that is about 1.4x larger (linear), and so using a focal length 1.4x longer, and then cropping as needed (or using a longer focal length and so forming an even larger image, and then downsampling).

For example to match the telephoto reach of the long end of the Olympus 12-100mm on a 20MP MFT body like the EM1 Mk II would, on a D850, need at least 140mm, cropped to match FOV, and to match the long end of the Panasonic 100-400mm would require at least 560mm.

That said, your proposed matching of that 12-100 and 10-400 by a Nikon 24-120 plus [discontinued] Sigma 150-500 is close but falls a bit short: some combination like 24-140 and 200-560 would be needed. At the long end, one option I see is the Sigma 150-600/5-6.3 DG OS HSM | C.

Some weight comparisons:

Olympus 12-100/4   561 g
vs
Nikon 24-120/4      710 g
    Just a bit heaver and "shorter" in reach.

Panasonic 100-400/4-6.3            985 g, $1800. C.f. 200-560
vs
Sigma 150-500/5-6.3               1780 g
    A lot heaver and a bit shorter in reach.
or
Sigma 150-600/5-6.3 DG OS HSM | C      1930 g
    Even heaver, with a bit more reach.

Okay! I get your point, BJL. In order to ensure that the D850 does not have a resolution disadvantage at the long end of the equivalent zoom, we need to use a longer focal lenth so that we need to crop the D850 image less, getting the same FOV whilst simultaneously achieving at least an equal pixel count of 20mp in both images.

That of course does increase the combined weight of the D850 system. I didn't know the Sigma 150-500 had been discontinued. If we use the Sigma 150-600 instead, the additional weight is only 150 grams. At 600mm, the D850 image, after cropping to the same FOV as the uncropped 400mm image from the EM1 MkII, would have more than 20mp, and each of those pixels is of better quality than the EM1 MkII pixel.

Lens quality issues are also a factor. It would be interesting to see a real-world comparison between the two lenses at the long end, on the different bodies. There might not be much difference at the long end but there sure will be at shorter focal lengths when the D850 doesn't need to be cropped at all, or might need to be be cropped only slightly to get the same FOV.

Adding up the weights of all these lenses and bodies, the 24-120 and 150-600 with the D850, and the 12-100 and 100-400 with the E-M1 MkII, the combined weights are; 2.12 Kgs for the MFT system, and 3.55 Kgs for the 35mm system.
That's an extra 1.43 Kgs for all the benefits of the much larger sensor with much higher pixel count most of the time when using zoom lenses, and better pixel quality.

As I understand, the main argument in favour of the Olympus 4/3rds system is that it provides sufficient image quality for your purposes, with an over all saving in weight and size, and therefore greater convenience of use. I can appreciate that.

If I were taking up digital photography as a new hobby, I would definitely consider opting for the Olympus 4/3rds system, now that they have a 20mp sensor with reasonably good quality pixels. However, in my current circumstances, already owning several Canon and Nikkor lenses, I would not be prepared to switch systems, unless the new system had no disadvantages, and only advantages.

P. S. Note that the EM1 Mk II has a pixel shifting high resolution mode matching about 40MP [Correction: about 50MP; I was thinking the EM5 Mk II], so if and when that can be used (tripod, good enough lens resolution, etc.) we are back fairly close to straight focal length equivalents, so needing a pair of F-mount lenses covering 24mm to about 800mm.

This is an interesting development. However, one way that I reduce the weight of my camera system nowadays, for the sake of convenience and the reduction of a lot of messing around, is not to bother with tripods anymore, unless they are really essential for a particular type of effect, such as the blurring of the water in a waterfall or taking shots at night, or in very poor light.

My impression is that pixel shifting requires a very stable tripod to avoid the slightest movement. Stable tripods tend to be heavy. That 1.43Kg of weight saving, which is the main justification for choosing the lighter MFT system, will be cancelled, will it not? 

Also, pixel-shifting is only successful with static subjects, a bit like bracketing exposures for merging to HDR, so it's use is rather limited.

Have I made an irrefutable case? 

P.S. Having done a bit of investigation on the pixel-shift feature of the E-M1 MkII, I'm getting the impression that the higher resolution is only significant when exceptionally good prime lenses are used. 90% of all lenses don't make the grade.

[Ray]

There's another factor which seems to be overlooked by the 'DR-benefit deniers'. 
Let's say one camera in the comparison has a DR of 13 EV, and the other camera has a DR of 10 EV, using the same standards of analysis.

Some folks might argue that 10 stops of DR is quite sufficient for their purposes. However, it's not necesarily the case that both cameras will produce acceptably good detail and low noise in those parts of the image which are at the sensors' DR limit.

The real advantage in practice, of the camera with 13 EV of DR, is that those parts of the scene that are underexposed by 10 EV, or 9 EV or 8EV will tend to be cleaner and more detailed than the same parts of the scene shot with the camera with a DR limit of only 10 EV.

Noise is something which increases gradually in accordance with the different degrees of light levels in the different parts of the scene.

The concept of PDR (Photographic Dynamic Range) attempts to address this issue, although I still prefer the engineering results shown in the DXOMark graphs because the comparisons are so easy to see, and cover so many different models of cameras, at both the pixel level and at a downsampled standard which allows comparison of the entire sensor.

I've actually taken the trouble to do my own tests, so I know just how deteriorated the image looks at the engineering DR limit of the camera, but I also know how much better the image looks from the camera with a DR limit of 13EV, when the image, underexposed by 9 stops, is compared with the same scene from a camera with a DR limit of 10EV, when that image is also underexposed by 9 stops.

Hope that's not too confusing. 

[HSakols]

This has been an interesting discussion.  Ray, I hope I haven't been too snarky. I'm no engineer and I'm kind of playing devils advocate because I'm trying to understand the real world differences.  Over the long weekend I only used my Olympus because I had to climb 4300 ft. in 7 miles to get to where I wanted to photograph. Yes, I would have liked a full frame, but I can no longer carry one that far.

[Slobodan Blagojevic]

Ray, in relation to the graph below, I think what proponents of m4/3 are saying is this: if you have to achieve the same DOF, you'd have to double the ISO on a full-frame camera, therefore shifting the Nikon line to the left, bringing it closer to the Olympus one. No?

[hogloff]

Ray, in relation to the graph below, I think what proponents of m4/3 are saying is this: if you have to achieve the same DOF, you'd have to double the ISO on a full-frame camera, therefore shifting the Nikon line to the left, bringing it closer to the Olympus one. No?

Why would one have to double the ISO. Most of my landscape images are taken at base ISO.

[Slobodan Blagojevic]

Why would one have to double the ISO...

That's not even a serious question, but I'll repeat what I said in the above post: "to achieve the same DOF."

[NancyP]

Do what you need to do to get the photograph you want. Not the photograph someone else says is "ideal" (unless you are a pro selling that photo), the photo YOU want.

Be grateful you aren't Curtis dragging around mules carrying glass plates and a portable darkroom through Southwest canyons. 
Of course, there are modern Curtis types driving truck-cameras... https://www.kickstarter.com/projects/1376821505/invisible-light-the-dolomites-in-ultra-large-forma (no endorsement - just a google search for "large format truck camera")

[armand]

Why would one have to double the ISO. Most of my landscape images are taken at base ISO.

That's not even a serious question, but I'll repeat what I said in the above post: "to achieve the same DOF."

While less likely for the dedicated landscape photographers I've been in this situation multiple times while hiking. Many times you see different opportunities and on long trails you don't have time to keep messing with your tripod and you just shoot handheld. A good stabilization system and more DOF for the same ISO/aperture can be quite helpful.