Night/Low Light (Manual) Focus issue

Discussion in 'Canon EOS' started by reish_lakish, Jan 24, 2012.

  1. I'm grateful for ideas, your technique/work-around, pointers or instruction. The problem is likely my technique; not the lens. I've just tested the lens (an EF 28mm 1.8 USM), in good light and bad, and it performs as it did the day I purchased it.
    These images tell the tale of a difficult night with a Banyan tree. Yes, the tree is hardly well-lit; but I have similarly poor results with better lit trees using manual and auto focus.
    The details:
    f/4, 30 secs, ISO 100. (This low ISO because significant noise is introduced even at ISO 200--with noise reduction on. I did shoot alternate versions at ISO 200, 400 and 800, stopping down with each change. No improvement in sharpness whatsoever; but quite a lot of noise in the shadows did result. I did not bother to process these.)
    Metering: highlight = patch at the tree's base; midtone off a card.
    Mirror Lockup: enabled.
    Noise Reduction: On.
    Tripod: True and trusted.
    Remote trigger: check.
    Wind: no wind.
    Earthquakes: none.
    Focusing technique: For this particular shot I tried both manual and auto focus. I stood about 3 meters from the tree. For auto focus, I aimed a flash light at the tree and targeted the beam. The auto focus read the same as my manual focus setting.
    The problem, illustrated:
    Screen shot 1
    : in Raw Therapee. Please note the focus problem in the magnified "pop-out." (CLICK.)
    Screen shot 2 shows a gruesome detail while post proc in Photoshop. (CLICK.)
    Final image: an attempt at salvage--two unsharp layers and significant cropping later. The picture is completely unsuitable for screen or print (CLICK.)
    Again, I suspect--like I've said many times before over expensive dinners--it's me, not it. And, again, please keep in mind that this problem is occurring with many (not all) of my night/long exposure shots--even those where the subject is better lit. The problem is inconsistent though my technique is roughly the same. Inconsistent as it is, it's becoming increasingly more common in my night shots.
    Thanks in advance-
  2. I did not look at all of your examples, but am instead answering based on my own experience with night photography. There are several options for getting good focus with night photograph:
    • Use a laser pointer to put a spot of light in the frame and manually focus on the spot.
    • Place a small light (I use a pocket LED) in the frame, focus, remove the light and make your exposure.
    • Use live view with exposure simulation turned on. In many cases - though not quite all - this may give you enough light without the previous two techniques to achieve focus manually.
    • Arrive early and focus before the light disappears. Leave the camera in place to make the exposure later without changing any settings.
    • Do slight focus bracketing among several exposures and take the best one as your final image.
    • Find a nearby subject that is better lit and about the same distance from your camera position as the tree. Focus and recompose.
    • Avoid the largest apertures, even if this requires a much longer exposure, since they provide narrower DOF and reveal focus issues more readily.
    • Try to shoot with shorter focal lengths when possible. Longer focal lengths are more susceptible to motion blur from even slight breezes vibrating the camera, and this is true even with excellent tripods and heads.
    • Don't worry about sharpness "problems" that are only visible at 100% magnification on your computer screen. The real issue is whether or not the image is sufficiently sharp for the presentation that you will actually end up with - a small jpg for web presentation can look share even if it is far less than perfectly sharp at 100% and even a fairly large print can be sharp with some softness visible at 100%.
  3. Just a couple of possibilities - cars passing by may twirl the air just enough to vibrate your tripod and give you slight unsharpness. If you press the remote release, then drop it and let it hang by the cable, the drop can again induce vibrations (hold in your hands for the duration of the exposure and try moving it as little as possible).
    If you can, go back to the same tree in broad daylight, use the same focusing technique and evaluate the results - that will tell you whether the problem is long-exposure vibrations or focusing. Speaking of which... do you use Live View focusing or do you focus through the viewfinder? Viewfinder focusing is not nearly accurate enough... and live view is difficult to use in extreme low light (my 5D Mark II won't let the shutter speed go below 30 (i.e., 1/30 second) so I'd have to take the ISO up to some 10000 or so (plus open the aperture wide) to even get a usable image on the screen). Also, if the problem is focusing, something on the tree should be really sharp.
  4. Thanks, Dan & Peter. I'm grateful for the quick reply.
    @Dan--terrific pics on flickr, btw. love the yosemite albums.
    • in lieu of a lazer i used a mag lite. (people in my neighborhood see a lazer dot and they run for their kalashnikovs. the red dot sends a signal. not kidding.)
    • no live view on my poor old 5D (classic).
    • wouldn't leave any camera in place for any length of time, out here, for reasons related to bullet 1. ;D
    • bracket focus: always a good idea. my defacto practice. the pic i posted was the best of the lot.
    • i seldom shoot wider than f/4; typically i'm between f/8 & f/16.
    • i shoot just two primes: wide @28mm & short-tele/portrait@85mm. when shooting long exposure landscapes or @hyperfocal distance it's the 28. so far, it's been plenty wide, esp for the hyperfocal stuff. and a decent lens for the money.
    • sharpness @100% is a problem here 'cause i'm printing--pretty much 1:1.
    @Peter, thanks.
    • Cars may be a possibility, though i made the effort to wait for cars for fear the headlights would ruin the image. it's a quiet road with houses along it vacant for a while now. an ideal place to shoot. (and, sadly, get shot at.)
    • i use a wireless remote trigger.
    • you're on to something with the daylight test. i re-shot the neighboring tree which sits in better street light and used only manual focus. the results are significantly better. i'm wondering if the problem is less lens and more sensor. (your suggestion about eliminating the long exposure part in the test is excellent. thanks.)
    • again, no live view. (and, agreed, viewfinder focusing--never mind composition--is a nightmare at night.) candidly (no slight to anyone reading this here in a canon forum, nor to canon) i decided against upgrading to the m2, waiting instead for a pro mirrorless camera. the time's come.
    • re something sharp on the tree: banyans are good for that. plenty of hard definition. better than Schwarzenegger in his glory days.
    I'm grateful for the ideas, Dan & Peter. You've raised good points to keep in mind for my future practice. I think my problem is light. I chose the wrong set of trees--too dark. Is it possible that the hardware/firmware has trouble interpreting a dark edge from a dark area? Like I say, tonight's results, made in better street light, are a world apart--though the settings and my distance from the subject were identical to last night's fiasco. I'm just saying.
    Thanks again, both. Enjoy-
  5. You might do a test shot at f/1.8 at some very high ISO. Don't worry about the noise. Play back the image at highest magnification to check your focus. If it looks good, you're probably pretty close. Then stop down to f/4, drop the ISO, and take your "real" frame. Better still, stop down to maybe f/8, as your lens won't be at its sharpest at f/4.
    Also be careful about your tripod. Wind, cars, and earthquakes aren't your only relevant motion sources. Is the ground soft/spongy? Have you jammed the tripod firmly into the ground? Have you weighted/bagged it?
    FAIW, I don't see much noise in a long exposure shot on my 5D, with noise reduction, even at an ISO of 400.
  6. Thanks, Sarah. I appreciate your help. Good suggestions--particularly those about the sort of ground on which to place the tripod. In this shot I was on a concrete sidewalk on a pretty un-trafficked street. My camera bag, a big Kata, hung from the hook off the bottom. It weighs about 6lbs without the camera.
    Re the noise, this is one noisy camera. I picked it up used from one of the larger, reputable, shops in NYC. Noise occurs even @ISO 200. I've had Canon look at the camera. The problem is known to them. I'll leave it there--but will say that, in my opinion, Canon is among the best companies I've worked with. Their service is excellent--as are their products. Why this particular problem, which they acknowledge, seems unfixable, is beyond me.
    Again, I'm grateful for your solid suggestions--as I am to Peter and Dan.
  7. I just looked at your three samples. Frankly, I'm not really seeing anything all that unusual, given that we are looking at 100% crops of night photographs. There is some noise... but nothing unusual. There will always be noise in digital captures at 100%, but this doesn't look like anything that would be visible in a rather large print.
    As far as sharpness goes, it is hard to tell much from your samples. Low contrast areas, especially when very dark or even underexposed, do not sharpen all that well, at least not with the same settings that might work for brighter and higher contrast areas of the scene. In photoshop you can try some selective sharpening of different versions of the image in layers and then use masks to combine the best portions. You might also try one of the methods of constraining the more aggressive sharpening to the higher contrast edges. (I'm fond of the tools in ACR for this.)
    As to the source of what you still might perceive to be a sharpness problem, there are only so many places left to look. You should perhaps try to eliminate as many variables as possible in order to make it easier to narrow down possible causes. For example, do you get sharper images from your camera/lens in the day time? If so, it is unlikely that the hard ware issue is involved, and much more likely that, whether or not you want to think so, that there is an issue related to difficulties in focusing in very low light conditions.
    One thing that frustrates those of us who try to help is when the poster sort of knowingly tells us that all of our thoughts are not possible explanations for the issue. At some point, we wonder then why the poster is asking. In some cases (though you haven't given us anything concrete that would confirm that this applies to you) I have seen a poster with an "equipment problem" absolutely stick to their belief that there was no answer or that the company had sold them bad equipment - in which case we probably can't help much...
  8. You should perhaps try to eliminate as many variables as possible in order to make it easier to narrow down possible causes.​
    Dan, this is sound advice and I appreciate it. Ditto your reinforcing the idea about the difficulties in low contrast.
    re this:
    One thing that frustrates those of us who try to help is when the poster sort of knowingly tells us that all of our thoughts are not possible explanations for the issue. At some point, we wonder then why the poster is asking.​
    Please be assured that my reasons for posting are legit. I'm looking for answers. Much of what you wrote in your first post is indeed sound and I said so. I think to say more may detract from the discussion, refocusing it on board etiquette. I'm happy to have that discussion off line. Insofar as others may search this thread in search of answers, I'd rather stay on tack. But do take me at my word that I'm grateful for the suggestions, think they're sound, and add much of the advice offered into my practice. I learn from it all.
    Thanks again. Best for the weekend-

Share This Page