Valentin Sama from the Spanish Dslrmagazine (Click here for the translation) tested the E-M5 sensor.
He first compared the new E-M5 with the previous m43 Olympus models. In short the noise performance of the E-m5 has indeed been improved but there is no real advantage in terms of revolution over the older 12 Megapixel E-P3 sensor. Valentin says the E-P3 per pixel sharpness is still a bit greater than the one from the 16 Megapixel E-M5 sensor.
Than Valentin made a comparison with two larger sensor APS-C cameras. The Nikon D7000 and Fuji X PRO 1. First he clearly reminds that both cameras are more expensive than the E.M5. The Nikon D7000 costs $1,500 (here at Amazon) and the Fuji X PRo 1 costs $1,700 (here at Amazon). According to Valentin JPEG testing results the E-M5 is on par with the Nikon D7000. And that’s already a surprising result!
The Fuji has a huge advantage because of the lack Anti Aliasing filter. The results are a greater clarity and less noise.
In summary: he says the E-M5 has indeed an improved JPEG quality although we have to see if that is an improvement is more due the new JPEG engine than the new 16 megapixel sensor.
One more thing: E-M5 with 12-50mm kit lens image samples at Robin Wong (Click here).
Rankings: Three different E-M5 kits are still on top of the Amazon US rankings (Click here).
E-M5 preorders (Click on the names of the stores to see the product page):
Black E-M5 body at Amazon, Adorama, BHphoto, J&R, Warehouse UK, Digitalrev and Amazon Japan.
Silver E-M5 body at Amazon, Adorama, BHphoto, J&R, Warehouse UK, and Amazon Japan.
Black E-M5 body with 14-42mm lens at Amazon, Adorama, BHphoto , Digitalrev and J&R.
Black E-M5 body with 12-50mm lens at Amazon, Adorama, BHphoto, J&R, Jessops, Amazon UK, Amazon Deutschland , Digitalrev and Amazon Japan.
Silver E-M5 body with 12-50mm lens at Amazon, Adorama, BHphoto, J&R, Jessops, Amazon UK, Amazon Deutschland and Amazon Japan.
93 Responses
you said “The Nikon D800 costs ” I think you mean D7000 right ?
“The Fuji has a huge advantage because of the lack Anti Aliasing filter. The results are a greater clarity and less noise.”
And more aliasing. And once aliasing has occurred, the only way to remove it is to blur it.
But I’m sure X-Pro 1 owners will enjoy looking at the false-resolution artifacts in their images and telling themselves they’re seeing “crisper detail,” just as Leica M9 owners have been doing all these years.
Are you speaking about the images from this test or in general? Because in the images in this test I see no aliasing in the X-Pro 1. And actually Fuji has created this new color filter array to prevent aliasing and moire.
And to me it does look quite better than the OM-D, though it’s just a small test and it’s not right to take too many conclusions out if it.
Fuji’s array pattern may be helping giving them the orbs. I do like the Fuji samples shown a little better.
The X10 has a totally different array pattern than the X-Pro 1. They are not even remotely the same besides both having red, green, and blue color filters.
Fuji’s new CFA combats moire but the 6×6 pattern is actually more susceptible to aliasing than a standard Bayer! Some extremely clever processing is required to determine colour and reduce aliasing artefacts, I’ve been very impressed with what I’ve seen thus far but there has been visible aliasing on more than a few images.
Bitch bitch bitch,get a life
So the low AA filter was cool until now for olympus until the e5 (they built the whole marketing around it, remember?) now it s getting uncool because fuji hits homeruns?
Gxr, nex, m9 all use basically no AA filter and all have superb resolution and iq
The m9 at low iso is still unmatched
NEX “basically” no AA filter? Eh noo…
NEX has no AA filter ????
On par with D7000 that slows my enthusiasm, as the d90 is still superior in many tests,,, biggest surpise would be if ep3 beats em5 in base iso, goodnight
+1.
Um.. not really. X-Pro 1 has zero issues with IQ. Tho that unfortunately doesnt apply on rest of camera. Its similar case as X100 or S5 or any other Fuji camera. Near perfect image quality (or very impressive in some department), but seriously flawed camera itself.
Tho if you can live with shortcomings (hopefully some of them are FW fixable), then it produces really good photos.
Btw. M9 does too, better than most camera will ever be able to and unlike X-Pro 1, it doesnt really have some serious issue.. (except price lol).
Your post is hard to understand. Are you sure that you understand the Moire effect?
yawn, what flaws does the X100 then mate ??
Hogan wrote a really good piece on this but I’m damned if I can find it. His reasoning went something like, the difference between the beauty but impurity of analogue sound vs the accuracy of digital.
For in the end what matters is the perception of the image, where in popularism the notion of the aliasing that exists contributes to the life, the worth, and the roundness of the image. Despite what the DxO junkies find worthwhile, I’m relatively pleased with what I see, rather than how the fecked up and twisted (FUAT) numbers crunch.
Thom, we know you’re out there
now you owe me a beer 😉
Yes, and don’t forget that the X-Pro 1 will only have more picture detail if the auto focus ever decides to lock in correctly. I personally think the comparison to the D7000 is much more significant as this is a benchmark camera. The Fuji has very limited use based on what the reviews have said and the lack of lenses.
Pana GH2 is very close to D7000 at least up to 1600-3200ISO
But GH2 have a much better video so in my opinion is a better camera.
Can somebody explain this to me please: “Valentin says the E-P3 per pixel sharpness is still a bit greater than the one from the 16 Megapixel E-M5 sensor.”
Thanks…
I don’t see any mention of what lens was used on the E-M5. I’m at a work computer and can’t extract EXIF data. Anyone know what lens was used?
No, because there’s no EXIF in photo composites 🙂
ZD 50mm f/2.0 macro. They use that lens for all their tests.
So why are the EM5 images so much softer than the EPL3 images?
I don’t get it…
So why are the EM5 images so much softer than the EPL3 images? Either Olympus have gone back to a shitty AA filter or there is something wrong with this test.
I would be VERY disappointed if the EM5 has a stronger AA filter than the EPL1.
Both great cameras. But having had an E-P3 I couldn’t bear to go back to something slower.
The e-m5 is sacrificing detail for less noise at both low and high ISO. Take a look at the ISO6400 in comparison with D7000. The D7000 is noisier but has greater detail than the E-M5.
I want to know the settings of the cameras. I always use noise reduction low with my E-620 and E-PL2. And what’s the problem that they can’t write which lenses they used? No time? Not enough space? Without knowing which lenses are used such a test is not really helpful.
NR settings? If it isn’t set to off or low then this is useless.
That makes sense. In many test results, I dislike JPGs from E-M5. In general, it looks good, but if you look carefully, many details are lost. E-M5’s raw is more prefereable.
Hi Steve, unfortunately that is also my conclusion , the E-M5 files are so processed that the JPEG result is pointless. Though I suspect that on the RAW files I have tested that NR is always present and cannot be fully turned off.
The D7000 does not cost $1500. That’s the price from some random retailer, and is much higher than MSRP.
i used the dpreview tool to compare the EM-5 with the d7000.
When comparing noise level(grain) and detail level, you can see that they are extremely close, however, the d7000 does preserve color tones much better than the em-5 at highest ISO settings.
At the highest ISO settings for the EM-5, some colors(such as red) appear brown or grey, in some parts of the image.
of course this is when comparing jpgs.
Still very impressed. 🙂
They were very specific about this
‘high saturation without ever stopping or burn any channel’
So may be an aberation.
The Nikon d7000 is onle $1200 at B&H. Not $1500… that Amazon price is inflated for whatever reason.
This is remarkable- a quantum leap for micro four thirds! Panasonic better step up their game. And it’s not just about the sensor. What about the five axis image stabilization and all the other innovative features. Olympus has FINALLY come out with a legitimate upgrade!
And it’d be much better if Oly or Panny will show another quantum leap in their glass too. Even though they release many M4/3 lens, only some of them are exceptional. There are no standard 2.8 zoom (i.e. equivalent to 24-70/2.8) or tele 2.8 zoom, (i.e equivalent to 70-200/2.8)
I would also make the argument that almost all of the fine details are lost in the ISO sample of the EM5 in comparison to the other 2 cameras. It looks mushy. Less noise artifacts, but a lot less detail to go along with it. The X Pro 1 looks amazing.
+100.
This is something I don’t like about E-M5.
For me, the five axis IS is an even bigger deal than the dramatically improved sensor performance.
XPro-1 has won me over for a while. But I know I’d have more fun with the OM-D.
Good luck with the Fuji quality control. Oh, and let’s hope there are no white orbs.
Make sure you have a very large storage card and lots of time to find the shot that’s in focus.
Does anybody have a pic of a 45mm 1.8 and/or 20mm 1.7 mounted on a silver E-M5 body? that would really help me deciding what body I should order..!
Google for four thirds matching simulation
thanks but I got a porn site 8]
@henno- I found this one of the silver/45mm. Haven’t seen one w/ the 20 yet. When I get mine I’ll post a pic!
http://www.flickr.com/photos/auxo/6875141299/lightbox/
Another-
http://www.flickr.com/photos/62942199@N08/6845559297/lightbox/
Here’s a good one-
http://www.flickr.com/photos/21122418@N07/6810642956/in/photostream/
In that same stream there are a few with the Panny 20 on a black EM-5.
I had the occasion to have a look at the X-Pro 1 today. While I can’t say anything about image quality, I must say that the whole camera concept is a huge disappointment for me. I specifically did not like the ergonomics of the camera. The body is quite large, and for that size the “grip” (so to speak) is way too small. The controls aren’t there, where you expect them to be mostly. The shutter speed and exposure control wheel are not tall enough to feel comfort and the rest of the controls are too scattered. Honestly, I think the optical view finder is a big disappointment, not worth it. It is quite small and does not really give enough advantage over the EVF. So, no point to use it, even the kit lens already “vignettes” the OVF. The EVF on the other hand, is much smaller than a GH2 view finder. Worst thing: the AF is really slllloooowwwww. It hunts and pumps – forget it. Worlds apart from GH2 and likely EM-5.
So given the small selection of lenses the Fuji X System isn’t really worth buying it. A real bad entry to the mirrorless system camera market, if you compare that to the beginning of m4/3.
The X-Pro 1 does not deserve the hype and enthusiasm it receives. While I could consider an EM-5 as a GH2 replacement (although not a compelling one), I could not do the same with the X-Pro1. The X-Pro 1 plays with respect to the system design in another league, very below of the m4/3 high end offerings.
> I had the occasion to have a look at the X-Pro 1 today. While I can’t say anything about image quality, I must say that the whole camera concept is a huge disappointment for me. I specifically did not like the ergonomics of the camera. The body is quite large, and for that size the “grip” (so to speak) is way too small
That’s what I said two weeks ago after holding one in hand. There’s really no grip at all and despite of body width of high end DSLR it’s as comfortable as holding piece of wooden plank in hand.
It’s clearly designed for analog era retro brick religion who of course advertise it as second coming of Jesus.
> Worlds apart from GH2 and likely EM-5.
If focusing speed with CDAF compatible 4/3 Leica D 14-150mm is any indication of focusing speed capabilities with native m4/3 lenses E-M5’s focusing should beat GH2 easily.
I had similar impressions when I handled it. I liked the cold metal feel but that was about it. And that cold metal feel is freezing in cold conditions and hot in sunny climates too.
Ergonomically, it feels dated, which is why I previously said I thought of no reason to have this design other than retro nostalgia. At least the E-M5 is comfortable to hold with the grip (I had a couple of issues without the grip, but no way near as bad as the X Pro 1). Autofocus was not so useful either. Like sticking a giant tamron lens on a dSLR.
Can’t argue that the photo quality of its images are outstanding though. For absolute image quality, it’s a fine choice.
+ 100000000000000000000000 and 1000000000000 more SO. TRUE THE FOCUS IS DAMN SLOWWWWW and i hate the ergonomics !!!!!!!!!
This isn’t funny… I sold my E-P1 out of frustration to buy a D90 which I then sold to buy a D7000.
I hope I don’t go full circle when they release the next version of the OM-D…
Don’t be a pixel-peeping moron. You can achieve PROFESSIONAL results with a D7000. Go out and enjoy it.
Dear Admin,
please be more gentle to all Olympistas! Think about their hearts and minds! Here in Göteborg several readers has thrown their computers out the window after reading your posting. At the moment the police are trying to persuade an Olympista NOT to throw himself out from his window on 10floor; he cannot face the fact that there may be cameras better than EM5.
Dear Olympistas,
I have a FT3 rumor for you: All other camera manufacturers are now considering shut down their camera operationa as the Olympus EM5 is superior to what they produce. Phase One has said they will rename their “IQ180” (highest rank at DXO) to “almost OlyIQ180”.
Do you feel better now?
What a moron.
I agree, this guy, K, is stupid. What is the point. People are just saying that image quality is not everything. I don’t mind the X Pro 1 deliver better quality, but i could still like EM5 for what it can do…
It’s his very small penis that makes kylberg so ignorant. Drink more beer, kyl — blurred vision makes it look bigger.
+1000
What has DXO got to do with the real world? Though I suppose it’s appropriate for the imaginary world.
Oh, my dicky ticker — I am fainting 🙂
(…swiftly, and with style…)
What are you talking about?
hehehe
OM-D seem on par until you bump up resolution bit up, otherwise it has same shortcomings as all m4/3s (or 4/3s) and same advantage. I dont think comparing m4/3s and APS-C is much useful, its simply different.
I had m4/3 camera, sold it long time ago and must say OM-D really pushes right button and I might even buy it, tho some really high quality lens would be nice, like 35-100mm f2 and such. Or at least decent AF for them..
http://egami.blog.so-net.ne.jp/2012-03-19
New Olympus patent 150mm 2.8 Macro
for 43 or m43????
@Robbie. Nice, i think this is a new M43 macro lens, and i will see then also water sealing.
(think on this, 300mm F2.8 macro FF have you, look this before) 😛
An Olympus Zuiko Digital 50 mm f/2 Macro was used on all cameras. No noise filtering at all in the cameras, sharpnes adjusted to normal.
Best
Valentin Sama
DSLR Magazine
Editor
Thank you very much 🙂
What aperture?
Diffraction would limit E-M5 image sharpness sooner than others.
You can’t use ISO6400 and ISO12800 to compare detail as there is too much noise and NR happening. DSLR Magazine did it correctly as you can see the heavier NR on the E-M5 at low ISO
The two conclussions I draw from the test are:
– The Fuji is clearly better for whatever reson already explained avove. I will not buy it because is to expensive, lack lenses, and autofocus is a snail.
– The Lightroom raw coverter for Oly cameras is a piece of sh…t.
Regards
No doubt X Pro is producing IQ above EM5 , class leading image quality fr all current mirrorless . That is the reason it having my attention . Let’s see if Pana / Oly ‘s next products can close up the gap in these few months .
There is no difference between APS-C and M43 in terms of PQ by now. Only difference is DOF and crop rate when legacy lenses are used.
Another difference is the collection of quality lens. APS-C (including Nikon, Canon, Sony) offers a mountain of great lens (L, Nano, Zeiss). Their users have options to choose from a variety of quality lens(mediocre, decent, very good, excellent lens. Of course, standard 2.8 zoom too (i.e. 17-55/2.8, 24-70/2.8, 70-200/2.8). M4/3 in fact also offers many lens too but only some of them are that great.
I see none of any 2.8 zoom from both Oly and Panny. (Of course, Panny already presented their 2.8 concept lens, but when exactly it will be available?)
you talk of APS-C cams as if they are of the same system. you forgot to consider you can use both 4/3 and m4/3 lenses on m4/3 bodies.
Can the JPEG engine of E-P3 update via firmware to be the same as E-M5 ?(They are both using TruePic VI)
I am not sure I understand, Mr. Sama.
Are you contending that there is no resolution advantage from the 12 Mpx to the 16 Mpx Oly sensor? Do you mean per pixel sharpness?
Is it at base ISO or all the way up?
If I get it well that might mean that Oly was correct in stating that 12 Mpx was where M4/3 was meant to be. So how to explain the improvement in noise in the OM-D?
It’s just a matter of better denosing, thaat however obliterates detail?
A bit confused, here.
Not clear, amalric… either for us. Do you see really more detail in the image respect the PEN sensor? Not for us…
Maybe Olympus is using the higher pixel count to improve high ISO rendering? The quality improvement is there.
But more detail?
Maybe you are right and better denoisig puts resolution on par with the 12 Mpx. sensor. Seems a reasonable solution…
I’ve compared 16MP GX1 to the current 12MP Pens and it’s difficult to see an improvement in details in real world pictures. However, you really need to use test charts to get accurate info on resolution.
I was unable to make head or tails of your sample, perhaps my look was too quick.
However it is perfectly possible that a higher density sensor gives better results at high ISO, and slightly worse at base ISO.
It depends from the physics of read noise, which to me is still a mystery, but it has been stated before.
Intersting find, I’ll cling to my Pens for longer, and save a lot of money 🙂
That 12MP sensor is five years old tech and wasn’t even very good already then so shouldn’t be hard to make all around better performing sensor.
If noise is higher (lower SNR) at base ISO situation isn’t going to get better at higher ISOs which need amplifying of even weaker signal.
Here you can do comparison from similarly processed RAWs and E-M5 gives at least same or mostly better details in fine (low or high contrast) detail spots than E-P3.
http://www.dpreview.com/previews/olympusem5/9
So, those people who said they wanted a better sensor?
They were 100% right by the looks of it.
Can you imagine how good the E-M5 would be if Olympus managed to get hold of a Fuji X-Trans sensor for it?
I’d buy one…
Fact: Image-Quality is good enough!
Now: we need features ! (Panorama, Remote, mulipble pictures, internal flash ;-), ….
Pro: size of the lenses (of the Olympus m4/3)
Am I the only one to get an impression that on the first photo everything is out of focus for E-M5?
Apsc will ever be superior compared to 4/3 as full farme is to apsc. Lower dynamic range for a smaller sensor size
I’m glad you guys are happy with being equal to Nikon D7000’s JPEG. Nikon’s jpeg engine isn’t very good and Nikon users usually don’t like the jpeg, but if that quality is good for you, then that’s all it matters.
I think the rationale is that ‘cropped’ sensors are hitting a limits to growth wall.
By increasing pixel density you get better performance at high ISO, but you are not necessarily getting more performance at low ISO. That was the original reason for changing the base ISO.
In theory all sensors could hit a ceiling. For instance the NEX 7 is said to be no better in pixel sharpness than the NEX 5n, and is much more noisy.
16 Mpx is as much as m4/3 can go, probably with the present technology. Indeed Fuji was wise to find a workaround by weakening the AA filter.
I wonder if sensors of all formats are now operating at borderline densities. This is hidden by the fact that denoising algorithms get better by the year, but they might have diminishing returns too.
“I think the rationale is that ‘cropped’ sensors are hitting a limits to growth wall.”
no such thing Am, there will always be the next big thing and all the possibilities haven’t been covered yet. However what we can say, is that at some point the need for more becomes less. For instance Ive never shot anything at ISO 6400 in my life, Im not sure I can even contemplate the circumstance for which I would need it, but then here we all are.
Realisticly the future isn’t sensor size RAW files or anything like that, we are on our way for better or worse to consumerism in photography, with 18 stop DR noiseless files that are just processed incamera.
We take heart that, at this point it is the photographer that matters. I know where I stand.
Riley.
I was referring to what Valentin found. No improvement over the E-P3 at base ISO. So what is Oly exactly selling?
BTW I am pretty sure that even FF is on a curve of diminishing returns. What is the advantage of ever increasing files if pixel sharpness doesn’t improve.
Possibly the E-M5 is a good bargain, not because it allows a few to shoot at 6400, but because it is a well rounded tool.
NEX 7, with 24 Mpx shows the opposite, and according to samples has not more resolution/sharpness. Seems to me that everybody is hitting the diminishing returns curve.
E-M5 pixel spacing 3.75 microns (17.3mm/4608 pixels)
Nex7 pixel spacing 3.91 microns (23.5mm/6000 pixels)
EPL3 pixel spacing 4.29 microns (17.3mm/4032 pixels)
D7000 pixel spacing 4.79 microns (23.6mm/4928 pixels)
XPro pixel spacing 4.82 microns (23.6mm/4896 pixels)
D800 pixel spacing 4.89 microns (36.0mm/7360 pixels)
Nex5 pixel spacing 5.09 microns (23.4mm/4592 pixels)
5Dv3 pixel spacing 6.25 microns (36.0mm/5760 pixels)
5Dv2 pixel spacing 6.41 microns (36.0mm/5616 pixels)
L8.2 pixel spacing 6.85 microns (27.0mm/3936 pixels)
LM9P pixel spacing 6.90 microns (36.0mm/5212 pixels)
5Dv1 pixel spacing 8.24 microns (36.0mm/4368 pixels)
D700 pixel spacing 8.46 microns (36.0mm/4256 pixels)
The one trick FF holds is shallow DoF, which at the same time with good optics. MFT benefits from a range of very fast lenses, some of dubious quality but nonetheless it is in the range of shallow DoF.
But there’s something else
In order to facilitate this we have to talk about AF too. This b/se one of the differences between FF and mFT is the magnification of the subject, where we both know mFT is 2x. This makes AF harder to accomplish, and this shows more with shallow DoF. Remember with the same lens mFT has half the DoF
P&O simply must get on top of this, while the potentials are there, they themselves haven’t offered the market enough in this capability.
@Riley “Remember with the same lens mFT has half the DoF”
How are you working that out? Do you mean the very same lens, or do you mean the same FF equivalent?
In any case, the same lens on both FF and m43 will display the exact same image towards the sensor, on m43 the capture area is far smaller than that image so only the centre of it is captured. The lens DOF has not changed at all.
You are down to make up things? CDAF is perhaps the most precise AF around…
Not the same can be said of dSLR’s PDAF. LOL