skip to Main Content

DxOmark tested the E-P3 sensor: Pretty much the same old sensor.


Image courtesy: Clubsnap.

Is DxOmark destroying the “legend” that Olympus used a new 12 Megapixel sensor? DxOmark (Click here) published their Olympus E-P3 sensor review: “We have tested the PEN EP3, the latest 4/3 camera by Olympus and the first results are now available. The Olympus PEN EP3 apart from changes on the exterior, it is pretty much the same sensor as the Olympus PEN EPL2 and Olympus E5.”

I would love to meet an Olympis enegeneer to ask him once and for all what they did and what they did NOT on the “new” sensor.

Preorder links:
Olympus E-P3 preorders at Amazon US, Adorama, BHphoto, Amazon UK and Amazon DE.
Olympus 12mm f/2.0 preorders at Amazon US, Adorama, B&H, Amazon DE and eBay.
Olympus 45mm f/1.8 preorders at Amazon US, Adorama, B&H and Amazon DE.

  • WT21

    You mean the rumor that 43rumors created? I didn’t see this “legend” anywhere else ;)

    • MikeS

      “Anywhere else,” indeed.

      From Olympus E-P3 press release: “Focus and shoot faster with the camera’s new 12.3-megapixel Live MOS Image Sensor, TruePic VI Image Processor and new FAST AF Tracking System.”

      Evidently, Olympus itself is a big believer of the “legend”.

      Also, I’d prefer that someone destroy the legend that DxOMark is at all credible.

      • And what exactly makes you say they aren’t credible? You’re just another guy with a bold claim on the internet, while DXOmark actually releases (most) of their methodology, and does their tests with scientific rigor. Do you?

        • DxO is also just a point of -scientific in this case- view…

        • Ross

          Who gives a stuff what DxOMark says. It’s the results from the camera that matters & if the camera producers results that I like, then I’m happy.

          • DXO measures “results from the camera.”

            • Nick Clark

              Not reeeeeeeally… DxO themselves state that they test pure RAW files and don’t take into account any in-camera processing. So they measure ‘results from sensor’, which is very different to ‘results from camera’.

              • Pure RAW? There is un-pure RAW? I’m no expert on this, but are you claiming DxO somehow pulls a different, “purer” dataset from the sensor than all of us using non-Olympus RAW developers?

                What I’m saying is that most likely DxO’s results agree 1:1 with those of anyone using Lightroom, for example. Olympus Viewer is another story as they can use proprietary data in the RAW, but nobody uses the software so it’s a moot point.

                • Do

                  AFAIK in the moment you can see a “RAW image” it’s an interpretation of the RAW data. DxO numbers are also some sort of interpretation of the data, but a numerical one. They may be congruent but but they need not be.

                  • Steve

                    Ah….what fun. When DXO says, “ace sensor” everyone is quick to quote it as proof…and when it doesn’t people start disputing their methodology, impartiality, etc, etc.

                    For sure, this test is bad news for Olympus, because, if nothing else, it’s really kicked their marketing in the balls. Higher DR ? It seems there’s barely enough DR for a jpeg at iso1600.

                    It’s a shame. I had hoped this sensor would achieve more than just a faster read-out.

                    • DXO only test sensor by clean and only use, red light, green light and blue light, when you use camera normal for exsample a blue wall, the is normal a light spectrum from UV to IR but only most blue light.
                      When Olympus lift up for use IR light and also use a bayer filter RGBIR not RGBG so is normal Bayer filter use in DXO test.

                    • Do

                      Of course it’s bad news for Olympus and I guess that the DXO measurements show what’s possible in theory and what’s not, i am more skeptical about it’s significance in real world. Most the time I am a jpeg shooter anyway (beside occasional adjusting white balance or exposure compensation in Oly viewer) – it’s only one year ago I bought my first Olympus and before that i read a lot of reviews and looked through masses of samples, and i found the dpreviev tests more significant than the DXO tests.

                  • BCK

                    Remember: All M4/3 cam’s have corrections, even in raw.

                    But it IS still a raw.

                    • re: Do above, this comment system is broken in so many ways it’s not even funny.

                      Do, for JPEG shooters DxO is indeed pretty much useless and can be ignore for real-world applications. For us RAW shooters not so much. I’ll stick with my E-PL1 which is identical to E-P3 in sensor performance.

        • MikeS

          “Scientific” is a stretch when it comes to DxOMark ratings. Though it’s commendable that they are consistent in their methods, the very fact that the ratings are backed by a private company – whose products are marketed (via DxOMark ratings, btw) and sold – automatically casts doubt on their supposed “rigor”. There is also the matter of the fact that they arbitrarily choose metrics (notably excluding resolution) in order to provide an arbitrary rating, based on an average of metrics (which assumes that the 3 arbitrarily-chosen metrics are equally weighted) with arbitrary scales (e.g. the 30db SNR threshold, or normalizing to a 300dpi 8×12″ print). There’s also the matter of the interrelatedness of each metric (e.g. dynamic range can suffer at higher ISOs, but the DxO score is based on the ISO that gives its maximum), which is explicitly not taken into account. I could go on and on.

          In the end, the primary fault of DxOMark ratings is that they’re taken too seriously, particularly the overall score. Instead of leaving scores separate, they are averaged for the simple determination of what is “best”, even when what is truly best depends on the intended use. That’s why the Sony a900 is rated slightly lower than the Nikon D3S, despite the fact that the two cameras are designed for vastly different purposes (one doesn’t use a 24MP camera primarily to print uncropped 8-by-12’s).

          On a side note, science is actually my line of work, so I do release all of my methodology when it’s relevant, and perform studies with scientific rigor.

          • They release the data they measured, one just needs to compare the graphics instead of look just at the scores (you can check them normalized for the 300dpi 8×12″ print or just the pixel data) that indeed might give some false idea.

            There is another issue with DXO measures… one that is noticeable, for example, in GH1 ISO 800 values: Some cameras cook the RAWs, some do it only at some ISOs (like some Pentax, DXO actually released some notice regarding that), some have options to disable that, some (GH1) have a glitch that cooks them applying NR in the RAW for ISO 800 when the settings for NR are 0 – I think if you set it to +1 it doesn’t apply NR at ISO800. o_O

            If there was some resolution measurement bundled in the data set it would help in detection of cooked RAW files, not to mention it would give another comparison detail, but, unless documented by the manufacturer, it’s basically impossible to ensure ‘pure’ RAW data analysis for all cameras, resolution is also lens dependent (you can actually see that data on the lens ratings).

            With that said, DXO does a great job measuring RAW performance of a camera, but that alone doesn’t define the camera, things like size, speed, lens range/prices, JPEG processing, etc. are often unique and should be taken in consideration when selecting a camera.

          • Valid criticism. But we have to work with what we got, not what should be. DxOmark is the closest we have to scientific rigor combined with a variety of tested sensors. It has its limitations, but I’d imagine you agree that the metrics themselves are well-documented, reliable and falsifiable. There are other items they could be testing, but they’re not – someone else has to do that.

            The way I use DxO is I ignore the scores altogether, and look at the graphs, comparing my current and past cameras in areas which matter to me (DR and tonality, mainly). For that it is vastly superior to eyeballing RAWs of different scenes, or worse yet, JPEGs as most here seem to do (I shoot RAW).

            This shows that my E-PL1 is all but identical to E-P3 where it matters to me. So it is up to camera features, handling, UI and size – and the last part is what might convince me to buy a second MFT camera in E-PM1 as a backup.

            And DxO doesn’t ignore resolution – its impact is included in the “print” 8×12 tests. It doesn’t really matter whether it is arbitrary or not – it’s a good standard test which is easy to grasp, and normalizes the results of cameras with vastly different MP count – which was an early criticism of DxO, addressed with this metric.

            • spam


              • I think you’re spot on there.

                Incidentally, the G3 test shows exactly the same thing (That it’s no different from the G1 really) so my GH1 will live a little longer, since unlike Olympus, we with Panasonics can’t use our camera JPEGs! :P

  • 1212

    in another article is written the high iso is so much better on the new e-p3… now what is true?

    how can high iso so much better when the sensor scores the same?

    “Overall, the new Olympus EP3 has been able to vastly improve the high ISO output and may finally creep up within the next iteration of the camera (or maybe even in RAW).”

    • > now what is true?

      The truth hasn’t changed for many years. DxO Mark is still irrelevant.

      • Lula

        because a dummy says so?

      • +1

        The sooner blogs stop posting DxO scores the better as far as I’m concerned. It seems to me their only purpose is to give camera nerds something to argue over.

        • cocute

          DxOMark is bad for those who want to hide things

          • dumbo

            yup, raw tests always hurt. now we are looking at real results not noise reduction software.

            when is olympus actually going to improve the hardware and not the software!

          • They’re pointless because if the results a camera produces are good to the eye, then that is all that matters. If people like to get all up in arms over bar charts then by all means have at it, but I think it’s safe to say you’re more tech geek than photographer at that point. From looking at actual photos taken with the E-P3 I see nothing that that will hinder my photography. If a bar chart says otherwise I really don’t care.

            • Uro

              Maybe to you, but I pixel peep and I see a lot of noise compared to Panny’s G3 or Sony NEX-C3.

            • Steve

              Some people’s eyes are better than others so a normative test like DXO has its place too.

              Plus it helps people who shoot raw to get a better idea of what they’ll be working with.

              In this case, it seems there won’t be much benefit to shooting RAW with the E-P3.

              For the money, E-P3 is not looking that good and the G3 is looking much better.

    • cocute

      yes ISO in E-P3 is better in JPEG images, in RAW is same in all pens

  • Read this today too. I don’t believe it for a second. The EP3 outdoes my EP2’s image quality and other sites found similar results.

    Chris Gampat
    Editor in Chief
    The Phoblographer

    • DXO tests for RAWs, JPEG performance doesn’t count. Has any other site done actual rigorous head-to-head tests on RAW DR, for example?

      I feel dirty just responding to someone who’s blog is the unpronounceable The Phoblbgher.

      • No one has because no one got the Olympus software with their units nor does Lightroom support it yet.

        Let me console you: you are dirty.


        • So why are you even commenting in this thread, and calling on false authority (ie. your and other sites which have only done JPEG tests WHICH ARE INVALID FOR RAW PERFORMANCE COMPARISONS).

          Don’t imply you know something about camera testing when even the most basic concept (that there’s no necessary causal relationship between RAW and JPEG performance) eludes you.

          This is why internet sucks – everyone can look like an authority by running a blog.

        • cowgaR

          you must be joking mate. You read m43 rumors, have tens of lenses at home, few dslrs, been through 43 system and jumped to micro recently, and tried Bible->DXO-Lightroom career, and yet YOU ARE CONCERNED WITH JPGs !$%£!

          stop joking please, these cameras don’t cost £200 pounds, even single lens cost twice as much! I HAVE NEVER USED JPG from any professional camera (not even my compact G10/12) as the difference is absolutely striking (once properly edited, one can’t deny horrible JPG artifacts, screwed noise reduction, 8bit color limitation and overblown highlights, shifted white balanced with not possibility to fix it and I can continue)

          do I really need to remind here, that NOBODY is shooting to JPG anymore with these cameras?

          I usually do not comment, but this made me angry as the post above claming *stupidity* as “it matter what is comming out of the camera”

          what is comming from camera is RAW file, I don’t care about the rest as no-chip, not even in hasselblad cameras, can’t replace DXO lens correction -> Lightroom -> Noise Ninja -> Photoshop combo, where editing one photo sometimes takes few hours (sometimes one minute) with a cost of few hundreds of pounds JUST for the software and hours of Quad Core i7 computing power to edit it, and you expect results from some, let me say

          “digic VIII ultra-sonic dual core processor with cherry on the top”

          inside the camera (with a performance of a calculator) to do the results INSTANTLY while you shooting?

          only noobs use JPGs, and only noobs care about the JPGs results from these cameras (and honestly, they have been horrible most of the time)

          If I want to use JPG, I buy myself some Panasonic TZ for £200, not spend thousands for equipment.

          Sorry guys for a rant, needed to stop reading and post something to relax… you eat marketing too much

          • Boooo!

            People forget that even such a “small” thing as correcting white balance is a destructive operation on JPEG files.

          • greyhat

            I agree that is a deception that DXO mark sensor measurements are bad news to Olympus (even if they are wrong): many enthusiasts look at DXO mark as a guide to buy or not to buy, and this affects sales.

            I may agree that you need RAW to get most of the camera, but saying that only noobs do JPEG, that is not true.
            If you nail the composition (no need to crop after), the exposition and white balance you can end up with really good results on a 12MP camera. It is similar to film, what you set in camera for the frame is 99% what you get on print, for good or for bad.
            Now many of us are used to handle/fix those settings later. Laziness, not expert enough or conservative could be some of the explanations.

            I do RAW in one of my cameras. Auto white balance is not good. Exposure measurement could be better.
            If I buy a smaller camera is to teach myself composition as a carry-always-with-me camera and not to spend more time at home processing RAW files.

            You can call Ken Rockwell many things, but he is not a noob and he only uses jpeg (except for M9).

      • Steve

        Chris, your tests really weren’t that good (low DR shots) and your conclusions indicate you don’t have a well trained eye (no comment on saturation or color accuracy). Either that, or perhaps Oly have been very generous to you ?

  • Jasper

    new series is a bad/good as the old ones.. no match for NEX.


    Olympus PEN EP3, a good micro 4/3 sensor already familiar to us

    Tests show the new Olympus EP3 provides similar measurements as the previous Olympus model.

    In spite of its goods results we still encountered some old problems:

    In high ISO (Lowlight ISO score of 536 which is a bit low in comparison with the results of the latest APS-C sensors (the score is 1083 for the Sony NEX C3 on Lowlight ISO, for example).
    The Olympus PEN EP3 has still the same behavior in low ISO for the dynamic range and the score at the minimum ISO is not increasing as expected (+1 EV in comparison with the next IS0).

  • Regular reviewers usually judge the images, like regular people but with extra care. Dxo judges (somehow) sensor performance, which has a connection to the final image, but it’s not as close as people think, and it ignores many other important factors (which they acknowledge).

    I would take this as verification that the sensor is similar to the other Oly sensors in the way Dxo measures, but not necessarily the same sensor (who can tell?). It could be that the changes to the sensor have more to do with autofocus than with noise and DR.

    • Sony nex is better ??
      NONSENSE !!!!
      With my epl1 sony cannot do nothing !!! In sony ,no estabilization system, lenses awfuly expensives,no phisical whell of photo, no flash,3 lenses-plastic ones ,is so hard to work in manual in the nex ,and the most i hate the shutter sound is awfull , buahhh
      In my epl1 or epl2 , i love it the shutter sound, in body stabilization of 2 axis,( so i can buy old optics and dont care ’bout stabilization- there’s a lot of adapters in a really good price)more than 30 optics 4/3 , flash included,is so fuckin simple work in manual,really beatifull colors , not ice colors of the nex , viewfinder of 100% of range etc etc etc!!!
      I test the sony nex5 per 1 week then i returned ’cause i dont like it at all !!!!

      • Hello Ernest

        Are you even paying attention to what is being discussed. Nobody here , DXO or anyone is telling you that such camera is better than another camera.
        This is all about the merits of the sensor , and ONLY the sensor. if you want to believe that your oly has a sesnor as good as the sony ones , dream on …


        • Jojojo so sorry you’re right we’re talking of the sensor !!!!
          So how can i start !!! With the fact that some guy says that he can use retro-engenering to measure a really complex sensor dont make me laugh , the retro-ingenering is really expensive for a particular ,also talking ’bout photography if the half of the sensor is prepared to catch IR that’s fuckin awesome , that means that i dont need to use IR filters, and the camera can a avoid parasite light !! Plus jiii a good protographer Never i mean NEVER uses more than 1000 Iso , u buy brighter lensen, or use the stabilization system with a low shutter speed but never rises to high the iso , or should i remenber u that one of the most important cameras in the world only uses a maxim of 2500 ( leica M9),sure the sensor is diferent. butt that’s the thing , oly is making a good things with the new sensor changing the type of photosites that a normal sensor uses , conserving 12 MP,and ignovating like always does , dont suprises u , if nikon or canon in a time comes with a similar system of autofocus like always happens !!!!!

        • Jojojo so sorry you’re right we’re talking of the sensor !!!!
          So how can i start !!! With the fact that some guy says that he can use retro-engenering to measure a really complex sensor dont make me laugh , the retro-ingenering is really expensive for a particular ,also talking ’bout photography if the half of the sensor is prepared to catch IR that’s awesome , that means that i dont need to use IR filters, and the camera can a avoid parasite light !! Plus jiii a good protographer Never i mean NEVER uses more than 1000 Iso , u buy brighter lensen, or use the stabilization system with a low shutter speed but never rises to high the iso , or should i remenber u that one of the most important cameras in the world only uses a maxim of 2500 ( leica M9),sure the sensor is diferent. butt that’s the thing , oly is making a good things with the new sensor changing the type of photosites that a normal sensor uses , conserving 12 MP,and ignovating like always does , dont suprises u , if nikon or canon in a time comes with a similar system of autofocus like always happens !!!!!

    • Steve H

      I was thinking the same thing, that the “new” part of the sensor is the autofocus. Was there ever confirmation that they added IR photosites and that’s how the autofocus works?

      If they are “leaving out” the IR filter over a portion of the photosites and then using that in the autofocus, that would be a “new” sensor design. A pretty simple sounding (but also radical in terms of vastly improving autofocus) redesign.

      • jim

        And If they have added IR sites but still kept the image IQ the same then I guess they must have improved output as it uses less sites to get the same IQ!

      • Yeahhh that’s true new type of sensor for sure !! Butt we always need to remember that in photography u need 3 important thing
        1.- good sensor
        2 .- good optics and
        3 .- good software engine thats makes a good brand or not!
        Also is important the stabilization system bufff in the ep3 u have it all^^


    well the dxo sensor test measures noise… so how can they say noise is the same, while others say high iso noise is much lower.. makes no sense.

    • Maciek

      they are measuring noise level and not how the noise affects the IQ

    • > the dxo sensor test measures noise […] makes no sense.

      The DxO misses important aspect that noise too has so to say a structure. “Better” the noise, easier it is to remove it. And except DxO nobody cares – everybody just wants the final picture.

      • Hawk

        i think olympus has payed some reviewers.

        • Digifan


        • sparedog

          they certainly censored some, i can tell you that much

        • OlyFan

          Just like Sony pays DXOMark?

      • Nick Clark

        Apparently a whole lot of people care, given that this post already has over 100 replies :)

      • This is the most interesting aspect, at least in Oly’s doings. Oly has constantly refined its denoising algorithms, so that they have earned almost one ISO stop per generation. Good denoising, better SNR, also ensures better DR at low ISO.

        All this is in the processing part, and the proof is that by shooting RAW even the best software with the best denoising is unable to much improve on Jpeg.

        DxO also underrates Oly’s ISO by using a different ISO definition from mftr. and so may be off even in RAW by one or two stops.

    • evenflo

      …because they MEASURE it, while the others just SAY it.

  • Maciek

    lets wait for more RAW IQ reviews from others.

    • Jaaaa not that but i use to work to nikon !! and nikos does that, olympus is ignovating all the time , and the brands always are complaning!!!
      Who was the first in put real live view in the reflex!!
      who was the one that came with the mirrorless solution!!!
      Who is the creator or the sswf ( super sonic wave filter for sensor cleanin
      And more and more always making things !! In 2009 canon says that there is n use for m4/3 camaras and mirrorless system that they never do such a thing !!!!
      And what happens !!! Now u got thAt canon in2012 are gonna join the mirrorless system !!!
      Jaaaa think ’bout it ^^

  • Simon

    dxo also concludes the LX5 sensor has a higher dynamic range than m43 or some aps-c sensors if I remember correctly.

    My only advice to prevent oneself from getting paranoid or depressed is a quote from Mike Kobal: let’s get out and take some pictures.

  • furb

    It must in the TruePic VI image processor vs. the previous versions…i think

    • Mr. Reeee


  • Everyone’s talking about an ISO leap forward in reviews, but anyone saw a RAW comparison? or they all JPEG? because i have no doubt that processing has been improved

    • Lucious

      correct you nailed it!

      JPG looks better but RAW are the same.

      hence the sensor benchmark from dxo does not show a better performance.

      and i don´t care about JPG…… i shoot RAW all time.

      • Digifan

        So you’re no pro than. A lot of professionals I know are interrested in the best jpg engine, they want to make money and not waste time. They want a live and have families, they don’t bury themselves in their offices struggling to et the best out of their RAW’s because the jpg’s don’t cut it.
        Man, RAW is for backup and special cases not for default.

        • Well, saying that RAW is just for backup and special cases is a bit too much…For consumers, thats true, but for most enthusiasts its the other way around…

        • Haha, 1st time I heard that. lol

        • Neonart

          Agreed! I use Raw for crucial paid jobs. When I’m shooting stuff for web or small prints or for pleasure the wonderful Oly jpegs do just great!

  • johnjoe

    Pretty much expected, the RAW image quality from the EP3 is identical to all previous m4/3 models.

    The “improvement” that you guys see in online samples mainly pertain to JPG output. The EP3 JPG engine does more aggressive smearing and sharpening, which give the false appearance of better high iso performance. I prefer older JPG engine to the EP3’s.

    • > which give the false appearance of better high iso performance.

      When DPR would do proper review, I’m pretty sure they will mention if camera introduces false details (== artifacts of aggressive sharpening (after aggressive blurring to cover the noise)). To date I never heard anybody complaining about the false details.

  • Joey

    Never really paid much attention to stuff like this before, but went over to have a look anyway. So unless i’m mistaken, DxO, by looking at ‘fundamental’ RAW have given what they call the same same sensor 5 different scores (ie E5 EP1,2,3 and EPL2)? Also reading their method i’m slightly sceptical, they say they reverse engineer software enhancements to get the ‘real’ image, but surely that leaves it open to interpretation, at what point can they be certain they have reached the core, or in other words, do they find what they expect to be looking for? As a non-technical person i admit i may have completely misunderstood, but it seems a rather pointless exercise?

    • Lucious

      no it´s not pointless!

      read the luminous landscape article about the DxO sensor test.

      a few self-called-experts here should read that article too!!!

      • Archer

        That article was written by DxO.

    • Hell yeah u can see it !!
      The af is awesome ,also the dinamic range , i always love the olympus colors cuz im a photographer, but thing ‘ bout it , if i dont need to use a infrarred filter , like in a simple reflex,, like in my nikon d90, because of the new IR lenses and the new sensor that avoid parasite light , that fuckin awesome !!! All the posibilities woww

      • The dynamic range is awesome? Sure… :D

    • There’s more to “sensor” performance than just the sensor itself. Noisier circuits, different heat dissipation characteristics of camera bodies, and other factors that can’t be controlled for can result in the same sensor in different bodies yielding slightly different results.

  • According to technical guys on getolampus, on my question “how they managed to improve noise perf. and image quality, is it the sensor or the true pic VI?” the OFFICIAL answer was:
    its processing! The TruePic VI is much more powerful then Version V.

    And after this test, i do believe its true.
    Just compare the available RAW files with JPGS. I dont know WHERE i did read it but they stated that: you wont get better results shooting RAW then shooting JPGS! So Oly has really improved its processing so its second-to none! I cant imagine what they would do with a better sensor like the G3 sensor :D:D *keeps dreaming on a Oly Pen Pro*

    • Lucious

      unfortunately RAW has more flexibility to offer then JPG.

      you might reach no better noise shooting RAW with these new models… but whitebalance, highlight recovery etc. are important points for me and why i shoot RAW.

      a better JPG engine doesn´t help me at all.

      a better sensor with a real lower noise would be better for me.

      • +1

      • chi

        agreed, i could never get the white-balance correct, so i have to shoot raw

      • Reto


  • Joey

    lol ok just looking further and apparently my 5Dii produced better images than a Hassey, now while i love my 5D i beg to differ, I think rather than produce these tables they should start selling whatever they’re smoking, because they could make a fortune :)

    • jrk

      Theoretically, the medium format cameras don’t allow for very high ISO. I saw an article about it somewhere but I forget where I read it.

      • Digifan

        Yeah, and your point? High ISO is paramount?
        If you NEED big files you NEED an LF/MF.
        That’s why 35mmFF still exists, it has it’s uses I won’t deny, but for specialty cases. Just as RAW is for backup and specialty cases. If your camera can’t deliver best jpg you are bound to RAW which means you wastea lot of time to convert and tweak your result.

        • That what you call “wastea lot of time” i call darkroom, and i don’t see anything bad about it…:)

  • Improved sensor for IR based AF, improved processing, barely same old live-mos technology… We will see in raw shoots. I still dream a Foveon-like sensor in Oly camera.

  • Raist3d

    The accusations that dxo is not credible are so hilarious. For one many here that note “other websites” show improvement miss te fact they were comparing joeys and not raws. Oops.

    Funniest is when others quote sites with favorable jpeg opinions using dxo’s own software! You would think dxo who wrote the software know best how to run it for a test, no?

    Finally there are several websites this time pointing in the direction of dxo anyway.

    It’s tiring to see the same sensor performance even if it’s a new/design whatever. But on this class of camera with the super spee it’s fine. This is not the e-5 (an e fivers rejoice, the e-5 still does a bit better)

    • Raist3d

      And yes by joeys I meant jpegs ;-)

      • jrk

        IOS autocorrect?

    • Esa Tuunanen

      Olympus should definitely get fully modern sensor from bottom to top fast instead of keeping tweaking old one but you should stop bowing to DxOMark as second coming of Jesus.

      I checked Luminous Landscape’s article about DxOMark’s methods and that automatic resolution “normalizing” for compensating differences in sensor resolution is questionable.
      It assumes that noise is random and can be averaged away perfectly when scaling down images of higher MP count weaker per pixel performance sensor but in bayer sensor camera noise isn’t anymore random at pixel level after demosaicing.

    • > The accusations that dxo is not credible are so hilarious.

      Hi! Haven’t seen you trolling here for a while!

      As I was one of those who made comments I beg to differ. Nobody in comments above called DxO “not credible”. They have a method. They are consistent. The problem with the method that it is not very relevant to the final IQ – because it is a raw data dump from sensor, not even demosaiced.

      And DxO even admits that to be true: they are interested in testing the RAW data. They are not after the best possible final image you can make out of the data, what DPR stuff always attempts – DxO are after the bits and grits of the RAWs.

      > You would think dxo who wrote the software know best how to run it for a test, no?

      Being a software developer, I can write a test which would say that 2*2 = 5. By your logic, it can’t be wrong, because since I, author of the test, know better how to run it and can confirm the results to be true. Duh.

  • I agree.

    “dxo also concludes the LX5 sensor has a higher dynamic range than m43 or some aps-c sensors if I remember correctly.”

    Correct. LX5 has higher DR score than GF2 for example, because it starts at a much lower base ISO (70 vs 150 on GF2), and that score is based on base ISO. However DR decreases much faster in LX5 and is significantly lower at the same measured ISO than GF2.

    DxO scores show the potential IQ that is attainable through RAW. So it is indeed irrelevant to many people but should not be misinterpreted.

    Moreover, the signal from the same sensor in different cameras is processed differently (read-out circuitry), that’s why all older m4/3 cameras get very similar but not identical scores (try EPL2, EP1, GF2, they all have the same sensor, only two-point difference).

    I quote dpreview, first page of their hands-on preview of the EP3:

    “The sensor is still the familiar 12MP unit but the camera’s processing is greatly improved.”

    That explains why high ISO JPEGs are improved: there’s a new processor, too. But raw quality is unchanged, as measured by DxOMark.

  • cocute

    In short, if you shot in jpg buy the E-P3,
    but if you shot in RAW use the E-P1 (is more cheap)

    • Steve H

      Except for the big autofocus boost.

      • jim

        and the touch controls, and the OLED hi res screen, and the built in flash and 1080 AVCHD….

        What have the romans done for us!

  • 1. DxOMark is a professional-looking site that, on the one hand, displays sometimes-wrong data very elaborately, while completely leaving out really important data items on the other hand, thereby repeatedly leading to meaningless results which would contradict all hard real-world results and experience. Don’t trust DxOMark.

    2. From the real-world RAWs I saw and scrutinized, the sensor does indeed not seem to be much different from any other Olympus 12 MP sensor, including the earliest ones in the E-30 and E-620. Visual differences along the different models so far amounted nearly exclusively to different amounts of antialiasing, resulting in more or less fine-detail resolution and sharpness. My impression is that the E-P3 sensor very much follows along the same line, image-quality-wise. (Which is something, for example, DxOMark does not take notice of at all, they simply don’t measure it. In practice, it can make a such a difference that 12 MP images from one sensor could actually resolve better than 16 MP images from another.)

    3. From everything which has been published so far, it is obvious, though, that the sensor really is brand-new and completely different than all earlier 12-MP-sensors – but only in the ways it provides AF functionality and is read out, all of which seems to have been significantly improved, thereby providing the fast AF that keeps being reported everywhere.

    4. If real-world differences like different amounts of noise may have been observed, that could have been through advances in the JPEG engine, which already had advanced significantly on its way from the E-30 to the E-PL2. Maybe it could have even been due to different noise reduction settings, though.

    • Jamp

      you really have no clue about the DxO test.
      you repeat the same crap all clueless noobs write.

      take some time and read the DxO whitepapers to their test.

      and stop boring us with your cluelessness…..

      • sparedog


      • What is in the “DxO whitepapers” that could explain how leaving out fine-detail resolution in a sensor can make a meaningful all-round sensor benchmark? What is in the “DxO whitepapers” that would explain their blatantly wrong usage of the term “ISO”, with their ISO measurings not adhering to those ISO/CIPA definitions all manufacturers adhere to, and, moreover, applying a JPEG-based ISO definition to RAW? When YOU read the “whitepapers”, you can surely give us a hint? Or who is clueless here?

      • spam


  • Scott

    I am sure the only difference in the sensor is the focusing advantage

  • Olympius

    Wow, the E-P3 scored a whole four points higher than the Canon G12 – 51 to 47! Way to go Olympus! You have those compact point & shoot cameras on the run now!

    See the results for yourself:

    But seriously, as a person who’s been able to compare raw output of different cameras and sensors, I have to see that DxO is usually on the mark…no pun intendeded. By looking at the RAW file itself, many of the processing tricks of the True-Pic engines are bypassed, giving you, quite literally, the “raw” material the engine has to work with. Since the introduction of the Olympus E-30, the 12mp RAW files from Olympus and Panasonic have hardly changed one iota.

    But there is an interesting exception: the 12mp multi-aspect sensor of the GH1, which got a DxOMark score of 64, and also demonstrates a full stop improvement in dynamic range over the 12mp sensor used in the Pens, E-30 and E-620.

    Having a GH1, what DxOMark states is very true, and is very visible in the RAW files–the GH1 sensor is one of the very best ever put in a 4/3 camera, even though the JPEG’s it produces leaves a bit to be desired.

    The only people who would have problems with DxO scores, are those who have never done camera RAW file comparisons themselves.

    Now where are all those people whining and complaining that IQ means nothing unless you compare RAW files? Could they be a bit embarrassed that Olympus has actually lost ground in this area, and that the E-P3’s ‘super sensor’ has not made any gains at all in real image quality? I wonder…

    I could tell just looking at the JPEG’s DPReview provided that there was no real improvement in RAW file IQ, other than what tweaks the True-Pic engine could get away with.

    These DxO Mark scores don’t surprise me in the least, but match exactly what I was expecting.


    • Raist3d

      As an olympus 4/3 rds shooter I have been chastised for simply mentioning jpeg as a valid high quality photography option for years. Inpredicted when the e-5 came along that the amount of jpeg shooters would increase. A bottom convenient in some cases ;-) but it did effectively happen

      yes, where are the “shoot raw” guys now? You see some but for some reason seems like shooting jpeg is now o-k.

      I agree with you. That said people should shoot in whatever works
      For them.

      • Olympius

        You know, that E-5 does produce some of the nicest straight-out-of-camera JPEG’s that you could want, from ISO 200 to 1600. Nothing wrong with that, as it can be a big time saver in post-processing.

        Of course DxO could care less about a camera’s JPEG performance, and they admit as much. They only look at RAW file performance, and to know one’s surprise, the E-5’s RAW files are no better than the E-30’s, or the E-620’s.

        But what’s so interesting about the E-5, and many of the other Olympus 12mp cameras, is that the in-camera JPEGS are often better than what you can get from the RAW file. That True-Pic V+ engine is tuned very well for JPEG creation in the E-5, and it might be worth the price of that camera to have that capability. Even though I’m not much of a JPEG shooter, I’ve certainly appreciated how nice Oly JPEG’s have been when I needed to go that route.

        For a lot of people, that’s all they need.


        • Raist3d

          Oh I know the E-5 pretty well. ISO 1600 isn’t as hot as you make it sound and can even in some situations have banding. At low ISO the E-5 is indeed fantastic- if you don’t bump into a dynamic range constrained situation.

          The main issue is that the E-5 at the $1,750 MSRP and self billed as a top pro camera invites comparison with what’s around, and it doesn’t quite pass the grade with the sensor on that end- unless what you care is ISO 400 or lower shots.

    • +1

      Agree with everything you say,
      unfortunately DR being my top priority i have sold off my GH1 and all m43, i figured that since GH1 was lunched all newer m43 sensors show a decline in DR while sony’s sensor doing better and better…i’ve jumped ship and now i see why

      • WT21

        Then why hang around and post on m43 rumor sites?

        • Steve

          I don’t have a mirrorless cam but I’m seriously interested.

          Unfortunately, DR is also very important to me and DXO & Dpreview tests have pretty much told me I’ll be disappointed with a mft cam.

          I’ll be waiting on Sony to release a wide zoom and the G 16-80 it seems.

          Shame, cos I’ve always liked Oly’s jpeg quality and the 9-18 looks great.

        • Olympius

          “Then why hang around and post on m43 rumor sites?”

          Obviously frosti7 has seen the light, and wants to inform those still living in the darkness that, in fact, other cameras with better sensors than the Pens, do, in fact, deliver less noise and greater dynamic range–and usually more pixels to peep at too.

          Consider it a service, because Olympus fan boys can’t be bothered with pitting their beloved Pens against any camera that might show up it’s image quality weaknesses.


          • WT21

            Sorry, you guys are really losers. I mean that in a loving way, of course. But losers none the less. Consider this a wake up call. You now have an opportunity to go out and turn your life around, and do something useful and positive. Or you can stay as losers. I’m sure Admin loves you for the hits you generate on his site.

      • Digifan

        I’d say get f.. out of here then and preach to a different public. I don’t care what you think, I make money with, what you call, POS cameras.
        I don’t care about DxO’s testing method and results, I care about the final output. I follow review sites to gather intel through example pics. I know how to capture a scene right the first time and don’t need the latitude everybody seem to need nowadays. I shot positive so I’m use to the risk of loosing DR by exposing wrongly. Just use oldfashioned lightmeters and follow old rules.
        Like in most trades craftmanship is disapearing and it shows in these forums.

        • Ross

          “Like in most trades craftmanship is disapearing and it shows in these forums”
          That is true. The old saying, “a poor tradesman blames his tools” is true with photography too. I see some fantasic photos from older cameras with older specs & small ISO range & yet the photos are great. It is because a real photographer took the photo. It’s like todays generation that can’t do mental mathematics, because they’ve become too reliant on the calculator (well that was 20 years ago) or the computer for the calculus. Now we have people who call themselves photographers that are too reliant on the latest capabilities of the latest cameras that complain if their photos aren’t perfect down to pixel level. Large poster sized prints need to be good, but how many of us put photos up on billboards? Too much emphasis is being put on the equipment & less on knowledgable methods in getting great photos or even the art of photography.

    • Low Budget Dave

      I usually shoot in jpg, so I can’t speak to the advantages of raw. I like the output from the Olympus firmware, and all the sample images look pretty good to me. Even though the Panasonic G3 has a better sensor, the sample images I have seen do not look as good (again, to me).

      I like the output of my Nikon D3100 a little better, but the Nikon is a little bit bigger. Neither one fits in my shirt pocket, but the EP3 might fit in a belt holster.

      So Should I look at the EP3 or the GF3? I have no idea, but the XZ1 is not nearly good enough, and the NEX5 with 16mm lens is too distorted for my taste.

      For all I know, the EP3 and G3 lenses are distorted as well, and the firmware fixes it before it gets to the raw files. It doesn’t matter to me, as long as I don’t have to fix it.

      If I were good at fixing raw files, then I could just buy the GF2 and shoot in nothing but raw.

      Here is what I want from a camera:
      – fast, accurate autofocus
      – good pictures, even at low ISO.
      – light enough to carry around without a backpack.
      – cheap enough that I don’t have to miss any meals.

      The Nikon 3100 is batting .750, but the EP3 might end up even better than that.

      If DxO tells me it is a dated sensor, that doesn’t really show up on my list.

    • jim

      I so wish the E-P3 had the GH1 sensor…. multi aspect too!

      • Reto

        I love my GH1 up to ISO 400. Above that it’s barely usable for quality pictures. Banding is really bad from ISO 800. So much for I would love the GH1 sensor in a forthcoming Pen. At least Olympus could improve the colours.

  • David

    I will never understand why photographers give a damn about DxOmark. It’s like saying a Lotus Elise isn’t a great sports car because it uses a small displacement Toyota engine. Who cares, it’s about how the whole package works together. Look at photos, taken by someone much more talented than you, to decide whether a camera produces good results. And if it’s not the end results you are focused on, then I’m not sure what you think photography is all about.

    • Raist3d

      Because great photographs can be taken with any camera and if someone is looking for an increase in iso dr and noise dxo provides for a fairly objective point of reference

      I totally agree with you that the while package counts. But this is just part of that evaluation ofte whole package.

      It also makes you wonder if you are spending the cash if there are better options for your needs. Whether ultimately more a bit more dr or a lot more dr (for example) has as much weight or not is up to each individual

      On the tech progress wise it also makes one wonder why the same performance for a few years now on those ends. Of course someone may or may not care about that

    • Thomas

      It’s like the Leica aficionados who spend $6995 on an M9 and $4000 on a lens and can’t take a photo any better than someone with a Nikon D50.

  • How can u say that i have all pens , canon 5d mkII ,nikon d90 and others im a photograf not a simple person like u !!!! there’s a lot of changes in the ep3 , af,dinamic range,iso,that u dont need infrared filters , jaaa who cares ’bout iso anyway with a correr. Use of the camara!!!
    How someone can say sony nex is better , u cannot compare legendary zuiko optics against sony plastic and expensive lenses !!!!

  • Boooo!

    We already knew that. It’s the same old sensor, that was obvious from the get-go.

    Now where is Thom Hogan and what happened with the findings of his friend that saw a stop of improvement, but only on his camera?

    • Thomas

      Question is, do the photos from the new camera look any better?

      • Boooo!

        The JPEG files do. However, despite owning an Olympus camera, I only shoot raw, because I need the extra control it gives me in PP.

        • Digifan

          Sorry but than you are a crappy photog, you’ld be better of camouflageing that with a better DxO rated camera then.

          • RW

            So you are saying Booo! is a crappy photographer because he shoots RAW????

            I’ve never heard such rubbish in my life. RAW is like using a socket set to tighten bolts rather than a crescent wrench.

            If you like to shoot jpegs – good for you. Doesn’t make him a crappy photographer because he doesn’t agree with your chosen limitations.

    • I was wondering about this too. Thom said that his reliable source claimed one stop raw improvement. How is this possible if the sensor is the same?

  • KI

    you m43 user are losers anyway.. no money to buy a real camera.. cheapos…

  • fta

    What the FFFFFFUUUU?????

    E-P1 Overall score of 55
    E-P2 Overall score of 56
    E-PL1 Overall score of 54
    E-PL2 Overall score of 55

    E-P3 Overall score of 51 ??????

    Either DXO mark or Olympus is smoking some heavy weed! The E-P3 scores LOWER than ANY m43 camera, including all of Panasonics? I give up…

    • DXOmark score differences of 3 or 5 (can’t remember, around there) are indistinguishable and irrelevant – it’s explained in the white papers.

    • mountainwalker

      Usually, newer cameras have higher scores ;-) Hence, this DXOmark eventually could have gone wrong.

  • che

    ok they are officially started to talk crap. Fact is that nex is not EVIL camera (its mirrorless not EVIL) It does not have Electronic viewfinder that we know… and LCD is not equivalent to EVF. Than again scores are not that matters, although i believe that Oly didnt change anything about sensor that i suspected allready.
    Thing is i just print a few photos on A4 matte paper in a lab from a Lightroom export to 1280 resize and 72dpi that i mistakingly didnt unchecked. Photos were stunning… Where’s that leave us…in all these MP war is just fucking HIGH ISO performance. Dinamic range really… i dont know anyone who doesnt at least color correct photos in lightroom… why do they need lab stuff and electronics to see the difference… its not seeable* (i just made that word)…
    I dont want to repeat myself but for the most part thats technical is all about lenses… then about the one behind the camera…then subject…composition…post…

    • in a month or so youll get nex-7, a 3 megapixels EVF and 24mpx sensor

      • Count me in…hahahaha…(just joking…yet)…

      • Dana Curtis Kincaid

        We’ll get a nex-7, probably with a crap-all lens and an unusable interface stolen from a PS3…

    • jim

      NO an LCD is a VF you find the view using it, hence its a VF – its just not a VF at the end of a bog roll!

  • Rachnaroch

    End of story for me, sadly. I like the E-P3 very much, and the 12mm f/2 even more, but there is no way I’m spending money in a camera with a sensor performance of cameras from 7 years ago like my first DSLR, a Nikon D70. I’ll stick to my GH2 and stop worrying about Olympu’s future models. I don’t care about them any longer. I was hoping for something new at the sensor level, first the 16 MP sensor from Panasonic, then, when we knew it was to be 12 MP again, a completely new sensor from Olympus (with in-sensor infrared phase detection no-less!), then the same sensor reworked … and after seeing the first reviews and test, this all we got: same old sensor and same performance of a few years ago. Hey Olympus, if you stick to 12 MP (something I can agree) at least give us 12 BETTER MP! What about improved DR for once? Bye, bye. I hope Panasonic can keep M4/3 live because I’m not counting on you anymore.

    • Exactly my sentiments. I’ll stick to my E-PL1. I might get an E-PM1 for backup and pocket camera, but there’s no reason for me to upgrade (except perhaps 1/4000s and orientation sensor).

      Good thing it leaves me more money to invest in glass – it will work great in the next generations of MFT cameras :)

    • Dan

      The problem is people keep equating sensor age with quality. How is thr Olympus sensor somehow worse than current sensors ? My EPL1 produces crisper images than my D7000 and holds its own until ISO 800. The D7000 does better beyond ISO 800, but thats about it. Reviews state the new Sony sensor has good dynamic range, but my camera clips highlights as do the new Canon’s.

    • Olympius

      Yes, but what about that super fast auto focus??? Doesn’t that make up for the fact that they haven’t been able to move the ball forward on image quality since the E-30?

      No, of course it doesn’t. Super fast focus on a camera that has slightly worse RAW dynamic range and noise then the E-P1 is not exactly what I was looking forward to.

      They should have just used the Panasonic 16mp sensor.


      • Dana Curtis Kincaid

        And AGAIN, Mister Troll, PanaLeica can barely make enough of those sensors for their G3 and GH2. LOOK at the supply chain problems you moronic sod!

        • jim

          Should have used GH1 sensors!

        • Plenty of GH2s in Europe.

    • 1+ I will stick to my GH2, too!

    • Low Budget Dave

      I would not worry about the sensor all that much. The Leica M8 scored in the 50’s somewhere. The Samsung NX scored in the 60s. The Samsung NX looks noisy and flat, while the Leica looks crisp and classic.

  • Spartacus

    The E-P3 scored worse than the E-Pl2 and E-5 as well. DxOMark and other reviewers should indicate what firmware they tested with.

  • che

    Its not even worth of trying to figure out the scores and how they calculate them.
    but i’m interesting in one that i saw on dpreview when comparing lenses on m4/3 and a FF and others is that best overall sharpness on all m4/3 lenses is between 3.5 and 5.6 not f8.0 and also on a chart grph it does not go much higher than a half of hight for any lens and on a ff its always f8.0 and goes much higher in that range untill after f11 WHY IS THAT (sensor size or something else) also, does that apply to my manual FF lenses that i use with my PEN.

  • mountainwalker

    I’ve tried DXO’s RAW-processing software and compared it to Olympus Viewer using images taken by an E-30. For me, it seems that Olympus Viewer does a much better job. The question is whether DXO is able to read out sensor data correctly. I don’t trust them, whatsoever. DXO should stop defaming Olympus.

    • How many RAW shooters use Olympus Viewer, and how many DXO/Camera RAW/Lightroom/Lightzone/Bibble/etc.? I bet a tiny fraction use Viewer. It is Olympus’s fault for not giving enough information to third-party (second party?) RAW software developers, and calling DXOmark’s tests “defaming” reeks of fanboyism.

    • KI

      you are an complete i**ot.

      • mountainwalker

        Maybe I’m an idiot. But I’m quite happy with my Olympus cameras and many people like my photos. Maybe the sensor is not the latest generation but it is a reliable tool in the right hands.

        • I’m very happy with my E-PL1 as well. That doesn’t change the fact that your OP was pointless and clueless, and that you are changing the subject.

    • simon

      Note that DxO don’t need to perform demosaicing for their tests, so the data extraction is quite unambiguous: get a bunch of 12-bit linear numbers and analyse their noise properties. For those interested, their site does a pretty good job explaining their measurements.

      They’re not ‘defaming’ Olympus, but merely stating what their measurements show: Olympus’ (micro)4/3 sensors have not seen significant improvement since the days of the E-3 (2007). I’m glad their measurements show that this is the case, so Olympus marketing cannot simply blahblah about their ‘new’ sensors.

    • Josef

      It’s not defamation if it’s true.

      Most of DXO’s measurements just look at the signal to noise ratio in a camera’s output. They simply measure how much good information exists before the noise makes it useless. This has a direct effect on both dynamic range and high ISO noise. Higher ISOs require amplification which in turn amplifies the noise. Olympus has simply improved the camera’s ability to use noise reduction techniques more effectively in jpg files. The raw files remain the same.

      DXO is not the end-all-be-all of camera comparisons, of course, but it is a useful tool to compare the color depth and dynamic range of different sensors. The fact is that Sony’s latest sensor designs have found a way to increase signal to noise levels while still adding more pixels. Olympus continues to use a sensor that is two (or soon to be three) generations behind.

      If you enjoy your camera, don’t let the DXO numbers get in your way, but don’t get mad at them for existing.

    • Olympius

      Olympus Viewer and their other software does the best RAW to JPEG conversions of any software I’ve seen, including ANYTHING from Adobe, DxO, Phase One, etc. The only problem is that Olympus software is a real pain in the rear end to use, I can work several orders of magnitude faster in Lightroom, and get to 99% of where I want to go. No so true with Olympus software–it’s like running a race while dragging an anchor behind you.

      As for DxO, go compare a 12mp RAW file from a Nikon D700 or D90 to one taken by any 12mp Olympus camera, then come back and tell us all how you don’t trust DxO Mark.


      • jim

        I find photoshop does a better job… but by the time your up on the Oly jpg its such a small amount of improvement that it becomes compleatly a waste of time…. so I just use OOC JPG… what a time saver!

    • Raist3d

      DXo is calling it as it is. Every single rating of every Olympus camera I own by them is pretty much in the ball park. I have been also trying DXO’s converter and no, it is not that bad at all. I still think Lightroom beats it at high ISO detail, their noise reduction is ok depending how far you want to go.

      The biggest win is the camera profile conversion: you can now have Olympus color with other cameras and vice versa.

  • zigi_S

    Man the e-pl3 looks sexy with the prime.

  • EP1

    Wow $900US for the EP3 with a 3 year old sensor? Oly is the value leader in cameras!!!

  • BS Artiste

    Just wondering, can the Pens record Raw+JPEG in the same shot?

    I always set my E-30 to capture in Raw+JPEG. The Raw files are for keeps, while the JPEGs are for quick sharing and viewing if needed.

    Do the Pens have enough processing and I/O throughput to record both Raw+JPEG at a high frame rate? Also, can the Pen SD cards keep up with the write speed of CF cards. I normally use 30 MB/s, 45 Mb/s, and 60 Mb/s 16 GB and 32 GB CF cards in my E-30.

    Before the 90 Mb/s and 100 Mb/s chips were available, I probably spent $800 to $1000 on the CF chips in the last two years for some trips to the Pacific NW in America and Canada and to London. I will get some 90 Mb/s and 100 Mb/s chips as soon as I decide on my next camera. I have to pick the new camera and determine whether it is SD or CF before purchasing more memory. I travel with a laptop, but I prefer solid state over a potentially failing harddrive when bringing back photos from my long vacations.

    • dumbo

      >raw and jpeg at same time
      all the pannys can, so would be stupid if pens cant :(

      • Skrytebane

        My E-PL1 can.

    • Digifan

      All the Olympus Pen’s and DSLR’s can RAW+JPG and there JPG’s can be super fine 4% loss, A lot of camera’s don’t have jpg’s that high quality.

    • Boooo!

      “I always set my E-30 to capture in Raw+JPEG. The Raw files are for keeps, while the JPEGs are for quick sharing and viewing if needed.”

      You do know that the RAW contains an embedded 1600×1200 JPEG, don’t you? :) That’s more than good enough for me, so I never found it necessary to shoot RAW+JPEG as well.

      On the E-5, I think the JPEG is even full-size, or at least quite a bit larger than the old ones. Might be the case with PENs as well.

  • Mar

    This is how automobile tests would look like if DxO tested cars.


    DxO is meaningless…

    • :-D :-D :-D

    • LOL :) just saw it on dpreview, but im afraid that as far as it concerns the olympus sensor DXO are right on the $

    • The DxO car test is meaningless – it doesn’t show which car won!

  • Preaching To The Converted

    I love comments on this site, if dxomark had published really great results for the sensor, all the comments written here would be ‘dxo is the boss! Their results are accurate’. Funny.

  • DXO_underperformingcompany

    DXO company is unreliable and underperforming. Lens correction software is one of their main businesses. And guess what…they only have like 2 lens/body combinations per camera manufacturer. haha It’s ridiculus. Try to find your lens and camera combination for their software. You may find 1 combination if you’re lucky. (well speaking from M4/3 POV, but other combinations are lacking, and DXO had 3 years to support the fastest growing camera market m4/3.) DXO = Pathetic, underperforming company, which can’t fully support their own lens correction software. Not very respectable.

  • Calvo

    relax: What is the agenda of dxo? Selling their raw converter and other tools: they are trying to accomplish this by claiming “we are professionals” And this is why they test the sensors… hm, they are evaluating their *interpretation* of the raw file, Than they claim our interpretation is better, closer to reality and so on

    But because they cannot read out the data stream from the sensor and obviously Olympus and all the others do not hand out their documentation… this is absolutely irrelevant. Yes they have done a great job by decrypting the raw files… they are close to the manufacturers version, which is terrific. But manufacturer has the natural knowledge of his raw format

    This is why most fotos look better in the original converter than in dxo… bytheway… is anybody here using dxo? No? So do not worry… if you worry you will die, if you don’t worry you will also die… so why worry?

    Dxo does this tests to get known and to “motivate” the manufacturers to cooperate.

    • Digifan


      • sorry my inglish

        it’s a little truth there
        i compare 40D with 50D, 60D
        the evolution of the sensors is just on the paper
        Olympus is just franckly

    • -100

      Again: How many RAW shooters use Olympus Viewer, and how many DXO/Camera RAW/Lightroom/Lightzone/Bibble/etc.? I bet a tiny fraction use Viewer. Therefore it doesn’t matter if the Viewer’s results are 1% or 10% better than other converters since no one uses it.

      Besides, you do realize that DxO has tested all Olympus MFT cameras, and they all have practically the same results. This shows that the sensor is essentially the same for FOUR stinking YEARS. Using Olympus Viewer or reading Olympus marketing material does NOT change that unfortunate fact.

      DxO’s efforts in encouraging manufacturers to give more information to them and others is a commendable effort in itself. I’d much rather use LR than Canon’s or Olympus’s half-assed RAW software which lack even a basic DAM, and again, I’m not alone. Competition is good, and monopoly over RAW processing is not good for us, customers.

      There is more to a camera than sensor, but four years and questionable marketing claims do not edify.

      • I did from time to time when I was still using my crappy E-PL2 and when I wanted to reproduce the same JPG conversion the camera would have done. If you want to get the same result as from the camera the OEM’s software is your only bet.
        Though to be honest when I was still using my E-PL2 I was shooting JPG most of the time coz the RAW was not good enough. With my GH2 it’s a different story. I shoot only RAW coz RAW is worth shooting with when using the GH2.

  • I was just looking at Dxo again and reminded that the whole “labs” thing is really there to sell Dxo Analyzer to reviewers. The Analyzer software is how they get their scores, using a raw file, after taking a series of photos. You can buy their kit and do it yourself, as some reviewers have (like Popular Photo). See here:

    As for the “white papers,” they are not very well written and I’m not sure if they are much more than short presentations, not serious articles intended for scientific publication.

  • I’d just like to point out that you could have a new sensor that closely matched the old one in some areas but improved on it in other areas. I wouldn’t be so hasty in saying Olympus were lying in their claims about using a new tweaked sensor. It sounds like exactly what has happened – improved AF and processing and the same sort of noise performance.

    • There were claims of improved dynamic range, IIRC, and that’s clearly not the case.

  • Olympius

    I find it absolutely hilarious how DxO Mark is only pointing out what is exactly the case: the E-P3’s RAW files are slightly worse than those of any other Pen before it, and so many people are blaming that on DxO! What a joke!

    There’s only one corporation responsible for the poor performance of the new Pen’s sensor, and that is Olympus.

    But, golly gee whiz, those new Pens sure do focus fast! Woo-hoo!


    • wife

      Yes, but now you can get crappy pictures much faster.

    • Dana Curtis Kincaid

      ” find it absolutely hilarious how DxO Mark is only pointing out what is exactly the case: the E-P3′s RAW files are slightly worse than those of any other Pen before it,”

      Frankly, I don’t trust DxO, and I don’t like you. I’ll take real-world result, Mister Troll.

      • jim

        Its clear as day that oly has no improvement in noise (at any ISO) or DR… and that this with regards to IQ is for all intents and perpouses the same sensor.

        DXO or not…. The long anticipated IQ increase has not been delivered… yet!

      • Real-world results reflect DxO tests. I guarantee that if you actually did “real-world” dynamic range tests on, say, cityscape at night, you’d get the same DR with E-P3 and E-PL1.

        This all for RAW, JPEG is a different story.

  • Bob B.

    Ya know…it just makes ya want to come up with the money to buy an M9…now doesn’t it.

  • Edwin

    How about looking at real life photos, and forget all those technical testing?

    • samshootsall

      Love your reply!! Let’s say that you shoot a professional style portrait, print and show to your client(s). DO you think that they want to know all the technical aspect of the E-P3 or any other camera? Just long as the pictures turn out sharp and Awesome, the client(s) will love the outcome not the camera.

      • Neonart

        @Edwin & samshootsall

        Amen! If the shots look good, and the prints look good, who cares what the ratings are?

        • And that’s why I’m sticking with my E-PL1. If E-P3/etc. had better IQ, I would have considered upgrading. But since there’s no compelling reason to do so, I won’t.

          • samshootsall

            I would only upgrade as I use to SUPER FAST AF…

  • Joe

    Funny, but after all these years DxO still hasn’t figured out how to decode an ORF.

    They aren’t getting output of the sensor, or even output of the A/D converter. They’re getting an encoded file and are missing all the algorithms to decode it properly.

    It’s all hogwash.

    • So is Lightzone/Bibble/Camera Raw/etc., which is what 99.9999% of RAW shooters use.

      Your point was, again?

      • Joe

        Some RAW converters have the decoding algorithms, some don’t. Some that don’t come close… DxO is not one of those.

        Their numbers are wrong when it comes to Olympus.

  • Dana Curtis Kincaid

    DxO is… Well, kind of a sham. I’m sure that they mean well, but if they REALLY wanted to test a sensor, and not a camera package, that would be a LOT of work. Actually well-nigh impossible.

    There are different tools for different jobs, and the most important thing is, are you happy with the end result?

    Can you take pix you are happy with? If so, then you don’t need a new camera every year or two. If you don’t like the pix, then try something different.

    A 1969 Porsche 911 isn’t as “good” as a 2011 Porsche 911, but it gets the job done, and is fun. Same thing with cameras. We are all trapped in this ridiculous capitalist must buy new stuff mode, and it is killing the planet.

    • napalm

      oh yes, obsolescense :)

  • Dana Curtis Kincaid

    “For sure, this test is bad news for Olympus, because, if nothing else, it’s really kicked their marketing in the balls. Higher DR ? It seems there’s barely enough DR for a jpeg at iso1600.”

    You know, that’s kinda weird, because I shoot with an E-520 at 1600 asa all the time, and really, I don’t mind the result. What’s bad news for Oly, along with every other camera maker, is pixel-peeping internet trolls that think a picture of a camera being taken by another camera is art.

    You people… Never happy. Well, soon there will 9 billion people on the stinking planet, and you’ll have a hell of a lot more to worry about than what the new flavor of the month at Baskin Robbins is, and if you can get a digicam at all you might not have the power to run your computer at night to download pix.

    Jaysus H on a flaming cross of Cheddar Cheese!

    • Do you believe in peak oil, too?

  • Dana Curtis Kincaid

    “Besides, you do realize that DxO has tested all Olympus MFT cameras, and they all have practically the same results. This shows that the sensor is essentially the same for FOUR stinking YEARS. ”

    And maybe DxO is secretly owned by Rupert Murdoch who has a beef against Oly… Who knows for sure? I don’t.

    Taking a look at the output of the E-5, I don’t find it lacking. I probably would not find the output of an EP-3 lacking. Hell, I could make good pix with a 2MP Fuji!

    I couldn’t make it into a 24×24 foot print, but damn it anyway, I don’t NEED TO. I figure 12MP in a nice, small package like the EP-3 will be quite kewl.

    • “This photo is private.” lol

  • Pierre T.

    May be the same sensor, but not the same image processor. E5 uses Truepic V+ image processor while E-P3 uses Truepic VI image processor.

  • Dana Curtis Kincaid

    AND, real-word again, which the blog author posted and then seems to have forgotten about:

    “The PhoBlographer (Click here) compared the high ISO results of the new Olympus E-P3 with the Canon 7D and Canon 5D markII: “Based on the image samples, the Canon 5D Mk II is still way ahead of the other two cameras. However, the EP3 is almost on par with the Canon 7D’s high ISO output results. Overall, the new Olympus EP3 has been able to vastly improve the high ISO output and may finally creep up within the next iteration of the camera (or maybe even in RAW).”

    Blog author! Are you trusting DxO or real-world results?

    • FFS not this again. Comments on this blog are becoming as clueless as those on Youtube.

      PHOBGLGRER IS TESTING JPEG OUTPUT, NOT RAW – and such tests are largely invalid. And I’m being generous with the word “testing” in this case.

      DxO reflects real-world results FOR RAW shooters in the metrics they test for (eg. dynamic range and tonality). JPEG shooters can just ignore the DxO results since they are largely irrelevant to you.

    • FFS not this again. PHOBGLGRER IS TESTING JPEG OUTPUT, NOT RAW – and such tests are largely invalid. And I’m being generous with the word “testing” in this case.

      DxO reflects real-world results FOR RAW shooters in the metrics they test for (eg. dynamic range and tonality). JPEG shooters can just ignore the DxO results since they are largely irrelevant to you.

  • whocares

    Lot of sadly deluded fanboys here the sensor image quality is exactly the same as previous pens i guess when they made the “new” sensor the forgot to boost image quality.What you have is the same old image quality in a far more responsive body live with it. The DXO whines are pretty funny from a bunch of ill informed fanboys suck it up guys maybe next “new” sensor will be better lol

    • Dana Curtis Kincaid

      Ya, CaNikon troll. I am a fanboi because I like the Oly lenses, and I don’t need a big white Canon lens as long as my arm to feel like a man. Add if you are too much of a coward to even use your real name… whocares is exactly right. We don’t.

      • Olympius

        The only troll on this message board is Dana Curtis Kincaid.

        Just sayin….


      • I like my Olympus camera and will stick with it. But I agree with whocares as well.

        It’s fine to stick a four-year-old sensor in a camera, they made great improvements in other areas, mainly AF speed and variety of bodies, What’s not fine is to lie to your customers by claiming improved DR and whatever else they did.

        Olympus’s camera division in such bad shape financially it was probably the only thing they could afford. Unfortunately it will ironically mean that those looking for the ultimate IQ will move to Panasonic or Sony when the next gen sensors come out.

      • whocares

        I am an Olympus user just not a deluded one, Is Dana even a mans name lol i suspect you are much more used to handling short white lenses

  • I’ll repeat what I said at DPR. DxO ISO measurement hardly applies to m4/3 cameras and especially to Oly. They underrate the manufacturer’s ISO values, and from that everything else ensues. SNR, and DR are off by one or two stops compared to reality.

    In addition since they cannot measure the processing of the signal, I assume, they miss all the improvements Oly has been refining in the last few years, especially in denoising. That explains that DxO is usually at odd ends with reviewers.

    By now most of us have learned not to trust DxO to form judgement, but for trolls it’s manna because they have a couple of bad measurements to feed on without lifting a finger.

    • Camera manufacturer ISOs are set by marketing department, not by engineers. This very issue has been discussed in a DxO article IIRC and elsewhere. Besides, it doesn’t matter: if you look at the graphs you can see that they are lines across the entire ISO range, so you can see improvement (or lack thereof in this case) across cameras which have different ISO definitions.

      Sure they miss the improvements Oly has done. But so does Lightroom and every other third party RAW program. Since 99.99999% of RAW shooters use non-Olympus RAW converters, your point is pointless.

      And surely you are aware that improvements in JPEG performance don’t necessarily imply RAW performance.

      • RAW fetishists make me merry. There is no way one can visualise a RAW without some kind of converter. DxO sell theirs, therefore it’s like asking a pub owner if his beer is good. Ah, the innocence…

      • Joe

        99.9999% of people who believe DxO’s numbers hold water would believe the moon was made of cheese if DxO told them so.

        DxO is a waste of time, especially when it comes to RAW.

        • cowgaR

          I have no faith in DXO numbers and I’m not keen to walk with a transparent sign of their site with results shouting on people, but for me DXO results are _much_ more relevant than olympus marketing department.

          and even more relevant than your comment.

          And no, moon wasn’t made out of cheese, it is an alien ship, dull inside, and there’s Bill Gates residency on the opposite site, that’s why we stopped going there in 70ties…

          yes, and I bought their software, after testing Aperture, Bible, Capture One…am I biased? I liked their results the best, as has been said, one can interpret results from RAW in many ways.

          Does it mean DXO numbers are completely off? No, maybe 80-85% precise, is enough for me though.

        • I read it on the internet by some guy named Joe so it must be true.

          Ignorance is a bliss, eh.

          • Joe

            Unfortunately I had spent part of my life, at one time, testing light-sensors in a lab, so I can’t claim ignorance. To that point, I can’t substantiate that claim on the blogosphere with any accountability.

            Fortunately, for my point, DxO’s results are just as unsubstantiated as mine. They have introduced too much variability in their tests. First, the RAW data from any camera is not a RAW as we would like to believe. It still undergoes processing from an analog signal to digital, interpolated and extrapolated for color information from the Bayer sensor (depending), compressed and converted to raw (among other things that a company with propriety information will not let out lightly).

            Next, DxO is trying to decode that RAW information without all the knowledge of how the RAW file is created (again, that proprietary info).

            Finally, they are running “scientific” tests on their decoded samples… actually, using “scientific” is loosely applying the term. Methodical is more like it.

            Basically, DxO should be saying “These scores represent which cameras perform best with our raw converter and testing system.” That would provide a more accurate assumption of the results… but they are trying to sell something.

            • That proprietary information is not available to Lightroom, Camera Raw, etc. either. As practically no one is using Olympus Viewer, your point is moot.

              • Joe

                You’re obviously and individual who is not of scientific thoughts by nature. The fact that you don’t understand the point I’m trying to make does not make it moot, it only makes your inability to follow along more obvious. Let me break it down one more time.

                Not being privy to the proprietary information on how to decode the ORF means you have to make up your own decoding algorithms. That, in itself, introduces a huge variable in the tests.

                What that means is that their tests have a large amount of assumed error that cannot be accounted for. By error I don’t mean mistakes, I mean variability in the posted results where the actual, accurate data, could be much different.

                I’ll leave you to your own devices from here on. I suggest you get some education on the matter before you spout your drivel just because you don’t understand.

      • whocares


        You mean because you do not understand DXO or cant be bothered to read the ins and outs on their website this make them wrong lol. Like or hate DXO it is the leading image testing software designer around and their software is used by many high end companies including top review sites { who you probably use } . You bitch in every thread about them while also stating you are a jpeg user . If anyone thinks the EP3 is a genuine advance in image quality the old pink tinted specs need to come off. It is a great step forward for AF and speed of use etc and this is why I will buy it . It is pretty pathetic how you fanboys have to jump to the troll attack as there is nothing but bull in your post

  • Olympius

    “Blog author! Are you trusting DxO or real-world results?”

    Or are you just wanting to trust those who give you the review results you want to hear, and dissing those you don’t like?

    Now that’s the real question.

    But here’ a REAL WORLD comparison for you: the Canon 7D vs the Nikon D300 vs the Olympus E-P2 at ISO 200 and ISO 1600…

    Which camera image has the worst noise and the worst dynamic range at ISO 1600? It’s not the Canon 7D nor the Nikon D300.

    The D300’s dynamic range, from a ISO 1600 RAW file straight out of the camera, blows the E-P2 clear out of the water–look at the angel’s reflection on the table top. The Canon 7D picks up the reflection too, but with a bit more noise. But with the E-P2, you get a black blob where the angel’s reflection should be.

    Amazingly, when you interpret the DxO Mark scores for those same three cameras, it states that the Nikon D300 will have the best DR and least noise from a RAW file, the Canon 7D will come in second, and the E-P2 will come in a distant third, JUST LIKE STEVE HUFF SAW IN THE REAL WORLD! No doubt Steve Huff and DxO Mark are in cahoots to deceive photographers around the world, because we all know that the Olympus 12mp sensor is just as good as the Canon’s 18mp sensor.

    Just because the goofballs at “The PhoBlographer” told us that the E-P3’s file quality is ALMOST as good as that of the Canon 7D is meaningless, because review sites were saying the same thing about the E-PL2! Of course “almost” is not the same as “equal to”, and certainly not as good as “better than” — but I guess we read into these things what we want to hear, isn’t that right?


    • Thomas

      Actually Steve Huff liked the 7D file better than the Nikon even though he saw a little more noise in the Canon, and said the Olympus was basically hanging in there.

      Then he bumped up the ISO to 1600 on all cameras and didn’t even comment on the “results”.

      He also said that he’s not really into SLRs when pressed to choose the Nikon V Canon, and that he could not wait to get his M9 back; a camera known to not have stellar high ISO performance.

    • The master

      Olympius: do you swallow or spit when you give Nikon a bj?

  • Kasino

    I notice alot of people here are saying that ‘the noise is reduced in the jpg and jpgs are fine’ and ‘RAW doesn’t matter, who cares if the RAW is the same as the EP2 – the jpg is better’. RAW is extremely important – it’s the actual image off the sensor. You can’t use marketing to hide it.The rest is just software interpretations of that information. Now if you take that RAW file you have basically two options – 1. Process it on the camera. 2. Process it on the computer. A computer will always win. You can do anything the camera software can do with your computer software and more. Much more. A computer is infinitely more powerful. So as far as i can see, RAW is really all that matters. Sure, if you don’t want to process every image by hand i can see why in camera processing is important. But even then, you can save a set up in Adobe camera raw that you like and batch process all your images. If we look at the lineage, aren’t we basically talking the same GF1/EP1 sensor here? The same GF1 sensor that is used in the GF3? Go buy a second hand GF1 for half the price with the F1.7 and you’re onto a winner.

  • Bobak

    ….and no one is surprised.

  • Ben

    It’s no more fun reading the comments :(

    • +10
      seems to now be full of the bitter, pedantic, trolls and sensor/dynamic range freaks
      if this carries on……………………..

      • Ben

        It really got downhill in the last few weeks. These guys must be getting a lot of money to be spending so much of their time (and wasting ours) talking out of their *ss here…

        By the way, in the Netherlands the E-P3 is now available for pre-order at Still rather expensive at €899…

        • EP1

          Wow… and I thought the US price was high.

        • The DxO results confirmed what I thought: it’s the same sensor. Therefore I will be spending my money on lenses. I might stop by my shop in Amsterdam to play with E-PM1, as it might be a great backup camera due to its size.

          Any news on when the 12mm or 45mm will be available in NL?

      • Kasino

        So, we should blindly follow the marketing, believe that its a better camera and buy it? Seriously, it’s everyones right to know the honest truth. If its the same sensor as three generations ago, don’t we have a right to know and mention it? Or are we simply labeled as trolls the moment we question the giant marketing machines that tell us the new product is better?

    • Andyoz

      Totally agree. It seems to have been the release of the 3 new products that has changed the tone. Before that I really enjoyed all the discussions here – everyone mostly got on and there was alot of exchanges of information and ideas.
      I think that admin should ask for all arguments to go to mirrorless rumors site.

    • Yeah, 43Rumors seems to have been overrun by clueless people claiming to know something about camera technology, or trolls. I’m not sure which is worse.

    • WT21

      agreed. Perhaps admin likes his 100+ hits when he posts such controversial stuff, but sooner or later it will drive away readers

  • bedo

    dxo has “destroyed” the EP3 … Here we have controversies! who is telling the truth about this new sensor?
    is new and very good or is the same e-pl2? (worse for dxo!)
    ISO performance better or equal to the previous camera? ….
    or the hard work of improving the place around the new processor?!

  • DingieM

    This just confirms that DxOmark says absolutely NOTHING about what a camera can do.
    It is what you get on paper or on your hard drive what counts!

  • spam

    What I don’t get is why anyone would expect an Olympus designed (or tweaked) sensor produced by Panasonic would perform different from a Pansonic deisgned sensor whith the same (more or less) size and pixel count.

  • Christoph

    Sorry for OT: The more the comments section here acts as a forum, the more this forum needs an ignore function…

  • Per

    DXO are measuring things in a consistent way, but I am not sure how relevant their results are.
    According to experienced photographers, judging from what they get out of their cameras, DSLR’s normally have DR within a range of 7 – 9 steps. Meaning m4/3 should be at the lower end and “ful frames” at upper end. Phase One IQ180 is estimated at around 12 steps. (source Luminous Landscape).
    So if DR is your main priority, m4/3 is the wrong way to go! DR will improve over time but the relation between sensor sizes will remain – at best gaps may narrow.
    Those who shoot RAW can exctract a bit better DR in post processing. Using HDR is another way to improve DR. -But note, manage thees processes with care! Using them the wrong way (too extreme) and you will get bad looking images.

  • Michael Devitt

    Wow, just a few DxO values and so much hate. The sensor doesn’t create art. It’s the photographer.

    • cowgaR

      wow, just a few kW less and so much hate. The engine/car maker doesn’t create a good lap/driving experience, it’s the driver.

      //in discussion which car is better, Nissan GTR (my favorite) vs Porsche Carrera Turbo costing twice as much (imo weaker, but praised by german reviewers indeed)

      e.g. HERE, I’m buying the camera, not the photographer. I want the best for some price point. But I got your point.

  • david

    Actually it looks like they did change at least the color filters; the red channel is less selective than it was in the E-PL2. That leads to the only real difference in the DxO scores between the cameras: in the color sensitivity. Effectively the E-P3 is slightly “red colorblind” compared to the E-PL2.

  • Dana Curtis Kincaid

    “You can see a full sharpness test in my full review, but either way it should be noted that the lens is superbly sharp when the camera is shooting at lower ISO settings. For a 12MP sensor, I’m actually quite amazed. I’ve used the Nikon D700, D300s and other 12MP cameras and I haven’t been able to see such amazing sharpness even despite using the good Nikon lenses. The fact that it’s beating even the D300s even astounds me.”


    “I’ve used the Nikon D700, D300s and other 12MP cameras and I haven’t been able to see such amazing sharpness even despite using the good Nikon lenses. The fact that it’s beating even the D300s even astounds me.”

    So, “Is DxOmark destroying the “legend” that Olympus used a new 12 Megapixel sensor?” Or, does it matter what DxO says at all?!

Back To Top

By continuing to use the site, you agree to the use of cookies. more information

The cookie settings on this website are set to "allow cookies" to give you the best browsing experience possible. If you continue to use this website without changing your cookie settings or you click "Accept" below then you are consenting to this.

What are Cookies?
A cookie is a small file of letters and numbers that is stored in a temporary location on your computer to allow our website to distinguish you from other users of the website. If you don't want to accept cookies, you'll still be able to browse the site and use it for research purposes. Most web browsers have cookies enabled, but at the bottom of this page you can see how to disable cookies. Please note that cookies can't harm your computer. We don't store personally identifiable information in the cookies, but we do use encrypted information gathered from them to help provide you with a good experience when you browse our website and also allow us to improve our site. You can watch a simple video from Google to find more information about cookies.

Cookies used by our Website
The 43rumors website,, uses the following cookies for the collection of website usage statistics and to ensure that we can . These are anonymous and temporary. By using our website, you agree that we may place these types of cookies on your device.
Read how Google uses data when you use our partners' sites or apps:
Google Analytics Cookie Usage on Websites: Addthis cookies:
Disqus cookies:
Vimeo cookies:
Youtube cookies:

Disabling/Enabling Cookies
You have the ability to accept or decline cookies by modifying the settings in your browser. Please note however that by deleting our cookies or disabling future cookies you may not be able to access certain areas or features of our site. For information about how to disable cookies in your browser please visit the About Cookies website.