Required Read

Okay, I don't have time to answer all of that
Don't have the time, or don't have the ability, since you realize that you made a number of factually incorrect statements that will be impossible for you to defend. Probably just easier to admit that incontrovertible fact and move on, but it's your call.
Sorry, good doctor, but you're wrong again.
Oh hey, and look, there you are talking about noise in 'dark shadows'. Dunno Doc, but I think you need to get your story AND your facts straight.
[/QUOTE]
 
Last edited:
I wouldn't disagree with any of your comments (which I don't see as disagreeing with me), although it would be interesting for someone to measure lp/mm for FF lenses converted onto m43.
You mean "converted" in the sense of adapted with a focal reducer or what?
If you re-read the first sentence of my next paragraph you'll see I'm talking about an adapter, rather than a focal reducer (specifically this: http://www.metabones.com/products/details/MB_EF-m43-BM1). I was looking to get a high quality Telephoto option so don't want the focal length reduced. This puts the 16MP of my GH4 squarely in the centre of the lens at the same pixel density as a 64MP FF camera. Hence my 70-200 f2.8 becomes a 140-400 equiv that I'd expect to be very sharp and the 300 is even more fun (but still big). According to etendue (e.g. http://www.clarkvision.com/articles/telephoto.system.performance/index.html ) it might be pretty reasonable as the f2.8 lenses are good wide open.
I'm wondering whether to get a metabones EF-m43 smart adapter (so a speedbooster without the lens) as a way to hook some of my very sharp Canon lenses on the GH4. The stabiliser would work but it's MF only. Would be interesting. I'm sure the 70-200 II and 300 have sharpness to spare (as they are still sharp with added 1.4x extenders, showing there is extra available, plus this route adds no glass so should be better than that). However the adapters really aren't cheap so I'm not sure. Certainly would be very challenging with some subjects, then again my old film stuff was all MF (ah, but nicer focusing screens though).
A well-made generic focal reducer, as the Metabones seems to be, would of course improve the resolution of an FF lens on an MFT sensor compared to what you can expect without the focal reducer. But it wouldn't reach parity with what you can get with an FF lens on FF since the focal reduction is 0.71 rather than 0.5. At 0.5 it might get close but a generic focal reducer would have to be really perfect to reach full parity.
 
I'm sure you knew I meant E-M1.

I'm not sure about the SNR graph as I think it's normalised in some way. The 1Ds mk II certainly captures loads more electrons than a E-M1 so will have less shot noise and the read noise isn't enough to matter.
 
I'm still here as just had a long phone call but must get to the shops pre rush-hour. I'm not sure why you feel the need for the rude tone but I'm not trying for anything like that so please don't think I'm arguing rather than discussing.

As I said somewhere above you'll get about 1.75 more stops of base ISO DR from any of the current FF Nikons, which is shadow noise. You'll also get circa 2 stops better mid-tone noise due to extra light gathering. Old FF cameras won't do as well for shadow noise but will do much better for for mid-tone noise unless someone altered Physics while I wasn't looking (are you suggesting the much larger number of electrons captured will in some way have more than their square root of noise?).
 
Really not wanting to talk much about equivalency, as it's been done to death... however focal length has nothing to do with the amount of light gathered that just the f-stop (well, strictly the T-stop) and that measures light per unit area. Hence at a particular f-stop a bigger sensor gathers more light in proportion to its area advantage over a smaller sensor.
As long as the quantum efficiency of the two sensors is matched.
As I mentioned below (in another post), not so much, as QE has improved about 1.5x over the last few years (2x if you pick your examples carefully, which is cheating of course) and (for example) FF is 4x the m43 area.
How sensors develop over time is irrelevant for a comparison between sensors of different size at any one point in time.
Also some m43 (e.g. E-M3) don't have great QE anyway (48%).
You mean the E-M1 (there is no E-M3 yet). The QE figures for current non-BSI sensors are very similar. But MFT sensors have better read-noise figures than FF sensors, which means that their DR (and thus signal-noise performance in the shadows) is better than we can expect based on equivalence reasoning.
Yes, I did mean E-M1, I'm not sure the read noise difference is that significant though, my GH4 is 2.5 electrons per pixel and a D610 2.9 and with a lot more well capacity. BTW I don't know about you but I find the sensorgen numbers for the E-M1 odd especially as it's supposed to be the same sensor as the GH4, they must be doing something unusual with it (16K capacity vs. 26K for the GH4).
This means a FF camera with a f1.4 lens captures about 4x the light a m43 camera will (and so has half the Photon Shot Noise,
And half the DOF, of course, which may also be important
Only with equivalent focal length, otherwise a very similar DoF (depending on whether you are looking at the whole photo or the pixels, but pretty similar either way).
Why should we compare at anything but equivalent focal length?
He made a blanket statement about f1.4 lenses and as he'd been having a pop at me for blanket statements I thought I'd mention the caveat he hadn't.
which is the main noise in photographs outside the deep shadows - BUT may be at a low enough level you don't care however small a sensor you have, or may not).
 
I'm still here as just had a long phone call but must get to the shops pre rush-hour. I'm not sure why you feel the need for the rude tone but I'm not trying for anything like that so please don't think I'm arguing rather than discussing.
Doc, I find your tendency to cut out conversation history that makes you look bad, then set up straw men and refuse to respond to various facts I have pointed out bordering on disingenuous.

It would be nice if you'd keep our conversation intact. The fact that you have refused to do that in the past couple of replies strikes me as simply an attempt to hide the evidence of the multiple errors I've caught you out in and squarely proven.
As I said somewhere above you'll get about 1.75 more stops of base ISO DR from any of the current FF Nikons, which is shadow noise. You'll also get circa 2 stops better mid-tone noise due to extra light gathering. Old FF cameras won't do as well for shadow noise but will do much better for for mid-tone noise unless someone altered Physics while I wasn't looking (are you suggesting the much larger number of electrons captured will in some way have more than their square root of noise?).
And now you're switching over to strawmen. How about you stick to what I said, rather than trying to rephrase improperly ;)

Since you refuse to answer the various factual points I have made, and which have incontrovertibly shown several of your statements to be factually incorrect, I will simply invite other readers to go back through our conversation history and read the posts of mine in which my evidence has squarely shown a number of your statements to be technically incorrect.
 
I'm sure you knew I meant E-M1.
I didn't - could have meant the E-M1. Could have meant the E-M5. Could have meant the E-3...
I'm not sure about the SNR graph as I think it's normalised in some way.
Normalized in what way? Maybe you should read DXO's literature on how they test and what their graphs mean.
The 1Ds mk II certainly captures loads more electrons than a E-M1 so will have less shot noise and the read noise isn't enough to matter.
Right, but the several statements you made that I proved, incontrovertibly, to be factually incorrect, are not changed by any of that.

I guess I can't blame you for cutting all of that conversation history out, since you don't come out looking all that great in it, but I don't consider that to be a really honest way of responding.
 
Last edited:
I would say including bracketing is pretty limiting, by restricting yourself to tripods and low-movement scenes, so not generally applicable. It's quite interesting seeing some of the bracketed scenes people shoot on the Thames at night. They think nothing moved so it's good but it isn't as the boats move up and down due to other boats that passed by a while back. That also screws with long exposures trying to make the water super-smooth of course.
Three questions:

1. What makes you think a tripod is necessary for bracketing? I do it hand-held all the time. This is not a problem as long as the longest shutter speed in the bracket can be hand-held without difficulty. The shots used for the stacked image below, for example, were hand-held.



2. What makes you think subject movement is a problem? You can tell the software to use a specific frame in the sequence to render a certain object or area. Alternatively, you can ask the software to make unwanted moving objects disappear.

3. Remember that the starting point you, not I, chose here is the one where shutter speed is not a restriction so that FF can get the same DoF as MFT without leaving base ISO by using four times as long an exposure time. Exactly what is it that we can't accomplish with bracketing and MFT in that scenario? As far as I can see, all the MFT user has to do to reach parity or better is to make sure the longest exposure time in the bracket is as long as that employed by the FF user.
Also your DR graphs would look quite different if you did DR for mid-gray (say) to get an idea of mid-tone image noise.
DR is by definition not about middle gray. It is a measure of SNR in the shadows, where the SNR is always weakest. This is therefore where most people find noise most troubling and therefore effectively sets the floor for acceptable image quality. Have a look at this report from a poll where people were asked to judge the quality of two images with regard to noise, one of which one EV better for midtone SNR, the other one EV better for DR. The large majority found the image with one EV better DR preferable as far as noise is concerned.
(They don't as they aren't measuring the noise in the light, although photographers usually have to include that, including me as my house no longer has any coal storage facilities.)
Not sure I understand what you have in mind here.
 
Last edited:
I wouldn't disagree with any of your comments (which I don't see as disagreeing with me), although it would be interesting for someone to measure lp/mm for FF lenses converted onto m43.
You mean "converted" in the sense of adapted with a focal reducer or what?
If you re-read the first sentence of my next paragraph you'll see I'm talking about an adapter, rather than a focal reducer (specifically this: http://www.metabones.com/products/details/MB_EF-m43-BM1). I was looking to get a high quality Telephoto option so don't want the focal length reduced. This puts the 16MP of my GH4 squarely in the centre of the lens at the same pixel density as a 64MP FF camera. Hence my 70-200 f2.8 becomes a 140-400 equiv that I'd expect to be very sharp and the 300 is even more fun (but still big). According to etendue (e.g. http://www.clarkvision.com/articles/telephoto.system.performance/index.html ) it might be pretty reasonable as the f2.8 lenses are good wide open.
OK. Sorry about the misunderstanding. My fault.

But without a focal reducer, the resolution prospects gets worse. Using just the center of the image circle from a lens made for a bigger format is generally inferior to using the whole image circle from a lens made for the smaller format. It follows that your FF 70-200/2.8 when used on MFT is highly unlikely to come close to the resolution provided by a high-quality MFT tele zoom like the new 40-150/2.8.
I'm wondering whether to get a metabones EF-m43 smart adapter (so a speedbooster without the lens) as a way to hook some of my very sharp Canon lenses on the GH4. The stabiliser would work but it's MF only. Would be interesting. I'm sure the 70-200 II and 300 have sharpness to spare (as they are still sharp with added 1.4x extenders, showing there is extra available, plus this route adds no glass so should be better than that). However the adapters really aren't cheap so I'm not sure. Certainly would be very challenging with some subjects, then again my old film stuff was all MF (ah, but nicer focusing screens though).
A well-made generic focal reducer, as the Metabones seems to be, would of course improve the resolution of an FF lens on an MFT sensor compared to what you can expect without the focal reducer. But it wouldn't reach parity with what you can get with an FF lens on FF since the focal reduction is 0.71 rather than 0.5. At 0.5 it might get close but a generic focal reducer would have to be really perfect to reach full parity.
 
Last edited:
I'm still here as just had a long phone call but must get to the shops pre rush-hour. I'm not sure why you feel the need for the rude tone but I'm not trying for anything like that so please don't think I'm arguing rather than discussing.

As I said somewhere above you'll get about 1.75 more stops of base ISO DR from any of the current FF Nikons, which is shadow noise.
Certainly not from the D4S and Df where the difference in comparison with the E-M1 is down to nearly nothing. The FF Nikons with higher pixel counts do well at base ISO but instead have their efficiency deficit at higher ISOs. Regardless of which FF sensor you pick, you will find that it does not realize the 2 EV advantage at the same exposure that equivalence leads us to expect across the entire ISO range. Some come close to that at base ISO and others at higher ISO but none comes close to it at low and high ISO alike.
You'll also get circa 2 stops better mid-tone noise due to extra light gathering. Old FF cameras won't do as well for shadow noise but will do much better for for mid-tone noise unless someone altered Physics while I wasn't looking (are you suggesting the much larger number of electrons captured will in some way have more than their square root of noise?).
 
Really not wanting to talk much about equivalency, as it's been done to death... however focal length has nothing to do with the amount of light gathered that just the f-stop (well, strictly the T-stop) and that measures light per unit area. Hence at a particular f-stop a bigger sensor gathers more light in proportion to its area advantage over a smaller sensor.
As long as the quantum efficiency of the two sensors is matched.
As I mentioned below (in another post), not so much, as QE has improved about 1.5x over the last few years (2x if you pick your examples carefully, which is cheating of course) and (for example) FF is 4x the m43 area.
How sensors develop over time is irrelevant for a comparison between sensors of different size at any one point in time.
Also some m43 (e.g. E-M3) don't have great QE anyway (48%).
You mean the E-M1 (there is no E-M3 yet). The QE figures for current non-BSI sensors are very similar. But MFT sensors have better read-noise figures than FF sensors, which means that their DR (and thus signal-noise performance in the shadows) is better than we can expect based on equivalence reasoning.
Yes, I did mean E-M1, I'm not sure the read noise difference is that significant though, my GH4 is 2.5 electrons per pixel and a D610 2.9 and with a lot more well capacity.
The well capacity is no bigger for equivalent photos so there is still a DR deficit for the larger sensor, just a bit smaller than it is in comparison with the E-M1. As to the significance DR and thus read noise more generally, see my other reply to you here.
BTW I don't know about you but I find the sensorgen numbers for the E-M1 odd especially as it's supposed to be the same sensor as the GH4, they must be doing something unusual with it (16K capacity vs. 26K for the GH4).
The same sensor can be utilized differently and there might also be slightly different versions of the same sensor. At any rate, there is nothing fishy about the E-M1 DR values as far as I can tell. I have checked them myself using black-frame data and my findings are well in line with those of DxO.

What is fishy, however, are the DxO (and therefore Sensorgen) figures for the GH4 at ISO 100. I just verified that the GH4 shows the same peculiarity as the GX7, GM1, and GM5, i.e., that it does not reach the saturation limit set by the ADC at ISO 100 (ISO 125 for the GX7 and GM1) because it runs into the limit set by the full-well capacity before that. In all likelihood, this means that the FWC isn't 26 K but something much closer to the FWC of the E-M1.

The DR values for ISO 100 may well be wrong too. When I checked the DR of the GX7 by means of black-frame data and the actual max signal, I found it to be about half an EV lower at ISO 125 than at ISO 200 rather than about 0.2 EV higher as DxO reports.
This means a FF camera with a f1.4 lens captures about 4x the light a m43 camera will (and so has half the Photon Shot Noise,
And half the DOF, of course, which may also be important
Only with equivalent focal length, otherwise a very similar DoF (depending on whether you are looking at the whole photo or the pixels, but pretty similar either way).
Why should we compare at anything but equivalent focal length?
He made a blanket statement about f1.4 lenses and as he'd been having a pop at me for blanket statements I thought I'd mention the caveat he hadn't.
But what is relevant here are comparisons at equivalent focal lengths, right?
which is the main noise in photographs outside the deep shadows - BUT may be at a low enough level you don't care however small a sensor you have, or may not).
 
Last edited:
I did reply rather than reply with quote rather than cutting it out, it was just getting very long and you were being deliberately rude and arguing special cases. It's quite rare for me to use Reply with quote as it gets crazy long after a while so you should more feel flattered earlier. The only reason I looked bad is your arguments, which I would not agree with in a lot of cases, due to being a selection of non-applicable, straw man (i.e. not arguing against my point but a different point) or only covering limited scenarios. Still feel free to live in a m43 is king world, a lot of people do. I like mine but find they come with limitations as well as advantages.
 
I wouldn't disagree with any of your comments (which I don't see as disagreeing with me), although it would be interesting for someone to measure lp/mm for FF lenses converted onto m43.
You mean "converted" in the sense of adapted with a focal reducer or what?
If you re-read the first sentence of my next paragraph you'll see I'm talking about an adapter, rather than a focal reducer (specifically this: http://www.metabones.com/products/details/MB_EF-m43-BM1). I was looking to get a high quality Telephoto option so don't want the focal length reduced. This puts the 16MP of my GH4 squarely in the centre of the lens at the same pixel density as a 64MP FF camera. Hence my 70-200 f2.8 becomes a 140-400 equiv that I'd expect to be very sharp and the 300 is even more fun (but still big). According to etendue (e.g. http://www.clarkvision.com/articles/telephoto.system.performance/index.html ) it might be pretty reasonable as the f2.8 lenses are good wide open.
OK. Sorry about the misunderstanding. My fault.

But without a focal reducer, the resolution prospects gets worse. Using just the center of the image circle from a lens made for a bigger format is generally inferior to using the whole image circle from a lens made for the smaller format. It follows that your FF 70-200/2.8 when used on MFT is highly unlikely to come close to the resolution provided by a high-quality MFT tele zoom like the new 40-150/2.8.
It's an interesting question and one I don't know the answer to. Canon have said they've been making lenses for a while now designed for high resolution sensors so it might be an interesting test to do (reports of 50MP cameras under test circulate from time-to-time, although Canon make a lot more prototypes than they release cameras). Also it's using the centre part of the optics, which will help somewhat. I'd just prefer it if someone else spent the money to try it out, as the adapter isn't cheap. Also remember the Canon lenses I mentioned are stabilised, which is good for video unlike the Oly lens (I have a GH4 so no IBIS - although a tripod works too). Having the stabilised 35-100 I wouldn't get the 50-140 as it doesn't offer enough more for me and it's larger too. The adapter would give me the option of a 600mm equiv with an f2.8 aperture for exposure for not a lot of money.
I'm wondering whether to get a metabones EF-m43 smart adapter (so a speedbooster without the lens) as a way to hook some of my very sharp Canon lenses on the GH4. The stabiliser would work but it's MF only. Would be interesting. I'm sure the 70-200 II and 300 have sharpness to spare (as they are still sharp with added 1.4x extenders, showing there is extra available, plus this route adds no glass so should be better than that). However the adapters really aren't cheap so I'm not sure. Certainly would be very challenging with some subjects, then again my old film stuff was all MF (ah, but nicer focusing screens though).
A well-made generic focal reducer, as the Metabones seems to be, would of course improve the resolution of an FF lens on an MFT sensor compared to what you can expect without the focal reducer. But it wouldn't reach parity with what you can get with an FF lens on FF since the focal reduction is 0.71 rather than 0.5. At 0.5 it might get close but a generic focal reducer would have to be really perfect to reach full parity.
 
I would say including bracketing is pretty limiting, by restricting yourself to tripods and low-movement scenes, so not generally applicable. It's quite interesting seeing some of the bracketed scenes people shoot on the Thames at night. They think nothing moved so it's good but it isn't as the boats move up and down due to other boats that passed by a while back. That also screws with long exposures trying to make the water super-smooth of course.
Three questions:

1. What makes you think a tripod is necessary for bracketing? I do it hand-held all the time. This is not a problem as long as the longest shutter speed in the bracket can be hand-held without difficulty. The shots used for the stacked image below, for example, were hand-held.



2. What makes you think subject movement is a problem? You can tell the software to use a specific frame in the sequence to render a certain object or area. Alternatively, you can ask the software to make unwanted moving objects disappear.

3. Remember that the starting point you, not I, chose here is the one where shutter speed is not a restriction so that FF can get the same DoF as MFT without leaving base ISO by using four times as long an exposure time. Exactly what is it that we can't accomplish with bracketing and MFT in that scenario? As far as I can see, all the MFT user has to do to reach parity or better is to make sure the longest exposure time in the bracket is as long as that employed by the FF user.
Also your DR graphs would look quite different if you did DR for mid-gray (say) to get an idea of mid-tone image noise.
DR is by definition not about middle gray. It is a measure of SNR in the shadows, where the SNR is always weakest. This is therefore where most people find noise most troubling and therefore effectively sets the floor for acceptable image quality. Have a look at this report from a poll where people were asked to judge the quality of two images with regard to noise, one of which one EV better for midtone SNR, the other one EV better for DR. The large majority found the image with one EV better DR preferable as far as noise is concerned.
I'm talking about mid-tone noise and so DR was a poor choice of word, should have said SNR.
(They don't as they aren't measuring the noise in the light, although photographers usually have to include that, including me as my house no longer has any coal storage facilities.)
Not sure I understand what you have in mind here.
Just saying, with an attempt at some humour, you can't ignore the noise in the light.
 
You started arguing with me so I reserve the right to state my points rather than chase yours. Which also means I'm not in a Straw Man position as that was where I started but you are.
 
I'm still here as just had a long phone call but must get to the shops pre rush-hour. I'm not sure why you feel the need for the rude tone but I'm not trying for anything like that so please don't think I'm arguing rather than discussing.

As I said somewhere above you'll get about 1.75 more stops of base ISO DR from any of the current FF Nikons, which is shadow noise.
Certainly not from the D4S and Df where the difference in comparison with the E-M1 is down to nearly nothing. The FF Nikons with higher pixel counts do well at base ISO but instead have their efficiency deficit at higher ISOs. Regardless of which FF sensor you pick, you will find that it does not realize the 2 EV advantage at the same exposure that equivalence leads us to expect across the entire ISO range. Some come close to that at base ISO and others at higher ISO but none comes close to it at low and high ISO alike.
You are correct on the base ISO difference of sensors designed to work at very high ISO and it's a fair point (although you picked two cameras with the same sensor so I'd say it's a single exception for Nikon - the Sony A7s is the same for low ISO DR). However that sensor does do a lot better than the m43 sensors at high ISO, so swings and roundabouts. I'd been looking at the more general purpose D610, D750 and D810 where 1.7 stops is about correct.
You'll also get circa 2 stops better mid-tone noise due to extra light gathering. Old FF cameras won't do as well for shadow noise but will do much better for for mid-tone noise unless someone altered Physics while I wasn't looking (are you suggesting the much larger number of electrons captured will in some way have more than their square root of noise?).
 
I wouldn't disagree with any of your comments (which I don't see as disagreeing with me), although it would be interesting for someone to measure lp/mm for FF lenses converted onto m43.
You mean "converted" in the sense of adapted with a focal reducer or what?
If you re-read the first sentence of my next paragraph you'll see I'm talking about an adapter, rather than a focal reducer (specifically this: http://www.metabones.com/products/details/MB_EF-m43-BM1). I was looking to get a high quality Telephoto option so don't want the focal length reduced. This puts the 16MP of my GH4 squarely in the centre of the lens at the same pixel density as a 64MP FF camera. Hence my 70-200 f2.8 becomes a 140-400 equiv that I'd expect to be very sharp and the 300 is even more fun (but still big). According to etendue (e.g. http://www.clarkvision.com/articles/telephoto.system.performance/index.html ) it might be pretty reasonable as the f2.8 lenses are good wide open.
OK. Sorry about the misunderstanding. My fault.

But without a focal reducer, the resolution prospects gets worse. Using just the center of the image circle from a lens made for a bigger format is generally inferior to using the whole image circle from a lens made for the smaller format. It follows that your FF 70-200/2.8 when used on MFT is highly unlikely to come close to the resolution provided by a high-quality MFT tele zoom like the new 40-150/2.8.
It's an interesting question and one I don't know the answer to. Canon have said they've been making lenses for a while now designed for high resolution sensors so it might be an interesting test to do (reports of 50MP cameras under test circulate from time-to-time, although Canon make a lot more prototypes than they release cameras). Also it's using the centre part of the optics, which will help somewhat. I'd just prefer it if someone else spent the money to try it out, as the adapter isn't cheap. Also remember the Canon lenses I mentioned are stabilised, which is good for video unlike the Oly lens (I have a GH4 so no IBIS - although a tripod works too). Having the stabilised 35-100 I wouldn't get the 50-140 as it doesn't offer enough more for me and it's larger too. The adapter would give me the option of a 600mm equiv with an f2.8 aperture for exposure for not a lot of money.
I am certainly not saying that what you are suggesting isn't worth trying or that you wouldn't be satisfied with it. It's just that the laws of optics makes it far easier for a lens made for a small image circle to reach a high lp/mm than for a lens made for a larger image circle. When both lenses are used on the sensor sizes for which they are made, that won't disadvantage the lens with the larger image circle on a per-image basis for reasons we have already discussed. But when you put the lens made for a larger sensor on a smaller one, it's another story.
I'm wondering whether to get a metabones EF-m43 smart adapter (so a speedbooster without the lens) as a way to hook some of my very sharp Canon lenses on the GH4. The stabiliser would work but it's MF only. Would be interesting. I'm sure the 70-200 II and 300 have sharpness to spare (as they are still sharp with added 1.4x extenders, showing there is extra available, plus this route adds no glass so should be better than that). However the adapters really aren't cheap so I'm not sure. Certainly would be very challenging with some subjects, then again my old film stuff was all MF (ah, but nicer focusing screens though).
A well-made generic focal reducer, as the Metabones seems to be, would of course improve the resolution of an FF lens on an MFT sensor compared to what you can expect without the focal reducer. But it wouldn't reach parity with what you can get with an FF lens on FF since the focal reduction is 0.71 rather than 0.5. At 0.5 it might get close but a generic focal reducer would have to be really perfect to reach full parity.
 
I'm still here as just had a long phone call but must get to the shops pre rush-hour. I'm not sure why you feel the need for the rude tone but I'm not trying for anything like that so please don't think I'm arguing rather than discussing.

As I said somewhere above you'll get about 1.75 more stops of base ISO DR from any of the current FF Nikons, which is shadow noise.
Certainly not from the D4S and Df where the difference in comparison with the E-M1 is down to nearly nothing. The FF Nikons with higher pixel counts do well at base ISO but instead have their efficiency deficit at higher ISOs. Regardless of which FF sensor you pick, you will find that it does not realize the 2 EV advantage at the same exposure that equivalence leads us to expect across the entire ISO range. Some come close to that at base ISO and others at higher ISO but none comes close to it at low and high ISO alike.
You are correct on the base ISO difference of sensors designed to work at very high ISO and it's a fair point (although you picked two cameras with the same sensor so I'd say it's a single exception for Nikon - the Sony A7s is the same for low ISO DR). However that sensor does do a lot better than the m43 sensors at high ISO, so swings and roundabouts. I'd been looking at the more general purpose D610, D750 and D810 where 1.7 stops is about correct.
Note the generalization contained at the end of the paragraph of mine you comment on. That generalization holds regardless of which FF sensor you choose to compare with. There is always an FF efficiency deficit such that the difference is far less than 2 EV at either end of the ISO scale. Sometimes it is at low ISO (e.g., D4S, Df, A7S, and all Canons), sometimes at high (as for the Nikon D610, D750, and D810). Sometimes, the difference is even negative. All Canon FF sensors do worse than MFT for base ISO DR.
You'll also get circa 2 stops better mid-tone noise due to extra light gathering. Old FF cameras won't do as well for shadow noise but will do much better for for mid-tone noise unless someone altered Physics while I wasn't looking (are you suggesting the much larger number of electrons captured will in some way have more than their square root of noise?).
 

Keyboard shortcuts

Back
Top