DxOmark -- why "only" 71 points?

[losangeles]

Why are the m43 sensors 20+ points behind the best sensors like the D600? Is it because the pixels on a FF sensor are larger and can be manufactured to have better dynamic range? Or does Nikon simply make better sensors than Sony/Olympus?

 

[Mssimo]

http://www.dxomark.com/index.php/About/Sensor-scores

You can read all about it here but it comes down to dynamic range and noise at high ISOs.

Full frame sensors get 4x the amount of light a any given F-Stop.

DXO is kind of weird, seems like medium format cameras do not do as well as full frame.
anyone can help me out with that one?
--
Mássimo

 

[Anders W]

Why are the m43 sensors 20+ points behind the best sensors like the D600? Is it because the pixels on a FF sensor are larger and can be manufactured to have better dynamic range? Or does Nikon simply make better sensors than Sony/Olympus?

The primary reason is that an FF sensor has about four times the area of an MFT sensor. It therefore accumulates more light (more photons) at the same exposure (same f-stop and shutter speed). The pixel size is of secondary importance. Even if we were to compare a 64 MP FF sensor with a 16 MP MFT sensor so that the size of each pixel would be the same on the FF sensor as on the MFT sensor, the FF sensor would be expected to do significantly better in DxO's tests.

 

[Anders W]

http://www.dxomark.com/index.php/About/Sensor-scores

You can read all about it here but it comes down to dynamic range and noise at high ISOs.

Full frame sensors get 4x the amount of light a any given F-Stop.

DXO is kind of weird, seems like medium format cameras do not do as well as full frame.
anyone can help me out with that one?

The reason why MF sensor don't do as well is that they are lagging behind in terms of technology. If you could put four D800 sensors together into one MF sensor, the MF sensor would beat the D800 sensor for exactly the same reason you point out earlier in your post: greater total amount of light on the sensor at any given exposure (f-stop and shutter speed).

 

[napilopez]

What Anders said.

Consider also that actually all the brands you mentioned currently have at least some of their recent sensors made by(although probably tweaked by the individual brands). The very best DxO rated sensor is made by Sony for sure, the D800(E).

Finally, kep in mind DxO scores don't tell the full story. They are mostly useful to compare to eachother from one reference point, and that the importance of the categorical scores may matter differently for different uses.

For example, DxO rates the GH1 as being better than the GH2. Most practical usage, however, will show the contrary. This is most evident in the ISO score, where the GH1 beats the GH2 according to DxO, but by virtually all usage accounts, the GH2 outperforms the GH1. The OM-D scores lower than the NEX-7 for low light, but even accounting for underexposure on the OM-D, it seems to slightly but noticeably outperform the 7 in all of my trials.

Then there are other things not factored. For instance the OM-D may be worse than a 5N in low light, but having very effective IBIS means that with static subject, your images will often come out cleaner than those of a 5N with non-stabilized lenses.

DxO is a useful tool, but not the full story, as they acknowledge themselves.

 

[seachicken2000]

Why are the m43 sensors 20+ points behind the best sensors like the D600? Is it because the pixels on a FF sensor are larger and can be manufactured to have better dynamic range? Or does Nikon simply make better sensors than Sony/Olympus?

The primary reason is that an FF sensor has about four times the area of an MFT sensor. It therefore accumulates more light (more photons) at the same exposure (same f-stop and shutter speed).

Personally, I would prefer to compare at the same photometric exposure; in other words the case where both sensors are recording the same image. I'm in the camp that believes a 25mm f/1.4 lens on m4/3 is effectively the same as a 50mm f/2.8 lens on full-frame. When using these lenses at their equivalent f-stops, the m4/3 lens will be creating a 2-stop brighter image on the sensor, and therefore the ISO will need to be lowered 2 stops to correctly record the same image.

So in this case you'd need to shift the E-M5 curves to the right about 2 stops when comparing to full frame, and it's a fair fight. Comparing at the same ISO is IMO giving the E-M5 an unfair handicap.

--
"The only thing that gets in the way of a really good photograph, is the camera"
Norman Parkinson

 

[Mr.NoFlash]

What Anders said.

Consider also that actually all the brands you mentioned currently have at least some of their recent sensors made by(although probably tweaked by the individual brands). The very best DxO rated sensor is made by Sony for sure, the D800(E).

Finally, kep in mind DxO scores don't tell the full story.

Especially, do not forget the lenses.

For the same price, m43 lenses can be of much better quality ( sharpness ) than Fullframe Lenses.

Perhaps thats not always the case, but thats just because currently m43 has only one big AND serious competitor ( Sony ) in the mirrorless field;
with more competition, the better lens quality/price ratio of m43 will be there.

 

[losangeles]

I disagree with this notion that m43 lenses are less expensive. I can get great used fast prime lenses covering most of my usage scenarios (say a wide angle, normal, and tele) for FF for $1,000. I could use the same lenses for m43 with an adapter but that would defeat the purpose (lower weight). The alternative would be to buy m43 primes, but that would set me back $2,500-$3,000.

I see it the opposite as you: m43 is the format for wealthy people.

Especially, do not forget the lenses.

For the same price, m43 lenses can be of much better quality ( sharpness ) than Fullframe Lenses.

Perhaps thats not always the case, but thats just because currently m43 has only one big AND serious competitor ( Sony ) in the mirrorless field;
with more competition, the better lens quality/price ratio of m43 will be there.

 

[MAubrey]

I can get great used fast prime lenses covering most of my usage scenarios (say a wide angle, normal, and tele) for FF for $1,000. I could use the same lenses for m43 with an adapter but that would defeat the purpose (lower weight). The alternative would be to buy m43 primes, but that would set me back $2,500-$3,000.

...which up until this past month was how much you'd pay for a FF body.

I surely didn't spend that much on my lenses. My trinity was exactly $1000 total (14mm, 25mm, & 45mm). If you're spending $2500 on a wide, normal, and tele, then you're doing something very, very wrong.

I see it the opposite as you: m43 is the format for wealthy people.

Only if you intentionally buy every lens a few hundred dollars above their MSRP.
--
--Mike

 

[Anders W]

Why are the m43 sensors 20+ points behind the best sensors like the D600? Is it because the pixels on a FF sensor are larger and can be manufactured to have better dynamic range? Or does Nikon simply make better sensors than Sony/Olympus?

The primary reason is that an FF sensor has about four times the area of an MFT sensor. It therefore accumulates more light (more photons) at the same exposure (same f-stop and shutter speed).

Personally, I would prefer to compare at the same photometric exposure; in other words the case where both sensors are recording the same image.

I agree that that's an interesting comparisons, though not the only interesting one.

However, my ambition here was simply to help the OP understand why the best FF sensors are 20+ points ahead of the best MFT sensor in terms of DxO ratings, and for this purpose, I think the answer I gave is correct and to the point. I didn't say that the 20+ difference is what everyone should use as the primary basis for choosing a camera, let alone the only basis.

I'm in the camp that believes a 25mm f/1.4 lens on m4/3 is effectively the same as a 50mm f/2.8 lens on full-frame.

So am I, as far as DoF, diffraction, and light accumulation (but not exposure) is concerned.

When using these lenses at their equivalent f-stops, the m4/3 lens will be creating a 2-stop brighter image on the sensor, and therefore the ISO will need to be lowered 2 stops to correctly record the same image.

Agreed.

So in this case you'd need to shift the E-M5 curves to the right about 2 stops when comparing to full frame, and it's a fair fight. Comparing at the same ISO is IMO giving the E-M5 an unfair handicap.

When shooting at higher ISOs and DoF is in short supply, I would agree. In this case, the MFT camera might even beat an FF camera as I showed here:

http://forums.dpreview.com/forums/read.asp?forum=1041&message=42531043

In other cases, such as when light is abundant (and there is no lower ISO or greater aperture for the MFT camera to go to) and when DoF is not in short supply or even undesirable, it's another story.

 

[DP13Photo]

My view is that µ4/3 lenses are across the board very sharp wide open, whereas all of my Canon primes are mostly soft wide open, at an often higher price. I have no idea what you are talking about. Full frame bodies and lenses are for the very rich and the very strong..... or the press.

Case in point. My Canon 85 f1.8 $369 (portrait lens) had to be replaced by Canon 85 1.2L. at 5 times the cost $2000.

Whereas the Olympus 45 1.8 at comparable price to the f1.8 Canon is sharp wide open and performs almost like the expensive L lens.

And lets not even talk about calibrating lenses or the difference between CDAF and PDAF where the number of in focus "keepers" is heavily in favor of µ4/3. Talk about cost.... get a grip on reality!

I disagree with this notion that m43 lenses are less expensive. I can get great used fast prime lenses covering most of my usage scenarios (say a wide angle, normal, and tele) for FF for $1,000. I could use the same lenses for m43 with an adapter but that would defeat the purpose (lower weight). The alternative would be to buy m43 primes, but that would set me back $2,500-$3,000.

I see it the opposite as you: m43 is the format for wealthy people.

Especially, do not forget the lenses.

For the same price, m43 lenses can be of much better quality ( sharpness ) than Fullframe Lenses.

Perhaps thats not always the case, but thats just because currently m43 has only one big AND serious competitor ( Sony ) in the mirrorless field;
with more competition, the better lens quality/price ratio of m43 will be there.

--
Dave

 

[Mike Fewster]

Why are the m43 sensors 20+ points behind the best sensors like the D600? Is it because the pixels on a FF sensor are larger and can be manufactured to have better dynamic range? Or does Nikon simply make better sensors than Sony/Olympus?

Nikon use Sony sensors in their top models. Some of the difference in performance between the same sensor can be down to model variation. Some can be because of the algorithm tweaking done by all manufacturers. Oly for example have excellent jpeg colour. Even in RAW, different manufacturers apply different amounts of sharpening. But there is always a cost, there is no free lunch. Improving one parameter is usually at the expense of another. For example, when Nikon began using the Sony sensor, they seemed to get a little more out of it with cleaner shots but more investigation at the time showed that the Sony cameras retained detail marginally better.
--
Mike Fewster
Adelaide Australia

 

[rhlpetrus]

2stop high ISO and 2 EV DR. It's 4x larger area at work, nothing surprising and nothing related to pixel size, D600=D800, basically.
--
Renato.
http://www.flickr.com/photos/rhlpedrosa/
OnExposure member
http://www.onexposure.net/

Good shooting and good luck
(after Ed Murrow)

 

[Anders W]

2stop high ISO and 2 EV DR. It's 4x larger area at work, nothing surprising and nothing related to pixel size, D600=D800, basically.

Not quite that simple I am afraid. It's two stops between the E-M5 and the D800 at base ISO yes (and hardly anything between the D7000 and the D800 in spite of the fact that the D7000 has less than half the sensor area). Further, the difference in DR at higher ISOs (which is a more informative measure than the "sports" score in my opinion) between the E-M5 and the D800 is only about one EV, not two. This is borne out by visual evidence too. Have a look here:

http://forums.dpreview.com/forums/read.asp?forum=1041&message=42531043

Finally, there is virtually no difference at all between the E-M5 and the best APS-C sensors when it comes to high-ISO DR, in spite of the fact that the MFT sensor area is significantly smaller.

 

[Great Bustard]

2stop high ISO and 2 EV DR. It's 4x larger area at work, nothing surprising and nothing related to pixel size, D600=D800, basically.

Not quite that simple I am afraid. It's two stops between the E-M5 and the D800 at base ISO yes (and hardly anything between the D7000 and the D800 in spite of the fact that the D7000 has less than half the sensor area). Further, the difference in DR at higher ISOs (which is a more informative measure than the "sports" score in my opinion) between the E-M5 and the D800 is only about one EV, not two. This is borne out by visual evidence too. Have a look here:

http://forums.dpreview.com/forums/read.asp?forum=1041&message=42531043

Finally, there is virtually no difference at all between the E-M5 and the best APS-C sensors when it comes to high-ISO DR, in spite of the fact that the MFT sensor area is significantly smaller.

All true. However, as you know, two systems with the same DR can have rather different visual qualities. For example, at base ISO the G12 (compact) has more DR than the 5D2 (FF), and I don't think you'd argue that the G12 has "higher IQ" than the 5D2.

 

[Anders W]

2stop high ISO and 2 EV DR. It's 4x larger area at work, nothing surprising and nothing related to pixel size, D600=D800, basically.

Not quite that simple I am afraid. It's two stops between the E-M5 and the D800 at base ISO yes (and hardly anything between the D7000 and the D800 in spite of the fact that the D7000 has less than half the sensor area). Further, the difference in DR at higher ISOs (which is a more informative measure than the "sports" score in my opinion) between the E-M5 and the D800 is only about one EV, not two. This is borne out by visual evidence too. Have a look here:

http://forums.dpreview.com/forums/read.asp?forum=1041&message=42531043

Finally, there is virtually no difference at all between the E-M5 and the best APS-C sensors when it comes to high-ISO DR, in spite of the fact that the MFT sensor area is significantly smaller.

All true. However, as you know, two systems with the same DR can have rather different visual qualities. For example, at base ISO the G12 (compact) has more DR than the 5D2 (FF), and I don't think you'd argue that the G12 has "higher IQ" than the 5D2.

I haven't looked all that carefully at G12 images and know next to nothing about its technical or other performance. Not all that interested in compacts. But I figure there must be a point where midtone noise becomes troubling and I am sure the 5D2 has less of that than the G12.

On the other hand, I am sure the E-M5 stands up very well compared to the 5D2 (and 5D3) for overall perceived image quality at base ISO. At higher ISOs, the FF cameras will be about one EV ahead when DoF is not in short supply and about one EV behind when it is.

 

[Danel]

Look at it this way, according to DXO, the Omd now performs better than the Canon APS-C cameras. If you had told me that a couple years ago I would have had a hard time believing it. Sony is so far ahead of Canon's sensors that even the 1" sensor on the RX100 is tied with Canon's APS-C cameras like the 7D and 550D. Olympus made a smart move teaming up with Sony for sensors. Suddenly, the Olympus micro four thirds cameras are at parity with many APS-C cameras. The Olympus engineers must be walking around with smiles on their faces after persevering for so long with the old Panasonic 12mp sensors.

 

[Polariser]

My view is that µ4/3 lenses are across the board very sharp wide open, whereas all of my Canon primes are mostly soft wide open, at an often higher price. I have no idea what you are talking about. Full frame bodies and lenses are for the very rich and the very strong..... or the press.

Case in point. My Canon 85 f1.8 $369 (portrait lens) had to be replaced by Canon 85 1.2L. at 5 times the cost $2000.

Whereas the Olympus 45 1.8 at comparable price to the f1.8 Canon is sharp wide open and performs almost like the expensive L lens.

And lets not even talk about calibrating lenses or the difference between CDAF and PDAF where the number of in focus "keepers" is heavily in favor of µ4/3. Talk about cost.... get a grip on reality!

Thank you Great input from a real user. The "calibrating lenses" bit has kept me away from DSLR's.

The 45 is a great lens for a small price.

 

[Bart Hickman]

Personally, I would prefer to compare at the same photometric exposure; in other words the case where both sensors are recording the same image. I'm in the camp that believes a 25mm f/1.4 lens on m4/3 is effectively the same as a 50mm f/2.8 lens on full-frame. When using these lenses at their equivalent f-stops, the m4/3 lens will be creating a 2-stop brighter image on the sensor, and therefore the ISO will need to be lowered 2 stops to correctly record the same image.

The way they compare it now, the FF sensor gets 4x the light because it is able to handle 4x the light and lenses that provide 4x the light are practical and economical to make. You have to give the larger sensor the full benefit of its size advantage.

Alternatively, you would change the FF sensor to F/2.8 (as you suggest), but leave the ISO setting the same. This underexposes the image, but only in a software sense. In a hardware sense, the "underexposed" FF sensor gathers the same amount of light and so sampled the same image as the "properly" exposed m43 sensor. I can boost the image in PP (or in camera) to display the image properly. In this case, the FF sensor and m43 sensor have the same total noise, but the FF sensor had 4x the dynamic range at the highlight end.

Either way, the FF sensor has 4x the DR and DxO would get the same relative answer.

You also see some latent advantages of the larger sensor appear here as well--eg., 50/2.8 can be had in a zoom lens on FF whereas F/1.4 zoom is out of the question on m43. Or you can have 50/2.8 in a macro lens on FF. The FF 28-300 F/3.5-5.6 converts to 14-150 F/1.8-2.8. I don't even know if such a lens is economically feasible. So your fair comparison starts to look even worse for the smaller sensor from a holistic system standpoint.

OTOH, the Pany 14-140 F/4-5.8 becomes 28-280 F/8-11.6 on FF which would be this long skinning sausage lens that's probably also not practical. Each system has their advantage. FF makes high performance practical. m43 enables a nice small system for performance hit. CX pushes it even farther towards small/lower performance. APS-C backs things off a bit towards more performance. So much choice!

Bart
--
http://bhimages.zenfolio.com

 

[Chez Wimpy]

On the other hand, I am sure the E-M5 stands up very well compared to the 5D2 (and 5D3) for overall perceived image quality at base ISO.

Sounds good to me. The 5D2 is a fickle basterd when banding rears its head at low ISO (and in landscapes this is exacerbated if you plan to do any PP at all), so I can pick up an OMD and landscape with my 14-45 without caveats. No tilt and shift, and no equivalent to the sunset spectacular Sigma 12-24 (the 7-14 is more expensive, but not as wide), but for the most part, I am sure the walkaround results will hold up, which means my 550D is moving to semi-permanent EF "digital teleconverter" duty. Heck, I am even impressed with the GH2 I picked up last week. The DR is still

--
-CW