Metering for dynamic range
It’s a cold hard fact, but in case you didn’t know it – photography is not ‘real’. We don’t capture reality as it is with a camera.
You might wonder what I’m on about, but let me explain further. Firstly, a camera sees in 2D, whereas we see in 3D. But also, our perception is a lot more different than a camera sees because we have far greater ‘dynamic range’. We are able to register detail in shadows and highlights that a film or digital camera can only dream about.
ISO 50, f4 = average reading of 1/4 second, 5 stop dynamic range where shadows are at 2s and highlights are at 20th of a second
This point is mainly the reason for why we often get images back from the lab, that looked nothing like how we perceived the scene at the time. It takes skill and patience to be able to get to a point where you are confident you are going to capture what you saw, and you may need to use neutral density graduated filter to squeeze the the entire range of tones from highlights right down to shadow onto your film/sensor.
So I thought I’d explain a bit more about dynamic range.
Our eye is capable of seeing over a range of 24 f-stops. But a digital sensor or film can see roughly 3 to 6 f-stops depending on the medium (negative film has a higher latitide and can often record over a wider range than slide film can). So it’s very common to want to record a scene that the camera simply cannot handle. You are either going to have an image with underexposed ground, or over exposed sky, or a bit of both.
Now, your camera simply meters everything and works out an average of 18% grey. For instance, if you point your camera at a black wall, the meter will make an exposure which will make the black wall grey. The same is conversely true for a white wall; your meter will make a white wall grey. This is why you need to add or subtract exposure compensation for situations when you know that the overall scene is too bright or dark.
So how do you get the right exposure?
Well the truth is that there is no such thing. It is purely about what you consider important to record, and what you are willing to sacrifice, but there are times when using an ND graduated filter will allow you to squeeze an entire scene which has a dynamic range greater than your film/sensor can record, and here’s how I do it.
I have a rather nice little Sekonic meter which allows me to take many spot readings around my scene and from that I can see the entire dynamic range that the scene contains. For example, I will take a spot reading of the darkest part of the scene and then a spot reading of the highlights in the scene, and from that I will perhaps see that there is a 10 stop difference between these extremes. I know that my Fuji Velvia film won’t handle this and that I really need to get the scene down to around 6 stops of a difference. If I put on a 3 stop ND graduated filter to compress the sky (highlights) down by 3 stops, I’ve reduced the entire dynamic range from 10 to 7 stops. Which is a bit more manageable, but I also have to determine what in the scene is going to appear 18% grey. Remember that when I meter the dark areas of the scene, the meter is telling me what the exposure would be if I wanted the dark areas to be grey. And the same is true when I meter the highlights. So I need to decide where in the middle of the range I want to be my mid tone. This is the bit that is subjective, but if I find for example some rocks in the scene that I think are 18% grey, I will decide that is my exposure point. I will then set this exposure manually (to stop the camera’s meter from changing the exposure when I add on the grad), and take the shot.
With an SLR, it’s a lot easier than this – sure it’s nice to have a meter that shows you the entire range so you can figure out what difference in exposure latitude you need to make, but often, the camera will work out a nice average for you – if you have already placed a grad on the scene anway. So it’s much more of a point and shoot approach to it, and if you are using a DSLR, you can check the exposure to see if you over did the grad – often for example, I tend to find that 3 stop grad is too much and 2 stops is more appropriate.
Thank you Bruce, that was incredibly helpful!
Mark
Comment by olwick — 29 August, 2008 @ 3:59 pm
Hi Bruce,
It’s one thing I had trouble with during the Patagonia trip – getting the DSLR to meter consistently when using ND grads. I never really worked it out but its metering seemed to go askew often when I put the grads on and I would find that the scene could be over or under exposed by as much as 1 stop.
I did a lot of take, check the histogram, add exposure compensation, take again. Not too sure why this was happening. I don’t know if the metering software has anything fancy that attempts to identify “sky” and adjust exposure for it. I assumed that there would be something I should be doing to get it to work properly, but it was easy enough to mess with exposure compensation.
Any thoughts?
Niall
Comment by Niall Connaughton — 29 August, 2008 @ 4:08 pm
Hi Niall,
I wasn’t aware you were having trouble with your exposures… normally a DSLR will make a nice average for you when you stick on the grad and you don’t really need to do anything.
However, I’ve noticed that my Canon 5D’s exposures vary wildly when using grads and full NDs…. I have found that the culprit is light leaking in the eye piece.
Here’s how to check for it: Make a reading of a scene with the grad on, and then put your hand up against the eye cup. Take the reading again. If it has changed, then this means that the meter was being affected by the light leaking in the eye piece.
Some cameras allow you to block up the eye piece by either using a curtain on the eye cup, or there is a rubber cover on the camera strap to place over the eye piece.
My own personal way of working is to compose the image on the tripod, and when I’m taking the reading (and the exposure too), I place my hand over the eye piece to make sure no light is leaking in.
I never had such trouble with other SLR’s, but for some reason the 5D seems to be particularly sensitive to this issue.
Comment by Bruce Percy — 29 August, 2008 @ 7:29 pm
Hi Bruce,
That could well have been the cause of my issues. I have used the ND grads handheld lately and didn’t suffer from the exposure issues, but definitely remember having problems in Patagonia, and still have all the badly exposed photos next to the correctly exposed, compensated ones to remind me of it. It’s also happened to me in Switzerland as well.
Next time I’m using the tripod and the grads, I’ll be sure to try your test to see if it is indeed causing a problem.
Thanks!
Niall
Comment by Niall Connaughton — 10 September, 2008 @ 4:45 pm
You should always check your histogram as well Niall. As much as a histogram is *not* the final word on a good exposure, it will give you an idea if you’ve lost shadow detail – the exposure will go of the left hand side of the display.
There is no such thing as the ‘correct exposure’. And each scene will do different things on the histogram – you just need to avoid clipping the histogram at the left (blocked shadows) and right (blown highlights).
Also, it’s best to try to get the histogram to the far right as possible – the more to the right, the more bits recorded. The more to the left, the less bits recorded. Once you have a good ‘right handed histogram’, you can then use Photoshop to pull the exposure down without losing the bits in the original file.
Comment by Bruce Percy — 10 September, 2008 @ 4:53 pm
Hi Bruce,
Thanks for this very helpful description of how you use your spot metre. As a recent convert to film (from digital), I was wondering if you could tell me where you would ‘place’ your brightest highlights in say a delicately hued sky with Velvia? I have tried placing bright highlights (with colour) about 1.5 stops above the midtone value, but I felt that there was more headroom there. From your article it sounds like you might go as far as 2.5 stops above midtone? Any advice woul be much appreciated.
I love your photography by the way. Stunning and inspirational.
cheers,
Ben
Comment by Ben Heaven — 5 January, 2009 @ 5:21 pm