October 22, 2014, 09:19:09 AM

Author Topic: Opinions on 5D III Noise Reduction for long exposures. To use or not to use?  (Read 2925 times)

GoodVendettaPhotography

  • PowerShot G1 X II
  • ***
  • Posts: 44
    • View Profile
Just curious to see who uses NR for long exposures on the mark iii, and what are the benefits and drawbacks in your findings! Thanks!
5D Mark III | 7D | 17mm TS-E f/4 L | 24-70mm f/2.8 L II | 70-200mm f/2.8 L II IS | 50mm f/1.2 L | 100mm f/2.8 Macro | 135mm f/2 L | ST-E3 | 600 ex (x3) | 2x TC III

canon rumors FORUM


Pi

  • 1D Mark IV
  • ******
  • Posts: 937
    • View Profile
    • Math and Photography
It subtracts a black frame for long exposures. This is very effective for removing hot pixels. It has a downside: in actually increases the read noise. The reason is that the latter is random (even the banding is, I believe), and subtracting it (not different from adding  in this case) actually increases it by sqrt(2). You can try it for yourself.

LR is very good with masking hot pixels, and I am not using LENR anymore. But for really long exposures, it might still be helpful.

emag

  • Canon 70D
  • ****
  • Posts: 319
    • View Profile
A long exposure noise reduction method for astrophotography is the use of a 'dark frame'.  In principal it's similar to LENR but the practice is different.  A number of images are captured with the camera lens/telescope cover in place (i.e. - dark frames), I typically take them when I'm finished with my imaging and am putting my equipment up.  The exposure length is the same as that used for the images ('light' frames).  The dark frames are averaged in post and a master dark frame created which is then subtracted from each light frame prior to stacking the lights.  Ideally, the darks are captured at the same temperature the lights are captured at.  I suppose it would be possible to create a library of master darks at various temperature ranges, but the sometimes (or usually?) random nature of the dark frame noise might preclude that.  In addition to the dark frames, a series of bias frames are also captured, similarly averaged and subtracted.  These are dark frames at short exposure taken to create a master bias frame that hopefully captures the readout noise which is also subracted from the light frames prior to stacking.  With the D60 years ago, Canon was first out the gate with a low noise sensor capable of astrophotography....the reason I and many others adopted Canon DSLRs for astro.

Reason I mention all this is that if you're contemplating a lot of long exposure work, you may find these techniques useful.  For exposures of a few seconds, LENR is probably more practical.  My astro work involves exposures of 90 seconds to 10 minutes, LENR is just not a good option at that point.

RLPhoto

  • Canon EF 300mm f/2.8L IS II
  • *******
  • Posts: 3507
  • Gear doesn't matter, Just a Matter of Convenience.
    • View Profile
    • My Portfolio
I don't use it. It doubles the exposure time and with LR, its easy to fix noise and hot pixels.

Mt Spokane Photography

  • EF 50mm F 0.7 IS
  • *********
  • Posts: 8864
    • View Profile
Its really a feature for Astrophotography or for still scenes.  Using NR for Astrophotography would probably also remove some of the stars, so long term noise reduction seems like the best choice.
 
Even for ordinary stills, NR blurs detail, so its a high resolution option.

oppfinnarn

  • SX60 HS
  • **
  • Posts: 1
    • View Profile
    • Lazy Art
I prefer to use 3rd party software for NR, like Dfine, and besides, taking pictures with long exposures is enough waiting without having to double the time between the actual exposures you do "for real"..

alejmr

  • Power Shot G7X
  • **
  • Posts: 10
    • View Profile
I personally don't use it.... LR actually is very good in removing those hot pxels.... You'll be saving a very good amount of time while on the field.
Cheers!

canon rumors FORUM


polarhannes

  • Rebel SL1
  • ***
  • Posts: 91
    • View Profile
I prefer to use 3rd party software for NR, like Dfine, and besides, taking pictures with long exposures is enough waiting without having to double the time between the actual exposures you do "for real"..

While I agree that it is usually more convenient to just use NR in PP, I don't agree on the 'just wait a bit longer' part. If you expose a picture of the stars longer than 30 seconds (at least to my experience) , star movement (or rather earth) can be seen which is not always intended. Of course this is just a special case.
"The camera is an instrument that teaches people how to see without a camera." - Dorothea Lange

niteclicks

  • Rebel SL1
  • ***
  • Posts: 98
    • View Profile
I actually out of curiosity ran a few tests on this last week after realizing my 5DIII had way more hot pixels than I cared to deal with. I was using iso 4000 temperature was about 85 deg. and 30sec exposures. I did not see a real difference, some ,but most of the hot pixels were still there , I did not try the more aggressive settings. I tried just taking darks with in camera noise turned on and again there was a dimming of the hot pixels but still very much visible, so I am not sure how effective it is anyway.

comsense

  • Rebel T5i
  • ****
  • Posts: 106
    • View Profile
Short answer: switch it off
For those still persistent about fixing it post by software, pls read the posts by Pi and emag again. While you don't need this feature except for certain long exposures, you cant FIX those one in post unless you have captured dark (= no light to sensor) images with identical exposure conditions (read emag). And I agree with emag that collecting dark images yourself gives you more flexibility in treating the noise.  Its like you can't do HDR post unless you have collected images with varying exposures and collecting images yourself is better than in-camera HDR. Just to add for people who are not trying to understand before shooting a reply; there are still thesis being written for finding most efficient way of background subtraction (which includes dark current/read noise) because its very critical to quantitative imaging.
« Last Edit: September 04, 2013, 05:01:28 PM by comsense »

J.R.

  • 1D X
  • *******
  • Posts: 1512
    • View Profile
Re: Opinions on 5D III Noise Reduction for long exposures. To use or not to use?
« Reply #10 on: September 04, 2013, 09:23:35 PM »
I actually out of curiosity ran a few tests on this last week after realizing my 5DIII had way more hot pixels than I cared to deal with. I was using iso 4000 temperature was about 85 deg. and 30sec exposures. I did not see a real difference, some ,but most of the hot pixels were still there , I did not try the more aggressive settings. I tried just taking darks with in camera noise turned on and again there was a dimming of the hot pixels but still very much visible, so I am not sure how effective it is anyway.

Did you try to remap the hot / stuck pixels?
5D3, 6D, 600D, RX100
16-35L, 24-70L II, 70-200L II, 100-400L, 50L, 85L II, 135L, 24TSE, 40, 100 macro, 18-55 II, 55-250 II, 600RT x 4
I come here to learn something new, not to learn how bad my gear is - I know that already ;-)!

niteclicks

  • Rebel SL1
  • ***
  • Posts: 98
    • View Profile
Re: Opinions on 5D III Noise Reduction for long exposures. To use or not to use?
« Reply #11 on: September 05, 2013, 10:28:36 AM »
Did you try to remap the hot / stuck pixels?

Yes I did. What really stinks is the first one I had that was stolen only had maybe ten now I have 50+ and that's at iso 400 and 10 sec.
« Last Edit: September 05, 2013, 01:07:45 PM by niteclicks »

canon rumors FORUM

Re: Opinions on 5D III Noise Reduction for long exposures. To use or not to use?
« Reply #11 on: September 05, 2013, 10:28:36 AM »