Long Exposure Noise - Parameter That Improve the E-M5

Discussion in 'Olympus Cameras' started by tradesmith45, Nov 20, 2015.

  1. tradesmith45

    tradesmith45 Mu-43 All-Pro

    Dec 13, 2012
    Oregon
    Thought I knew the noise performance of my E-M5 until I did some 3 min. exposures for foregrounds for aurora shots - wow was I wrong. And in camera Noise Reduction does not solve this problem. Found Brendan Davey's helpful work over hear & learned Oly m43 cameras are among the worst for long exposure noise.
    Long Exposure Sensor Testing.

    I suspect IBIS may contribute to this problem by limiting the amount of heat sink mass that can be used. Davey's tests of the Sony A7R & A7R II which has IBIS support this idea. He hasn't tested the A7S II which could provide confirmation.

    This is an important issue for me since I so love night landscapes. This is the 1st m43 issue that has caused me to seriously consider another system - Fuji.

    That's the bad news but may have some good news. Davey's work is helpful but doesn't provide enough info to know what shooting parameters might lessen the impact of long exposure noise. Night landscape shooting experience has shown me that an exposure of 2 min. @ f2 & ISO 3200 gives just adequate exposure of rocks/trees on a clear sky star light only night.

    So followed the method used by Davey - put the body cap on the E-M5 - I shot a series of captures to examine which might be best combination of ISO & exposure time. In LR, boosted all these images +4 EV & cropped the to 1300 pixels on the long side. No additional adjustments were made. Here's what I got w/ in camera Noise Reduction Off & Noise Filter @ Low (think this only effects jpgs).

    22866050880_218aff2ab3_b. 30to120.jpg by tradesmith45, on Flickr
    23135729846_8262e6c57a_b. 150to480.jpg by tradesmith45, on Flickr

    As expected longer exposure & higher ISO produces more - lots of - noise.
    (As you can see the 180 sec. images have less noise. But I'd turned the camera off to do some chores after the 150 sec. sequence so the camera was probably cooler when i started again w/ >=3 min. tests.)

    I then sorted these into a group that provided equivalent exposures around 2 min. @ ISO 3200 to see if longer time/lower ISO was better than shorter time/higher ISO. Then I ran all these through DXO Prime NR set to 80 (near the max of 100).

    As you can see w/o any PP NR, ISO 1600 @ 4 min. is the best. There seems to be a sweet spot. But w/ the use of Prime, ISO 12,800 @ 30 sec is the best. That's sort of a surprise. I'd earlier found that the combination of in camera Noise Reduction & DXO Prime makes things even worse @ ISO 3200 & 3 min. Even asked about this at DXO support. They responded that Prime was not designed to handle long exposure noise. I've found it great for exposures up to 30 sec. so maybe this will work? Have never tested night landscapes at ISO 12800 so will have to do that to confirm this. DXO can certainly get rid of noise but what happens to the image??
    22793861749_61dd10e6c8_b. Eq Exp NR Comparison.jpg by tradesmith45, on Flickr

    For these really long exposures, in camera Noise reduction is very ineffective. Davey found the same.
    23135733006_d2b9815c27_b. Camera DFS.jpg by tradesmith45, on Flickr
     
    Last edited: Nov 21, 2015
    • Informative Informative x 3
  2. tradesmith45

    tradesmith45 Mu-43 All-Pro

    Dec 13, 2012
    Oregon
    So some of you could help with this discussion. Davey has not tested any Pana cameras. Any of you willing to slap the body cap on your camera & do these tests?

    As mentioned, I shot RAW, NR Off, brightened the captures +4 & cropped to 1300 pixels. Would love to know if there is a standout long exposure noise m43 camera out there.
     
  3. tradesmith45

    tradesmith45 Mu-43 All-Pro

    Dec 13, 2012
    Oregon
    BTW, I plan to replicate these test to see if the ISO 1600 sweet spot is confirmed.
     
  4. tradesmith45

    tradesmith45 Mu-43 All-Pro

    Dec 13, 2012
    Oregon
    Forgot to mention it appears there is a bit of a noise hit when going into Extended range at ISO 6400. ISO 5000 seems much cleaner that expected for only a 1/3rd stop difference.
     
  5. barry13

    barry13 Super Moderator; Photon Wrangler

    Mar 7, 2014
    Southern California
    Barry
    Hi, just curious... approximately what was the air temperature when you did these tests?

    Barry
     
  6. 50orsohours

    50orsohours Mu-43 All-Pro

    Oct 13, 2013
    Portland Oregon

    Thanks for the info!

    On another note, is it just me, but I can not see the appeal with Fuji when you can get a SONY A7II or A7s.
     
  7. tradesmith45

    tradesmith45 Mu-43 All-Pro

    Dec 13, 2012
    Oregon
    Probably over 70. To make sure no light got into the camera, I put in my camera bag.

    But I first ran into this problem shooting Aurora at 20-25F.
     
    Last edited: Nov 21, 2015
  8. tradesmith45

    tradesmith45 Mu-43 All-Pro

    Dec 13, 2012
    Oregon
    Check Davey's tests. Fuji has less long exposure noise than any Sony. A 3 lens Fuji kit is much lighter & they have several great fast UWA & WA lenses- f2 & f1.4. But they don't have any long teles. So I'm interested in the rumored Oly fast primes.

    When I tried an A7R, I was pretty disappointed in the controls. DPR noted these problems too. Have yet to try a Fuji.
     
    Last edited: Nov 21, 2015
  9. Levster

    Levster Mu-43 Top Veteran

    I now only have a GM5, but even relatively short exposures produce terrible results. The noise is probably something to do with the electronic shutter, but I found it unusable for 30 sec exposures, let alone anything longer. I'm started to play with 3 min exposures @ 1600 ISO with my E-M5ii, so I'll post the results once I get a good night!
     
  10. Levster

    Levster Mu-43 Top Veteran

    Wow, I didn't expect there to be that much of a difference between Fuji and Olympus sensors!

    image.
     
  11. tradesmith45

    tradesmith45 Mu-43 All-Pro

    Dec 13, 2012
    Oregon
    Yes, Fuji & a couple Nikons are the best currently available right now.

    The EM5ll is thought to have the same sensor as the mark 1.
     
  12. Levster

    Levster Mu-43 Top Veteran

    This isn't due to heavy noise reduction is it? Most reviews suggest a 1 stop advantage in ISO under normal conditions, it seems crazy that there's such a marked difference with long exposures.
     
  13. tradesmith45

    tradesmith45 Mu-43 All-Pro

    Dec 13, 2012
    Oregon
    Long exposure noise is a different critter than short exposure - heat build up in the sensor. All Davey's tests used no NR.
     
  14. tradesmith45

    tradesmith45 Mu-43 All-Pro

    Dec 13, 2012
    Oregon
    The Fuji & Nikon D750 both use custom Sony sensors but are better performers than other cameras using Sony. Has to be something about how the sensor is mounted & the processors. The X-Trans is claimed to have a noise advantage because less signal processing is needed due to the lack of a comb filter. But there are several poor performing cameras w/o those too.
     
  15. Levster

    Levster Mu-43 Top Veteran

    Do you know which of the Fuji cameras share the good sensor? I've seen earlier models, such as X-M1, for only £130 second hand. If it's only models such as X-T1 then that would be outside of my budget to try. I'm hoping for a clear night tonight to try some 3 min exposures with my E-M5ii. I wouldn't switch to Fuji from mu43, but I could be tempted to switch my GM5 for a Fuji.
     
  16. tradesmith45

    tradesmith45 Mu-43 All-Pro

    Dec 13, 2012
    Oregon
    I barely understand the Fuji line of cameras so can't help. Davey tested the T1 & A1 & they performed the same. Check the Fuji boards for more details about camera differences. I'd check both sensor & processor models.
     
  17. Levster

    Levster Mu-43 Top Veteran

    I pulled the trigger and bought a Fuji X-M1, which is pretty much bottom of the range and cost £130 second hand. The difference between the RAW output at ISO1600 @ 180s with my E-M5ii is vast, but so far I have only compared darks and I haven't had a chance to try it on my scope. This is the difference between the two RAW files, with no exposure adjustment:

    180s_ISO1600_E-M5_Vs_X-M1.
    I've put both files on Dropbox, with the E-M5ii here:
    Dropbox - PB260001-1.ORF
    and the X-M1 here:
    Dropbox - DSCF0006-6.RAF
     
    • Like Like x 1
  18. tradesmith45

    tradesmith45 Mu-43 All-Pro

    Dec 13, 2012
    Oregon

    Wow, I can't thank you enough for this!

    Here are your results+mine for ISO 1600 & 180 Sec exposures @ +4 EV & cropped to 1300 pix.
    22705693094_04503318a2_b. DSCF0006-6-Edit.jpg by tradesmith45, on Flickr

    There has been speculation about whether the E-M5 & E-M5II share the same sensor. Looks like not. Clearly, this is not a censor size issue but rather something to do with mounting/heat sinks.

    For me this is a very big deal. I've been planning to buy a Mark II to get focus stacking/bracketing & selling????? my Mark I. Your result almost guarantee I won't do that. Been considering a move to Fuji for landscape work. These results make that much more likely!! But I'd have to figure out what to do for birds & wildlife because Fuji doesn't have anything for that.
     
  19. Levster

    Levster Mu-43 Top Veteran

    I really like my E-M5ii for day to day use, so couldn't see me selling it. I've also picked up a Fuji 35mm f/1.4 to try this weekend. I could maybe see myself going with Fuji for wider shots and sticking with Olympus for long shots. Or if I don't get on with the lenses then I'll keep the Fuji for Astro shots only. The Fuji lenses are pretty expensive, with their only 90mm equivalent being £600.
     
  20. tradesmith45

    tradesmith45 Mu-43 All-Pro

    Dec 13, 2012
    Oregon
    Yup, expensive lenses & very limited selection - almost no 3rd party support.

    Also been looking at IR hot spots issues. The XF 10-24mm isn't usable, the Samyang 12mm hot spots some & no reports on the16mm f1.4. So may not be any UWA options for IR except the XF 14mm.

    For starry landscapes, the XF 16mm has bad coma until f2.8 & there are mixed reports on the Samyang 12mm.

    So not only expensive lenses but not so flexible.