Evolution of m43 Image Quality

Define hardware performance?
to compare a normal raw shooter will use the same illumination (did dpreview finally start using the same illumination to the dot ?), same exposure time and aperture and lenses that have equal transmission (T-stops) @ that aperture, stopped down (to address vignetting), and the best gain (not equal nominal ISOs dialed on camera) for each camera and raw converter (and raw converter shall not be like ACR/LR that are known to use hidden corrections, etc for different cameras).. then you need either to downsize an image from a camera w/ more mp up upsize an image from a camera w/ less mp... then you can draw some conclusions... dpreview examples miss so many points here, that they are valid only for jpeg shooters... on top of that you shall remember that beyond DR (which is about readout noise mostly) there is S/N above deep shadows (when readout noise is overcome by photon/shot noise and things are mostly about sheer light collecting surface - there is no replacement for displacement) and quality of color separation (see something like SMI @ DxOMark, albeit the single number for a given spectrum is a very crude indicator), and so on...
 
Well, I was very happy with the quality of the e-m5, especially with the 45 1.8 and the 75mm. But then I looked at the A7r....

I guess you really can't have it all.
 
Also, does the E-m1 today beat the IQ of the original Canon 5D?
Yes. I have a Canon 5D sitting beside my E-M5, which is sitting beside my Sony A7. The Olympus has certainly surpassed the Canon. The Canon can have slightly cleaner files, particularly at higher ISO's. The problem is that there's so little latitude for editing the RAW files without introducing banding. The E-M5 files are vastly superior to the 5D's in terms of their flexibility and in LR that invariably results in far superior IQ for the final printed result.
 
I'm not a troll. I've posted many message before. I own 2 e-m5's, and more m43 lenses than I can fit into a (big) bag. I own no dSLR's, and my Canons and their lenses and my other similar bodybuilding equipment long ago left my house.

I am very happy with the 4-m5, and waiting for the next iteration after the E-M1, or maybe the one after that.

Still, the other day I wandered past B&H's Sony booth, and tried out the A7R, more for curiosity than anything else. The IQ was, um, really good.

I know the routine - few lenses, and when they come, they will be big, especially, zoom telephoto lenses. I left that behind me, the 70-200, with 1.4x extender, when my Canon's left me.

My question: When in the future will m43 have the IQ of today's A7r?
For my usage, the future is already here. As illustrated by the link below, the best MFT sensor beats that of the A7R by about 1.5 EV for equivalent photos (same DoF, same shutter speed).

http://www.dpreview.com/reviews/ima...&x=-0.11008016822184253&y=-1.0330123796423658

Hard to say how long it will take for FF mirrorless to catch up. ;-)
Also, does the E-m1 today beat the IQ of the original Canon 5D?
Do you know how to process raw?

http://www.dpreview.com/reviews/ima...&x=-0.11008016822184253&y=-1.0330123796423658
Yes I do. But judging by your link, you have no clue about how to evaluate hardware performance.
Define hardware performance?
Are you saying that you don't understand what that is? If not, why are you asking?
It is easy to apply NR to the RAW, in camera, but it won't improve the JPGs. My link clearly shows that +2EV is about the right noise-level (same), note - the A7r image has not even been down-sampled yet - another way to extract/preserve detail and reduce the noise further.

Your link does not hint of any of that.
Why should it. It shows what it was meant to show: Hardware performance. You are showing something I am not interested in at all (the OOC jpegs).
If it's the hardware performance we are interested in, it's better to look at this:

http://www.dpreview.com/reviews/ima...&x=-0.11008016822184253&y=-1.0330123796423658

But yes, they are pretty similar at the same ISO. Therefore, MFT has the 1.5 EV advantage that I mentioned.
Looks to me that they are rather similar at +2EV, won't you agree (even before resizing the FF one)?

http://www.dpreview.com/reviews/ima...=1&x=-0.3636745958733079&y=-0.868893692277461
The hardware performance is clearly very different for equivalent photos, just as I originally pointed out. Before downsampling the FF image, MFT is about 2 EV ahead. After downsampling to 16 MP, it's about 1.5 EV.
If your point is about hardware efficiency - yes, I agree, we had those discussions, and smaller sensors are more efficient. But to my point, a larger sensor still performs better. Maybe not the full +2EV, but still rather close to it, and it shows in the results.
Well, your point is simply wrong since the evidence clearly shows the contrary.
Anders, are you really trying to tell us that FF and m43 sensors perform the same way? Image IQ wise?

I can go out there and shoot side by side with similar results?

I am sorry, but I do trade DOF for noise on FF. So I am not shooting equivalent at all. Rather, I am getting very usable high-ISO images out of larger sensor cameras.

And I bet every FF shooter does the same thing, with a lens that is as fast (+2EV) and at at least one stop higher ISO (+1EV). (assume stabilized lenses).

We hit on some of this in Daniel's post: http://www.dpreview.com/forums/post/53007968

and I like his conclusion: http://www.dpreview.com/forums/post/53099150

So, back to your and my findings: even IF the sensors perform similar, the image (IQ) is dissimilar - FF shooting is simply different from m43 shooting.

Can we leave it at that?
 
Corpy2,

I do not care whether you troll or not,

My problem with you is total absence of evidence that you really have something do with photography. How can you prove that you have a camera (any brand) and do photography?

Hope you can prove that you're relevant and worth a discussion

Regards

S.

Camera in bag tends to stay in bag...
Why has he/she to prove anything to you to be "worth a discussion"?
And U 2
 
I'm not a troll. I've posted many message before. I own 2 e-m5's, and more m43 lenses than I can fit into a (big) bag. I own no dSLR's, and my Canons and their lenses and my other similar bodybuilding equipment long ago left my house.

I am very happy with the 4-m5, and waiting for the next iteration after the E-M1, or maybe the one after that.

Still, the other day I wandered past B&H's Sony booth, and tried out the A7R, more for curiosity than anything else. The IQ was, um, really good.

I know the routine - few lenses, and when they come, they will be big, especially, zoom telephoto lenses. I left that behind me, the 70-200, with 1.4x extender, when my Canon's left me.

My question: When in the future will m43 have the IQ of today's A7r?

Also, does the E-m1 today beat the IQ of the original Canon 5D?
Corpy2,

I do not care whether you troll or not,

My problem with you is total absence of evidence that you really have something do with photography. How can you prove that you have a camera (any brand) and do photography?

Hope you can prove that you're relevant and worth a discussion

Regards

S.

Camera in bag tends to stay in bag...
Oh, my. You've certainly progressed the thread to the wand waving stage quickly, have you not.

Strange as it may seem, I actually do not live to prove anything to you. I'm not interested in the length of your wand.

But since you are so interested in mine, I'll accommodate you, since I wouldn't want to be the cause of your jumping on the guy next to you at the urinal. Before I got my first 35mm camera in 1971, I had a Polaroid, with 2 f stops, one for colour, and one for black and white, and I discovered that if one used the one for color with B&W film, one could take some remarkably low-light pictures in dim lighting.

Based on your message to date, I'll have to assume that was around the time your grandmother first met your grandfather in 6th grade.
 
I'm not a troll. I've posted many message before. I own 2 e-m5's, and more m43 lenses than I can fit into a (big) bag. I own no dSLR's, and my Canons and their lenses and my other similar bodybuilding equipment long ago left my house.

I am very happy with the 4-m5, and waiting for the next iteration after the E-M1, or maybe the one after that.

Still, the other day I wandered past B&H's Sony booth, and tried out the A7R, more for curiosity than anything else. The IQ was, um, really good.

I know the routine - few lenses, and when they come, they will be big, especially, zoom telephoto lenses. I left that behind me, the 70-200, with 1.4x extender, when my Canon's left me.

My question: When in the future will m43 have the IQ of today's A7r?
For my usage, the future is already here. As illustrated by the link below, the best MFT sensor beats that of the A7R by about 1.5 EV for equivalent photos (same DoF, same shutter speed).

http://www.dpreview.com/reviews/ima...&x=-0.11008016822184253&y=-1.0330123796423658

Hard to say how long it will take for FF mirrorless to catch up. ;-)
Why did you specify 3200 ISO for the em1 and 12800 for the A7r?
 
I'm not a troll. I've posted many message before. I own 2 e-m5's, and more m43 lenses than I can fit into a (big) bag. I own no dSLR's, and my Canons and their lenses and my other similar bodybuilding equipment long ago left my house.

I am very happy with the 4-m5, and waiting for the next iteration after the E-M1, or maybe the one after that.

Still, the other day I wandered past B&H's Sony booth, and tried out the A7R, more for curiosity than anything else. The IQ was, um, really good.

I know the routine - few lenses, and when they come, they will be big, especially, zoom telephoto lenses. I left that behind me, the 70-200, with 1.4x extender, when my Canon's left me.

My question: When in the future will m43 have the IQ of today's A7r?
For my usage, the future is already here. As illustrated by the link below, the best MFT sensor beats that of the A7R by about 1.5 EV for equivalent photos (same DoF, same shutter speed).

http://www.dpreview.com/reviews/ima...&x=-0.11008016822184253&y=-1.0330123796423658

Hard to say how long it will take for FF mirrorless to catch up. ;-)
Also, does the E-m1 today beat the IQ of the original Canon 5D?
Do you know how to process raw?

http://www.dpreview.com/reviews/ima...&x=-0.11008016822184253&y=-1.0330123796423658
Yes I do. But judging by your link, you have no clue about how to evaluate hardware performance.
Define hardware performance?
Are you saying that you don't understand what that is? If not, why are you asking?
It is easy to apply NR to the RAW, in camera, but it won't improve the JPGs. My link clearly shows that +2EV is about the right noise-level (same), note - the A7r image has not even been down-sampled yet - another way to extract/preserve detail and reduce the noise further.

Your link does not hint of any of that.
Why should it. It shows what it was meant to show: Hardware performance. You are showing something I am not interested in at all (the OOC jpegs).
If it's the hardware performance we are interested in, it's better to look at this:

http://www.dpreview.com/reviews/ima...&x=-0.11008016822184253&y=-1.0330123796423658

But yes, they are pretty similar at the same ISO. Therefore, MFT has the 1.5 EV advantage that I mentioned.
Looks to me that they are rather similar at +2EV, won't you agree (even before resizing the FF one)?

http://www.dpreview.com/reviews/ima...=1&x=-0.3636745958733079&y=-0.868893692277461
The hardware performance is clearly very different for equivalent photos, just as I originally pointed out. Before downsampling the FF image, MFT is about 2 EV ahead. After downsampling to 16 MP, it's about 1.5 EV.
If your point is about hardware efficiency - yes, I agree, we had those discussions, and smaller sensors are more efficient. But to my point, a larger sensor still performs better. Maybe not the full +2EV, but still rather close to it, and it shows in the results.
Well, your point is simply wrong since the evidence clearly shows the contrary.
Anders, are you really trying to tell us that FF and m43 sensors perform the same way? Image IQ wise?
Obviously not. I am telling (and showing) you that MFT sensors perform better within the DoF range I, and quite a few others it seems, are ordinarily interested in using.
I can go out there and shoot side by side with similar results?

I am sorry, but I do trade DOF for noise on FF. So I am not shooting equivalent at all. Rather, I am getting very usable high-ISO images out of larger sensor cameras.

And I bet every FF shooter does the same thing, with a lens that is as fast (+2EV) and at at least one stop higher ISO (+1EV). (assume stabilized lenses).

We hit on some of this in Daniel's post: http://www.dpreview.com/forums/post/53007968
Yes. That's why I am surprised that you are still taking issue with obvious facts.
Well, you need to learn to expose so as not to clip any channel and use suitable RAW-converter settings. That's true for every camera and has no bearing on the hardware-performance issue we are discussing here.
So, back to your and my findings: even IF the sensors perform similar, the image (IQ) is dissimilar - FF shooting is simply different from m43 shooting.
Sure. That's what I have been telling you all the time. If you want extremely shallow DoF, an FF camera has advantages. If you don't it has the disadvantage I mentioned (along with some others in the form of higher bulk, weight, and cost).
Can we leave it at that?
Absolutely.
 
Last edited:
I'm not a troll. I've posted many message before. I own 2 e-m5's, and more m43 lenses than I can fit into a (big) bag. I own no dSLR's, and my Canons and their lenses and my other similar bodybuilding equipment long ago left my house.

I am very happy with the 4-m5, and waiting for the next iteration after the E-M1, or maybe the one after that.

Still, the other day I wandered past B&H's Sony booth, and tried out the A7R, more for curiosity than anything else. The IQ was, um, really good.

I know the routine - few lenses, and when they come, they will be big, especially, zoom telephoto lenses. I left that behind me, the 70-200, with 1.4x extender, when my Canon's left me.

My question: When in the future will m43 have the IQ of today's A7r?
For my usage, the future is already here. As illustrated by the link below, the best MFT sensor beats that of the A7R by about 1.5 EV for equivalent photos (same DoF, same shutter speed).

http://www.dpreview.com/reviews/ima...&x=-0.11008016822184253&y=-1.0330123796423658

Hard to say how long it will take for FF mirrorless to catch up. ;-)
Why did you specify 3200 ISO for the em1 and 12800 for the A7r?
Because in order to produce the same image with the A7r as with an MFT camera, the lens of the A7r (and any other FF camera) must be stopped down two stops more (for equal depth of field). That costs a corresponding increase in ISO on the A7r, everything else equal.

If you are fine with the shallow DoF you will get with FF if you shoot at the same exposure (same f-stop, same shutter speed), then the A7r will instead be about 0.5 EV better than the E-M1 with regard to shadow noise. So whether FF pays off depends on your personal take on the DoF equation. Speaking for myself, I am rarely interested in having less DoF than I can get with my fast MFT primes wide open, but this is just a matter of personal taste and photographic objectives.
 
Last edited:
Your comparing apples to oranges... You can't compare a full frame sensor to Micro Four Thirds. The fact you can get 5D like images out of a Micro Four Thirds camera should be more than good enough for most people.

Anyway... I never understand this ISO nonsense... just get a lens with a decent aperture and ISO performance will be the least of your worries... Most of this debate is trivial nonsense which can resolved by buying a fast lens that is sharp wide open and learning how to shoot properly.
 
Last edited:
Seriously, shoot with a compact for a while and learn how to shoot within stringent ISO limitations and then come back to this discussion. Once you've done that, also come back and compare the prices of a full frame camera vs. a Micro Four Thirds option that is 4x cheaper... Anything else lacks cogency in terms of this discussion.
 
If your point is about hardware efficiency - yes, I agree, we had those discussions, and smaller sensors are more efficient. But to my point, a larger sensor still performs better. Maybe not the full +2EV, but still rather close to it, and it shows in the results.
Well, your point is simply wrong since the evidence clearly shows the contrary.
It really does not. It only shows how a certain portion of an image may or may not look like under certain conditions. Not even for the same size and the amount of detail in it. It completely disregards what the whole image will look like (mid-tones, highlights, gradients), and whether such condition will exist in real shoot to begin with. Again, in most scenarios it simply does not.
 
This comparison is only valid if you have to have an exact depth of field (do people care whether it's 4m or 4.8m?) and an exact shutter speed (again I'm rarely fussed between 1/125 and 1/180 unless stuff is moving or the lens is long). I think you can't use this as the only valid comparison unless you're determined to make m43 look good at all costs. It's one way to compare, but in reality not a choice people would make.

Take a bunch of typical scenes, photograph them as best as the cameras will do and see what the trade-offs are. I suspect the FF bodies will be way ahead. Except you'll enjoy carrying them around a lot less.

P.S. The Canon 6D at 12,800 is a more interesting comparison. I suspect both FF bodies would benefit from DXO9 too.
 
I'm not a troll. ...

Still, the other day I wandered past B&H's Sony booth, and tried out the A7R, more for curiosity than anything else. The IQ was, um, really good....
How could you tell?

Are you saying you took some pictures and right there and then viewed them on a huge monitor or printed them out?

Please don't say you looked at them on the LCD ...
 
This comparison is only valid if you have to have an exact depth of field (do people care whether it's 4m or 4.8m?) and an exact shutter speed (again I'm rarely fussed between 1/125 and 1/180 unless stuff is moving or the lens is long). I think you can't use this as the only valid comparison unless you're determined to make m43 look good at all costs. It's one way to compare, but in reality not a choice people would make.
At some point (e.g. 1/60 and 1/125) people will start to care. The point being made is that the M43 shooter does not need to raise the speed to 1/180. They can stay at 1/125 to match the FF and benefit from the additional exposure.

This ends when the M43 goes out of range, usually with the aperture. Hence Anders keeps saying that if one needs or can tolerate the wider apertures on FF for which there are no M43 equals, his comparison no longer applies.
 
Looks to me that they are rather similar at +2EV, won't you agree (even before resizing the FF one)?


I have to say that the difference is much bigger than I expected, even after I resize:



d022485e4e80416288392683994e5e0b.jpg.png
 
Well, I was very happy with the quality of the e-m5, especially with the 45 1.8 and the 75mm. But then I looked at the A7r....

I guess you really can't have it all.
I have a picture of a glacial moraine taken almost ten years ago with my Olympus E-3 that works well as 17' x 22' show print. That's with a decade+ old 5 MP FT sensor.

Glacial Moraine taken on hike up Field Mountain to the Burgess Shale (Canadian Rocky Mountains)
Glacial Moraine taken on hike up Field Mountain to the Burgess Shale (Canadian Rocky Mountains)

Clyde Butcher hauls huge 8"x10" view cameras through Florida Everglades swamps and enlarges the plates on a 3/4 ton enlarger into beautiful wall prints.

<iframe width="560" height="315" src="//www.youtube.com/embed/LTfqyqjgc44" frameborder="0" allowfullscreen></iframe>

Most of us do want to be able to take along a variety of lenses and maybe even a second body, BUT most of us do not want to haul an "only" 65 pound bag;

And big and determined as he is; hwe still needs a loving, loyal unpaid assistant, Most of us, even if we are blessed with a loving partner, do not have a partner who will - watch the video-:

and most of are willing to make sacrifices for our photography, but not quite dedicated enough to suffer malaria a dozen times.

So we make sensible compromises. The Sony is a very interesting camera. Come to think of it (my wife is a professional musician, and I'm quite familiar with their sound equipment) the sort of triumph of miniaturization that is a hallmark of that company. However, it, too, seems to have its drawbacks. Reviews have been mixed: some have hate the (un)ergonimics. Also, given the optically-necessary-for-full-frame size of the lenses, the net saving in system size and weight. And then there is their cost, which also becomes problematic when related to Sony's relative profusion of different mount and standards (unless one has thousands in pocket change).

For me - and I speculate for many others - the small size and weight of mFT's, as a system in relation to the very acceptable photo quality it produces, as well as the many (affordable!) lens and body options it offers are the perfect choice, much in the same way that 35mm came to be, 50+ years ago.

--
erichK
saskatoon, canada
Photography is a small voice, at best, but sometimes one photograph, or a group of them, can lure our sense of awareness.
- W. Eugene Smith, Dec 30, 1918 to Oct 15, 1978.
http://erichk.zenfolio.com/
http://www.fototime.com/inv/7F3D846BCD301F3
underwater photos:
http://www.scubaboard.com/gallery/showgallery.php/cat/500/ppuser/5567
 
Last edited:
Hi Erich,

I am curious what lens you used for the moraine shot? It looks very sharp edge-edge.
 
Perusal of the DXO Mark website will allow anyone, including Corpy2, prepared to pull out the figures to:

a) hazard a guess at the answer to the first question. I guess 4 years. This is the time it took from the Nikon D2Xs in 2006 (score 59) to the Nikon D800 (probably same sensor as Sony A7r) in 2012 (score 95). Of course by then full frame will deliver scores around 120 or whatever but who cares, there is always something "better".
D2x was APS-C, so it should be comapred to newer APS-C. They have been hitting the 82 mark, about 14 points below the D800e at 96. I doubt that, w/o some trick like multiple sampling, ASP-C or m43 will ever get near D800's performance.
b) answer the second question, (yes), the Olympus EM1 scores 73. The original EOS 5D scored 71 in 2005.
Canon has been much slower than other sensor developers re QI parameters, especially at base ISO (DR and color depth). Still, I doubt the EM1's IQ actually beats the 5D's, in actual performance.
In the meantime those of us who use a recent model M43 camera can enjoy image quality better than that offered by full frame cameras just a few years ago and good enough for just about any purpose.

I don't necessarily think DXO Mark scores are the final judgement of image quality, in fact they are probably not, but they are useful for revealing trends, as here.
The highest scoring m43 is reaching now the level of APS-C of 5-6 years ago (D90). It's ok, pretty good actually, but to say it will keep going up and reach the level of present FF ignores the basic Physics and technological limits of making sensors.
 

Keyboard shortcuts

Back
Top