P3 PL3 PM1: dark situations settings and strategies

Discussion in 'Olympus Cameras' started by pheaukus, Jul 2, 2012.

  1. pheaukus

    pheaukus Mu-43 Regular

    178
    Jun 22, 2012
    Hi,

    I am new to the PEN PL3 and wonder, what are your strategies for getting the best images (or image quality) out of dark situations?

    So far I tried the following:

    • Bright
      Use ISO up to 1600, NR off, IS on if needed
      (even though the pain begins around ISO 800)
    • Darker, NR on
      ISO up to 3200, apply low NR and sharpening
    • Darker, B+W
      Switch to Monochrome, ISO up to 3200
    • Darkest, Grainy B+W
      ISO up to 6400, Grainy B+W art filter
    I am not very happy with the outcomes of the latter three scenarios.
    I do understand that they are good for different types of shooting and and that I have to find out how to use each of them best.
    Still, I hope that with little tricks one can push the limits of each scenario.


    Questions:
    1. What are your camera settings for dealing with low light?
    2. What are the effects of over- or underexposing on noise, detail, shadows and highlights?
    3. Are there ways to change how the camera deals with its dynamic range?
    4. What else can be done?
    5. How do these settings affect your shooting?


    Answers to above questions, so far:
    1. Camera settings for dealing with low light
      A: Beyond what can be read in the manual, there are no magic tricks. Much more can be achieved by improved use (see answers to 2, 3 and 4)

    2. Effects of over- or underexposing on noise, detail, shadows and highlights
      A: there is more information in the highlights, so your goal is to expose as far to the right as possible if you want maximum headroom (ETTR)
      (Explanation: The sensor is a linear device that tries to capture 'brightness', a logarithmic distribution. This discrepancy is reflected in the levels of brightness stored in RAW. A sensor covering a DR of 5 F-stops would have 128 discernable levels of brightness in the darkest F-stop 'bucket' but 2048 levels in the 'brightest' bucket.)

    3. Ways to change how the camera deals with its dynamic range
      A: Expose to the right (ETTR) Q: At same shutter speed and aperture, choose ETTR + higher ISO or normal exposure + lower ISO?
      and/or
      A: get the exposure right just for the areas considered most important

    4. What else can be done
      A: more careful metering of your subjects (spot metering, spot highlight metering)
      A: use fast prime lenses fully open
      A: shoot in RAW to get maximum flexibility out of your sensor
      A: apply noise reduction in post-processing
      A: long exposures
      A: use exposure bracketing (to a. get the right exposure b. learn about the effects of exposure on RAW PP results and c. find out if and how ETTR works for your camera)

    5. How do these settings affect your shooting
      A: create a still scene, let people pose
      A: rocking motion and sequential shooting to find focus and get more sharp images at full aperture
      A: sophisticated spot metering techniques
     
  2. mattia

    mattia Mu-43 Hall of Famer

    May 3, 2012
    The Netherlands
    Dynamic range is a property of the camera sensor+electronics package - there's more information in the highlights (hence 'expose to the right') than in shadows, so your goal is to expose as far to the right as possible if you want maximum headroom. Of course, there are plenty of (artistic) reasons to deviate from this rule.

    I think it also depends on expectations - ISO 3200 will never be all that clean compared to base ISOs. And that's often fine. I tend to apply any noise reduction in post-processing (i.e. let lightroom or DxO Optics do it), and shoot in RAW where possible to get maximum flexibility out of your sensor. Also, particularly with a m43 camera, shoot with as fast a lens as possible (i.e. not the kit zooms)
     
    • Like Like x 1
  3. pheaukus

    pheaukus Mu-43 Regular

    178
    Jun 22, 2012
    That's a quick reply!

    That I believe to understand :smile:. Still there may be changing sensitivities over the width of the DR and different ways to map it in-camera.

    I am sorry but this raises more questions for me...
    - With 'more information in the highlights', do you mean that that is a more receptive area within the sensor DR? Or does it mean that one just allows more photons to hit the sensor, thereby reducing noise?
    - If ETTR shifts the histogram graph right, is it the same as overexposure?
    - And finally, in order to keep shutter times low, I have a choice between a) doubling the ISO or b) 1.0 overexposure/ETTR. I don't yet understand when one is favorable over the other.
     
  4. RnR

    RnR Mu-43 All-Pro

    Sep 25, 2011
    Brisbane, Australia
    Hasse
    Low light situations are interesting for me too. I hate to use flash. It seems to destroy the soul of the scene, or maybe I'm just a flash nub :D I've just been looking at some pictures from a couple of recent low light situations, and this is my tactics for next time.

    1) Get 2-3 people to pose with the 20mm at f1.7. The scene is still, so the iso should be reasonable with IBIS.
    2) Use my konica 50mm f1.4 to shoot from a little ways away in BW at high iso. Try to get smiles, gestures etc. Use high speed seq shooting mode and rock slightly forward to get at least one photo with half decent focus. Save up for a konica 57mm f1.2 :D

    Caveat - I'm still very much still a beginner.
     
    • Like Like x 1
  5. pheaukus

    pheaukus Mu-43 Regular

    178
    Jun 22, 2012
    I fully agree on not using flash. I actually like that my camera does not have a built-in flash...

    I'm still waiting for my Nikkor adapter, will try the rocking when I get it. I believe people will smile more ;)
     
  6. David A

    David A Mu-43 All-Pro

    Sep 30, 2011
    Brisbane, Australia
    In low light situations, the areas you are interested in are often brighter than the surroundings. For example a month or two ago I spent some time shooting a tree under a high streetlight. The tree was in the centre of a small traffic island and also catching light from the lights of passing cars. The actual exposure to capture the lighting on the tree, and also the car lights, was much shorter than what my meter in either averaging or centre weighted mode indicated.

    If you've got that sort of situation you can spot meter the lit area and then use a bit of positive or negative exposure compensation depending on how you want the tone of the area you metered to be rendered. With an Olympus body turning Live View Boost off in the gears menu allows you to see the effect of exposure compensation on the screen/in the viewfinder (if you use a viewfinder) and you can actually just adjust exposure compensation until that area is correctly portrayed on the screen/in the viewfinder for you. I was using the PL 25mm F/1.4 and I was getting the shots I wanted in late dusk at shutter speeds of 1/25 and 1/40 second from F/1.4 to F/2 as the last traces of sunlight were fading, and that was using ISO 200.

    The point of that story is that large dark areas in a scene will fool the meter if you're using the averaging or centre weighted modes. Using the spot meter and metering the brightest areas and exposing to get them right will end up giving you the scene as it was, with some bright areas and the rest of the scene looking dim to dark. That's what you should be aiming for unless for some reason you're trying to get detail in the shadow and dark areas.

    The other good guide in that type of situation is the highlights and shadow display. That will start to indicate blown highlights in areas which are about 3 stops brighter than the mid-grey the meter is calibrated for. If the area showing the blown highlight indication is a genuine highlight in an area where you want to keep some detail, then you want to use +2 stops EV above the standard spot meter recommendation or -1 stop EV below the highlight spot mode recommendation. For normal caucasion skin tones those exposure compensation values are +1 stop above the standard spot meter recommendation or -2 stops below the highlight spot mode recommendation.

    If you want to play a quick game to test this out, set your camera to spot mode your exposure compensation to +1 stop, and then meter and shoot caucasian faces on a TV screen in a dark room. The exposure will be shorter than you think, you can use ISO 200 if you have a reasonably bright TV screen, and you will get a good exposure of the screen with good skin tones on the face and the part of the room you can see around the edges of the screen will be quite dark. Try doing an average meter measurement and/or a centre weighted mode measurement and see what exposure settings those methods recommend. They will be longer than what you get using spot mode and measuring the face of someone on screen, and the results you get will not be as good as you get using the spot method I outlined.

    I've read a bit on ETTR and tried it myself but I've never really been satisfied with the results I've got. I tend to be much happier when I get the exposure that gets the tonality and luminance right for the areas I'm most concerned about. Since I shoot RAW I can recover a certain amount of blown highlights with the E-P3, and also get detail in shadow areas in post processing. I don't find pushing the exposure to the right in camera and then pushing it to the left in processing to get the tonality right gives me as good a result as I get when I've exposed correctly for the parts of the scene which are important to me. YMMV but I will add that in a sub-tropical area like the one in which I live, it's hard to shoot outdoors with the E-P3 without getting the highlight and shadows display showing blown highlights so exposing to the right is rarely an option for me outdoors and when the light is dimmer I'd definitely rather keep my ISO setting as low as possible with the E-P3 because of the noise I get at ISO 1600 and above. Once again that means getting the exposure right for the areas I consider important rather than increasing the ISO setting even further to give me an option of pushing the exposure to the right at the cost of noise.

    Learning to meter correctly and how to get the right exposure for the parts of the scene that are most important for you will let you keep your ISO settings as low as possible and that gives best results in my experience. And by keeping your ISO setting as low as possible, I'm not suggesting you shouldn't go above 800 or 1600 or whatever. You go as high as you need to get the shot you want but no higher, and if you learn how to expose correctly for the parts of the scene that are important to you you may find that you can actually get the result you want at a lower ISO setting than you expect.
     
    • Like Like x 3
  7. David A

    David A Mu-43 All-Pro

    Sep 30, 2011
    Brisbane, Australia
    1-digital sensors take more data from the brighter parts of the scene than they do from the darker parts, and more data means more information. There's always more information in the highlights, provided they aren't blown because once you blow them there is no data captured.

    2- shifting the histogram to the right may or may not be overexposure. Whether or not it is depends on what the meter is measuring. Still, shifting the histogram to the right will sooner or later result in overexposure. If you follow ETTR and move the histogram to the right when taking your photo, you then have to do the opposite and move the histogram to the left in processing to get the right brightness values in the final image.

    3- Doubling the ISO lets you use the same aperture at half the exposure speed of the original ISO setting, or the same shutter speed and close the aperture down 1 stop. Whichever one you choose, you're still giving the same equivalent exposure as the meter recommends at the lower ISO setting. Dialling in 1 stop overexposure or exposure compensation doubles the exposure which is a very different thing. Doubling the ISO does not double the exposure if you're following the meter's recommendation, it just lets you shoot with half the shutter speed or with your lens stopped down 1 stop further and get the same exposure as you would have got at the lower ISO. The reason for doubling the ISO setting is to get that shorter shutter speed or smaller aperture for greater depth of field. Adding 1 stop exposure compensation just doubles the exposure so all parts of the scene will be rendered brighter in the final image.

    In some cases you want to go the exposure compensation route, for example to get detail in the face of a person with the sun behind them rather than having them appear in the image as a silhouette. In other cases, such as if you're trying to shoot a moving animal or bird in dim light, you double the ISO in order to be able to get a higher shutter speed to prevent the subject's movement from resulting in a blurred image.

    There are a number of good books available on exposure and it would be worth your while to read one in order to get a better understanding of how the 3 elements that make up exposure (aperture, shutter speed, and ISO setting) relate to each other, and also to learn how the different meter modes work and what is involved in deciding what is the correct exposure for a given scene. There's far too much involved in all of that for anyone to give you a good answer here.
     
    • Like Like x 2
  8. mattia

    mattia Mu-43 Hall of Famer

    May 3, 2012
    The Netherlands
    Shooting in low light situations doesn't afford much opportunity to really ETTR the way you do for, say, landscape photography, where you want as much detail in there and you don't have to worry about things like shutter speed or ISO. I shoot very, very close to the edge and squeeze my gear for all its worth (often wide open, often violating the 1/focal length shutter speed rule, making full use of high ISO capability) when shooting low light.

    'Right' exposure isn't always the theoretical optimal exposure for a shot either. As for the dynamic range issue. Luminous Landscape explains the 'why' of it best:

    Expose Right

    Short version: there is more information stored in the right hand side of the histogram than in the left, so if you haven't blown highlights, it's better to push down exposure in post rather than try pulling it up. I also find that with modern cameras, shooting RAW, it's better to simply shoot at higher ISO if you need it, because pushing to higher ISO by increasing exposure by 1 stop in software will tend to give you worse results. Sometimes that's the only feasible way to get a shot, but it's not optimal.
     
    • Like Like x 1
  9. fredlong

    fredlong Just this guy...

    Apr 18, 2011
    Massachusetts USA
    Fred
    I second everything David A said. I'd only add, don't underexpose! In low light situations with light sources in or just outside the frame it's really easy to underexpose your subject. Let the lights get blown out (unless you're photographing the lights).

    High iso often gets a worse rap than it deserves because the images are underexposed. In most cases you'll be better off increasing your iso instead of underexposing a lower one.

    Fred
     
    • Like Like x 2
  10. pheaukus

    pheaukus Mu-43 Regular

    178
    Jun 22, 2012
    I read the article and it was an eye-opener. At first I thought that ETTR is just conventional overexposure. The article you refer to states ETTR actually matters because of the way how the sensor receives and stores information: The sensor is a linear device that tries to capture 'brightness', a logarithmic distribution. This discrepancy is reflected in the levels of brightness stored in RAW. A sensor covering a DR of 5 F-stops would have 128 discernable levels of brightness in the darkest F-stop 'bucket' but 2048 levels in the 'brightest' bucket. That makes sense; I hope I can make use of this in praxis...

    I just tried to PP some images I took in a stage setting (black walls, low artificial lighting, dimly lit people) where I chose -0.7 underexposure to keep shutter speed fast and avoid blowouts of stage lighting at ISO1600. This resulted in a blue/violet veil of noise on all dark tones and skin shadows were noisy as well. Knowing I will not be able to save the blacks I now managed to rescue the shots (somewhat) by correcting exposure +0.7, reducing black -40 and gentle NR. Good to know that I did exactly the wrong thing and my camera can do much better in the future!

    Thank you very much!
     
  11. pharaviel

    pharaviel Mu-43 Veteran

    313
    Jun 20, 2011
    Reggio Emilia, Italy
    Daniele Frizzi
    When I started doing low light photography (4-5 years ago) I found very useful to keep a slow burst setting and bracket the exposure with 3 frames. Sometimes the center one was good, sometimes the underexposed one would have less motion, sometimes the overexposed one was better quality (because: ettr) if there was no moving subjects.
     
    • Like Like x 2
  12. pheaukus

    pheaukus Mu-43 Regular

    178
    Jun 22, 2012
    Yes bracketing can help! I will try this to find out how my camera reacts to
    different situations... mille grazie