EM1 Long Exposure Noise Problem? – Update

You may or may not have been following an interesting debate that has been hitting the various forums on the internet over the last few days.  Users (or a user) have been complaining that the EM1 shows too much noise when taking a long exposure shot of 60 seconds or more.  The user who initially stated this also mentioned that they had turned off their noise reduction filter which on a Olympus camera is not the same as a noise filter in that the noise reduction filter takes a frame of equal length of the initial 60 second frame to map out any pixels that may need subtracted from the original photo to try to eliminate the noise that would usually occur at long exposures like these.

The argument is that long exposure shots of 60 seconds or more should not have this much noise when the noise reduction filter is switched off.  Below you’ll see in my own tests I was able to reproduce the exact same issue shooting with the EM1 for 60 seconds at an ISO of 200.  Theses are straight out of the camera RAW files converted to JPEG in order for me to upload them for you to see with your own two peepers.

OLYMPUS DIGITAL CAMERA

 

Straight out of the camera with Noise Reduction set to auto (remember that noise reduction is not the same as a noise filter which could degrade the image quality slightly)OLYMPUS DIGITAL CAMERACrop of the above photograph with the noise reduction filter set to auto

OLYMPUS DIGITAL CAMERA

Noise Reduction set to off

OLYMPUS DIGITAL CAMERACrop of above image with noise reduction off and no post processing (read zero noise filter applied at all)

Well?  The user was right, there is noticeable noise when the noise reduction setting is off.  The amount of noise present is no surprise to me or probably any one who has done extensive long exposure photography before.  The noise is easily corrected in most post processing software of your choice.

So if this is something that can be corrected in post processing why is everyone getting their undies in a bunch?  Good question.  Originally the person who brought this situation to light mentioned that he was only concerned because he does a lot of fireworks shots and he does not have the time to wait around for the dark frame that is taken immediately after the 60 second shot to subtract the hot pixels and what not and reduce the noise…  Good argument however 60 seconds for a fireworks shot seems awfully long if you ask me.  Perhaps adjust your aperture to reduce the time.  I’ve found that 30 seconds is adequate with an aperture of f/11, anything longer than that and the fireworks end up over exposed.  Though every situation is different and I’m sure there are loads of great 60 second firework shots out there.

Don’t have the time to bother with shooting a dark frame and then doing a little post processing magic?  Why not just turn your ISO up a bit and reduce your shutter speed?  The EM1 handles higher ISO like a champ.

I guess I’m trying to say that 99.9% of people taking 60 second or longer photographs will most likely leave the Noise Reduction on auto and ISO low therefore making this a non issue.  For those of you that do long exposure as well as high ISO astrophotography the EM1 will not please you.  If you’re passionate about landscape astrophotography or star trail photography and you’ve already made the plunge you can sell your EM1 on Amazon or eBay… you’ll probably get close to what you paid for it since it’s in such high demand.

Update

Here is a crop of the same image from above that was taken without the Noise Reduction filter on.  It literally took 10 seconds to eliminate some of the noise from this image using Lightroom 5.2.

OLYMPUS DIGITAL CAMERA

Updated shot with lens cap after version 1.1 update to EM1… ISO 3200 for 30 seconds with lens cap.  Unprocessed straight out of camera raw file converted to JPEG to upload to the site.

OLYMPUS DIGITAL CAMERA

More from John Barbiaux
The Camera You Use Always Matters
Gone are the days where a photographer could kick up dust, hang...
Read More
20 replies on “EM1 Long Exposure Noise Problem? – Update”
  1. says: Jeromy

    Your fireworks argument is completely wrong, sorry. Whether 60s or 30s or 5s: Shooting with dark frames results in 50 % effective shooting time and risking to miss a great shot – regardless of the time of every single shot. And by the way, bracketing in low light situations doesn’t allow automatic dark frames. Quite a common scenario for night/low light shooters.

    There might be people out there who don’t want to investet time and hassle having to play around with manually made dark frames and post processing if they could have much cleaner images with a E-M5 level sensor.

    In my view you try to psychologically rescue the E-M1. But fact is this cam is not suitable for enthusiastic low light / long exposure / bracketing shooters – that is if the first reviews concerning hot pixels turn out to be correct.

    1. says: John Barbiaux

      Thanks for your feedback but you don’t need to shoot a dark frame after every shot… Just take a couple and use them in post processing. It doesn’t take very long to process… The time you spent writing your comment I could have finished cleaning the noise out of your photo for you.

      Some people may use bracketing but it’s far from common and typically used for HDR or as insurance if you’re not too good at getting the setting just right yet. Don’t worry it will come with time.

      The article is written to put your opinion into perspective, you over generalize when you assume everyone needs to shoot long exposure and fireworks the exact same way you do. I’d actually argue you are over complicating both. Try shooting in RAW as opposed to bracketing, you’d be surprised at the latitude you’ll get in regards to exposure adjustments.

      As for fireworks if you don’t believe me that you don’t need a minute long exposure and bracketing then perhaps you’d trust this guy: https://photography.nationalgeographic.com/photography/photo-tips/photographing-fireworks/

      Or, if you’d like to really get your point across then I’d like to invite you to use the Submit an Article link in the main menu at the top of the page and do a detailed tutorial with images you’ve taken using the settings you recommend and help others learn more about your specific process.

      I really do appreciate your feedback and hope that you’ll consider sharing your process with us, take care.

      1. says: Anonymous

        >> you don’t need to shoot a dark frame after every shot… Just take a couple and use them in post processing. It doesn’t take very long to process… The time you spent writing your comment I could have finished cleaning the noise out of your photo for you. <<

        Could you very briefly outline your method of post processing using manually taken dark frames? Many thanks in advance!

      2. says: John Barbiaux

        Hi, I am currently working on a star trail tutorial where I will go over this in a fair amount of detail for you (using the EM1 none the less) so you should bookmark the site. In the mean time here are a couple resources for you: https://starcircleacademy.com/2012/10/darkframes/
        and for an alternative method of light trail photography with dark frame subtraction (probably the “work smarter not harder” method)
        https://www.brianchapmanphotography.com/blog/2009/04/15/dark-frame-subtraction/

        Remember that there is generally no single “right” way to do any one thing in photography, if you talk to one photographer who says they do two 30 minute exposures for light trails another will argue that stacking more exposures at 30 seconds apiece is the best way. Neither is wrong, both can work.

        Also, if you’d like to try taking a series of long exposure shots and some dark frames here and there and then combine it all check out this software: https://www.markus-enzweiler.de/StarStaX/StarStaX.html (it’s free)

  2. says: Serge

    Thanks a lot for your article!

    My biggest concern now is not as much of a noise, but to answer the following questions first of all:

    1. Is it a common problem with all E-M1 units or found only in first October-November shipments (cannot tell for sure till new shipments arrive, right :o)?)?
    2. Can we expect the issue to be fixed or it’s here forever because of the way E-M1 has been engineered to implement a new Dual-Fast AF sensor and with AA filter removed?

    These questions have not been answered yet by Olympus.

    Why is it important for me to have the answer? Because depending on the answer I can make a decision whether to replace my unit with another one (in another batch), send it back to the shop for a full refund, or keep it knowing there is nothing can be done about the sensor noise issue / trade-off.

    One more thing. I guess, it’s very important to have quality images with NR turned OFF in these situations:

    Fireworks sequential shooting
    Night shooting in Continuous mode with tracking at high ISOs
    Astro-photography
    Night HDRs

    Do you still think that a Pro and Flagship camera (claimed by Olympus as delivering the best ever image quality in M43 world) can allow such a trade-off?

  3. says: Serge

    Sorry, John there is one very important thing I missed in my previous post.

    The debate around unacceptable noise on E-M1 is heated up not as much by the sensor quality of E-M1 by itself, but by comparing it to the performance of E-M5’s one, that shows much better results in all NR OFF and even NR ON (!) situations.

    There is a whole bunch of practical evidence of E-M5 better performance in this zip file:

    https://bit.ly/1aH3j0v

    Also refer to evidence given here:

    https://www.mu-43.com/showthread.php?t=53621thread

    No-one seems to understand yet why E-M5 performs better! My guess is that implementation of a new Dual AF sensor with no Low Pass filter required this trade-off on E-M1. I can live with it, but I want clarity of proof from manufactures that this trade-of is an inevitable “evil” for a Dual Fast new sensor.

    Cheers!

  4. says: Andrei

    This a problem for all cameras. Just tested my D60 @ ISO 1600, 30sec and f22 (no NR) and the result is quite the same (a bit better because I guess the bigger sensor). I will try this with my father’s D90.

    For me this is not an issue because when I shot night photography I don’t shoot in complete darkness (I shot city lights, cars etc)

    Next spring I will get the E-M1 with the 12-40 f2.8 and the 60mm f2.8 macro. I will build a new system based on the E-M1 in the future because for me the advantages of MFT are way bigger than a system based on the Nikon D610 for example or the Pentax K-3 (other cameras I’ve considered).

  5. says: terry

    The issue is, it appears the older camera, the E-M5 handles this noise better.

    New camera performing worse than old camera with respect to long exposures.

  6. says: Serge

    Thanks for all your feedback, John!

    Very, very helpful!

    However, the problem, let me iterate, with E-M1 is as simple as that:

    Why E-M1 performs much worse than E-M1 when shooting in low light situations with NR OFF (i.e without dark-frame subtraction), while the Olympus E-M1 PR hype was stating that the sensor has been improved and is best in its class?

    Cheers!

    1. says: John Barbiaux

      Good point, that I can’t answer but only say that hopefully Olympus will address the issue. In the meantime for those of us with the EM1 there is a workaround. Is it a little more work? Sure, but that’s why it’s called a workaround ;).

      I certainly understand folks frustrations, you want your camera to be a tool you can use for every situation and certainly would like a camera considered a successor (by some) to perform better than the camera it’s succeeding right?

      The best solution I can give anyone and the same solution you’d here from many a pro photographers with much more tenure than myself is this: don’t sweat it, sell the camera and buy the one that fits your style/needs the most. Makes sense right? If your initial reaction to that is “but I want THIS camera” then you have to reassess whether your interest lies more in the gear you use or the art of photography.

      I’ll keep the site updated if I hear of any solution in the future.

      1. says: Serge

        Thank you, Jonh!

        You nailed it down by reading my thoughts strait out of my head :o)!

        I still hope it’s October-November batch that misbehaves. However, this yet to be investigated and confirmed by my Olympus provider. This hope still keeps my unit in my grabs :o).

      2. says: John Barbiaux

        Haha, anytime. I trust they will come up with something. Reminds me of the slight recall the em5 had bc of cracked LCD where one of the tiny screws held it in. Mine was one of the culprits.

  7. Thanks for your effort in writing up about it, however I think you’re underestimating an issue with this camera that will turn out to be a real problem for a whole group of users. The camera is pretty much pitched towards long exposure and astrophotography, after all their global E-M1 site clearly points that out in the ‘Gear’ section page 7. https://cameras.olympus.com/omd/gear#

    On the contrary, in reality, with the current firmware and hardware (and according to their tech support will be no upgrade or fix for the discussed issue) the E-M1 is only to some degree capable of shooting long exposures. That is when keeping ISO down and exposure times as short as possible while keeping NR on Auto or ON at all times, like Olympus recommends for this camera. On the other hand, the camera was specifically designed with extreme low light photography in mind, evident by looking at some small but smart features like the dimming of the LCD/EVF when shooting in LiveTime, possibility of setting up the camera for up to 30 minute in-camera timed exposure without using cable remote, and in the LiveTime settings you’re encouraged to set the ISO as high as up to ISO1600.

    While these LiveTime settings (30 min exposures and ability to use ISO1600 or even higher in BULB for astrophotography) was fully feasible option with the E-M5, now with the E-M1 you would be totally SHOCKED at the results from ISO3200 at 30 seconds, and believe me there’s no way you would get away with that in post. All the while the trusty old E-M5 continue to impress with fully usable results at the exact same settings. The Panasonic GH3, which I’m using a lot, has virtually the same sensor as the E-M5 and I uploaded two sample JPEGs where the two are exposed at 60 seconds, ISO3200 and using no DFS (LNR) or noise filtering (all set to minimum or OFF). I also throwed in an E-M1 60 sec exposure at ISO200 with the same settings for reference.

    On a side note, I noticed a few more quirks with the E-M1 during the shooting of the below test images along with the GH3. While the GH3 viewfinder showed a perfectly usable image, with a quite correct Auto WB interpretation and exposure estimation (by taking resort to slower refresh rates) the E-M1 viewfinder was nearly impossible to use for focusing at such low light, and the Auto WB setting was severly off, I also tried setting WB manually but there’s something strange going on there. Results with the E-M1 was very different also between LiveTime and BULB and at different apertures of the same low-light scene, whereas the GH3 produced nearly identical WB and image at all settings. In my memory, the E-M5 produced very similar results as the GH3, I have been using the two of them side by side for a year, but I decided to let the E-M5 go for funding the new E-M1… Anyways… on to the samples:

    Olympus E-M1, 60 sec at ISO200, NR OFF:
    https://www.dropbox.com/s/04kca6mool789c1/E-M1_60sec_ISO200_LNRoff.JPG

    Panasonic GH3, 60 sec at ISO3200, NR OFF:
    https://www.dropbox.com/s/i1l5kzw5aw1xyt4/GH3_60sec_ISO3200_LNRoff.JPG

    Olympus E-M1, 60 sec at ISO3200, NR OFF:
    https://www.dropbox.com/s/k651gnb21htuck4/E-M1_60sec_ISO3200LNRoff.JPG

    Olympus E-M1, 60 sec at ISO3200 – NR AUTO:
    https://www.dropbox.com/s/sgmuv7plrxee81k/E-M1_60sec_ISO3200_LNRon.JPG

    These are the results, see for yourself which camera you would pick for your long exposure and/or astrophotography needs, of which a lot of buyers had in mind when buying the E-M5. I work in a professional camera store and I meet customers with interest in the MFT-system every day. I would not recommend buying this camera if long exposure work or mainly astrophotography would be interesting to you or something you may pursue in the near future. Save some bucks and buy the E-M5, E-P5 or the Panasonic GH3, or wait a few months and follow these discussions and see what Olympus comes up with, they might have to change their minds on this one.

    Regards,
    Martin

    (Landscape and aerial photographer based in Sweden)
    http://www.martinwallgren.com

  8. says: Fiacre

    Please, don’t say that noise filters (in camera or post precessing) can avoid the problem. They can mask it, but these filter always attack details and textures. A good camera with this price must have a good result from the begening, at least as good as an E-PL5 or E-M5, older and less expensives.

    “I guess I’m trying to say that 99.9% of people taking 60 second or longer photographs will most likely leave the Noise Reduction on auto therefore making this a non issue. For the rest of you, shoot a dark frame and you can subtract most of the noise out in post processing in no time at all so worry not.”

    Are you serious ?
    Do you know it is a 1400 euros camera ?
    Firstly, lot of people using long exposures can’t use dark frame substitution (Timelapse, fireworks, etc…). And the obligation to take dark frame at same temperature of each speed used, and the use of a specific software (Lightroom can’t do that) for each frame is a silly solution for a “pro” camera. Incredible.

    Secondly, even with dark frame substitution by the camera, try to get back some information from the shadows (RAW or jpg) from a long exposure at same temperature from E-M5 and E-M1 (there are plenty of samples on the web…). You will see the noise problem with the E-M1 (or “poor performance”, as you want). THE PROBLEM IS STILL THERE EVEN WITH DARK FRAME SUBSTITUTION. Silly story for a 1400 euros camera.

    As i do a lot of long exposure, i will avoid this camera until the problem is solved. But i’m afraid it won’t be only a software fixed by firmware, as it looks like a termal problem with the sensor, it could by heated by an electronic board (or PDAF sensors ?). That would explain why an older camera with nearly the same sensor can perform better for long exposure. My message is not to attack Oly, as i’m a king of Oly fanboy but just to say the problem is real and not solved by noise filter or dark frame subsitution. 🙁

    1. says: John Barbiaux

      I am serious! It sounds like you haven’t had a chance to use the em1?! I’d recommend at least putting the camera through its paces before jumping to conclusions. Do you shoot raw? Then you will need to do some noise reduction. I’m sorry you feel it degrades your images too much but I’ve never had an issue with it. The sample on the site that I took 10 seconds to adjust on Lightroom didn’t lose too much detail in my opinion (imagine if I’d have taken my time with it). I can’t give you any info on firework photos bc I have not tried that with the camera as it’s normally a once a year thing here in the US and not really of interest to myself (not to say it’s not great).
      Use noise reduction, don’t use noise reduction, I don’t care… Just go use the camera you have to take the photos you can and enjoy.

  9. says: Chris Smith

    OK. My two cents.

    From doing a lot of reading and using a rented E-M1 over the weekend, I think that you need to change your position on this. This is a real issue and this camera should not be recommended for long exposure photographers. Other than that, it seems to be an amazing, revolutionary, camera.

    Even in your last image where you reduced noise in Lightroom, I can see little pink fireworks in there. It’s up to the photographer whether this is acceptable or not.

    The bottom line seems to be that if you can use dark frame subtraction on the camera then you’re going to be OK. But if you’re shooting time-lapse, star trails, fireworks, etc., this is not the camera for you.

  10. I totally agree with Chris Smith. And I’m still waiting for a response from you John regarding my previous comment. What do you think about the high ISO 60 sec exposures I linked to? Still piece of cake to have them cleaned up in post? What are your thoughts on the WB issue at long exposures?

    The whole point is that we’re talking about a year 2013 flagship camera, aimed at PRO photographers, some of which used the E-M5 as a step stone into the realms of MFT along with a second system, ready to swap out their heavy gear bag. I remember very well how my D300 performed at star trails and extreme exposures. And still it had some trouble with amp glow, not to mention the D80, they both outperformed the E-M1 at anything above 60 seconds NR OFF. The E-M5 is in another league altogether.

    I own the new camera, been shooting with it for a while now, bought it for my own dear money. I love the new features, the handling, the huge highres EVF, the robustness, the WiFi connectivity, improved video (but miles apart from the GH3) oh and the improved IBIS of course. But I came from the E-M5 and while the improvements are not life-changing, the noise issue is. I would say that the signal-to-noise ratio is somewhere near 4 full F-stops worse than the E-M5/GH3 at 60 seconds ISO3200. That is when NR is ON with the E-M1 and OFF on the E-M5/GH3. That’s like going from ISO200 to ISO3200 in one go. Although those figures being on the extreme side of exposure, in landscape astrophotography it is common sense.

    The E-M1 really is a masterpiece of a daylight camera, but honestly how many of today’s flagship cameras would pull off great sales with that advertisement? It’s unfortunate that Olympus have decided to advertise the camera as a great astro/low-light camera, that’s where the main problem is. It was designed from the ground up as a great all-round performer with the icing of the cake for long exposure work — LiveTime. But they smoked the sensor.

    If you were working in a professional camera store like I am, you’d be amazed at how sensitive the regular customer may be, or even the aware beginner, about things that may pose a threat to their full creativity and money investment. Then we have the pro customer who needs to make sure that the camera will perform in all those situations where he/she many find himself with the camera, in order to make money for his living!

    I find the greatest problem with the E-M1 is that you cannot trust it. Will it deliver? Where’s the turning point? The sensor issue is very elusive, exponential in its nature. I lost inspiration to take it out at night altogether because of this. Yesterday I also figured the new sensor has more issues than we see on the surface. https://www.flickr.com/photos/cablefreak/11160180444/

    All manufacturers have had their share of problems. Nikon service is still receiving lots of D600’s with oil smudge on the sensor. D800 had issues with AF and it still surfaces on the service desk. But those are repairable problems, we had no real problems selling the D600 nevertheless. But with the E-M1 it’s different. It gets raving reviews all over the world, but the dirty little secret will reveal itself in matter of time. And Olympus officially have no solution but to use NR at all times.

    Now, since you certainly have great skills at battling noise in post, I give you credit for that. Would you mind showing us all how to clean up an ISO3200, 20 sec exposure (general setting for a Milky Way above landscape shot) so that it matches the rest of the competition in terms of IQ and printability? You can use NR ON if you like, or…. you’re going to need to. 🙂

    Cheers!

    1. says: John Barbiaux

      Martin, I’m not sure why you need me if you have the EM1. Give it a try and write up an article about the Em1 and it’s long exposure with high ISO issues and submit it to Submit an Article for review. Great response by the way, I appreciate your passion.

      Also, I’d highly recommend a different camera!! If you find that your camera isn’t delivering the quality you expect in an area of photography you’re obviously passionate about then why keep trying? I am currently working with the Nikon Df and can assure you it’s got zero long exposure issues and really excels at the higher ISO that you’d use for astrophotography.

      I would be happy to show you the processing for landscape astrophotography photos and if you bookmark the site you’ll be emailed when the tutorial is finished.

      Enjoy.

  11. says: Joe

    Anyone know if this has been resolved at all yet? I’m making a new camera purchase soon. Landscape astrophotography is my #1 priority, usually around 30 second exposures at high ISO, but I need something lightweight so don’t want to go with something like the full frame Canon 6D since it is just an enormous weight penalty (I already have a 7D anyway). Should I still just pull the trigger on the EM5? Or is the EM1 any better these days?

    Thanks for the info, I learned a lot from this page.

    1. says: John Barbiaux

      Joe, this is just my two cents but I had similar issues when searching for my primary camera (I review lots of gear but I have one camera that is my workhorse). Unfortunately, the EM1 is not suitable for landscape astrophotography at high ISO (think 30 seconds with an ISO of 3200 or 6400), the sensor just can’t do it. The EM5 is capable but again, you won’t get the same results you would from a full frame camera. The solution for me was the Nikon Df, full frame and small. The price is tough to swallow but it’s the first camera that I’ve used that handles everything I throw at it and then some. Here is a link with a couple landscape astrophotography shots (one at ISO 5000 and another at ISO 10000) https://photolisticlife.com/2014/04/14/3-months-with-the-nikon-df/

      So the short answer to your question is I’d skip the EM1, the EM5 would work, a full frame camera is going to yield the best results.

Comments are closed.