How to correctly expose in this situation.

[sanger]

Is there any way of having both the sky and in this case the train line exposed correctly together.
Both photos were taken at 6.40am and one was set for the sky and the other the line.
Are there tricks to getting the whole photo exposed correctly.
Is this where the graduated filters come into play ?

[ameerat42]

Well... (Takes a breath.) Yes, and No, I suppose.
Let's look at the basic exposure details for each shot.

For Pic 1:
...
Focal Length: 55.0mm (35mm equivalent: 82mm)
Aperture: ƒ/10.0
Exposure Time: 0.167 s (1/6)
ISO equiv: 100
...

For Pic 2:
...
Focal Length: 55.0mm (35mm equivalent: 82mm)
Aperture: ƒ/10.0
Exposure Time: 6.000 s
ISO equiv: 100
...

Only the shutter speed varied, and so that's about 6 stops between 1/6sec and 6 sec.
(Ie, approx 6 doublings of the original shutter speed - 6 stops.)

If you are shooting in (8-bit) jpeg only, then I reckon that you would have LITTLE chance of
getting a good overall exposure in a single shot. However, the specs of you camera state that it
can shoot in 12-bit raw, and I reckon you might give it a go win that mode, picking an exposure setting
somewhere between the two here (and I'd guess a bit more towards the shadows as you'd probably be
able to correct highlights better than shadows.)

Of course, take a series of shots about that median exposure, and WITH CAREFUL PROCESSING OF THE RAW images
you might just be able to fluke something.

I had a recent post here where I managed to get a normal shot from way overexposed using a 12-bit raw file.

Am.

[John King]

Lovely photo. I can see why you took it.

Some suggestions:

Use your metering system differently, perhaps. I would have spot metered around the brightest and darkest parts of the scene, then worked out a manual exposure if the camera could not work out how to get the result you wanted. On one occasion, I took a number of shots of a piccy before working out what was fooling the camera, then took this:



Always remember that you have the best light meter ever invented in your hands. It not only takes light readings, but also shows you a photo of what that light meter reading will give you ... .

i.e. Matrix metering (whatever it's called on your camera) will only get you so far. You need to know how the other metering modes available to you work also - spot, centre-weighted average, high/low key metering, and maybe etc.

You can also take exposure bracketed shots and combine them later in PP.

If you shoot RAW, you can 'develop' the RAW in multiple different ways, then combine them as layers in post. e.g. combine your two images shown here to give a better result.

Bryan Peterson's book "Understanding Exposure" is very simple, cheap, and covers lots of things about exposure, as one would expect. None of it requires you to be a brain surgeon, thankfully!

[thegrump]

My 2 bobs worth. I just twerked the colour, contrast, and sharpness, just a little in Fastscape

DSC_0104a.jpg

DSC_0104.JPG

[ameerat42]

TG. You've lifted the shadows, but where's the sky detail: color, clouds?

You can do a certain amount with these images, but the dynamic range is restricted.
Where you might get the tones up, the color will fade, and perhaps VV.

About all you can do is raise the shadows, a bit at a time.

[John King]

Gidday Sanger

Beautiful cat, BTW. Looks very much like our two villains.

As these are practically identical framing (obviously almost certainly on a tripod), copy each photo to its own layer in PS, then merge the two together after aligning them if necessary. Either that, or let PS automatically align and merge the two layers.

I have to go out, but will give it a go later if I have a chance.

[ameerat42]

Is there an (almost) inevitable shift in emphasis from exposure to PP?

[John King]

Is there an (almost) inevitable shift in emphasis from exposure to PP?

Not for me, Am. I'm a great believer in getting things as right as possible in the camera. As you worked out, this should be well within the DR capabilities of the D3100, so attention to exposure details would appear to be at least part of the answer.

However, exposure bracketing is definitely an option. Even my ancient 2003 Olympus E-1 could bracket up to 5 shots. I think my E-M1 can manage both exposure and focus bracketing up to 999 shots. Should be sufficient!

It gives me a headache just trying to work out what parameters I can use!!

Now, I really must go to the chemist ...

[thegrump]

am .... You are right, I was so busy trying to get the rocks correct, I forgot about the bigger picture.
John .... I also would prefer to get it right in camera, but if we don't ( and that is my case most of the time ), there is always the computer.

[Cage]

Hi Sanger

I had a fiddle in Photoshop with the Darker version as there was no recoverable sky detail in the light shot, not with a 400kb file anyway. Are you shooting in RAW ? It gives you heaps more data to play with.

I found in trying this type of shooting that it helped to take lots of shots at different exposures, unless of course the clouds are zooming across the sky. For example 1/2 sec on the dark shot may still have given you lots of recoverable data in the sky and less to do in the dark areas.

This is a quick play with levels and some colour tweaking but it was too dark to get all the details in the rocks and the trees look a long way off. With a RAW file to work on it may have been a better result.

It looks terrible I know and only posted to give you an idea of how much data can be recovered in PP.

[ameerat42]

Gidday Sanger

Beautiful cat, BTW. Looks very much like our two villains.

...

It can't be the Cheshire variety, as I can't even see its

[arthurking83]

Is there any way of having both the sky and in this case the train line exposed correctly together.

Like already mentioned: HDR

Are there tricks to getting the whole photo exposed correctly.

The best way to answer this is like AM said. Yes and No .. ie. 'kind of'

Is this where the graduated filters come into play ?

Maybe is probably the best answer here. It's not a clear cut yes or no in this scene as the grad will invariably blend into the mix of the scene(in this instance both the tree tops and the hill top on the RHS of the image.
Grads are handy for some situations, but they have their limitations too(and therefore weaknesses, or drawbacks).
The other issue with using a normal grad is that in this scene, the brighter part of the upper half(ie. the sky) is brighter at the bottom of this half, than it is at the top. So in reality a reverse grad would(theoretically) be more ideal.
A reverse grad has a darker strip in the middle, which then gets a bit lighter as it graduates towards the top of the filter.
The drawbgack of the normal grad is going to be even more excessive tho with a reverse grad.

BUT! .. where the actual answer lies, rather than the yes/no/maybe answer given, is in knowing what your camera is capable of and working with that to extract the most out of it.

So as already said, you could do a HDR and get a balanced image in PP .. but that doesn't help you learn about exposure, and more importantly what John referred too in using Spot Metering.

(I'll try my hand at an actual answer now):
From memory the D3100's NEF files are capable of really recovering about 3-4 stops of shadow detail and maybe 1-1.5stops of highlight detail.
In reality, it may be capable of a little bit more(or maybe just a little bit less!) but lets just use these values as a guide.

Am made the comment that you had about 6 stops of exposure difference, but I think this value is incorrect. It may be close, so in a quick give and take situation, it may work well enough to get a good balanced image.
This is where John's remark re spot metering comes becomes important.
What we don't know in these images is where you spot metered from(which is important).

With spot metering, what you ideally want to do(and sometimes this may need to be quick) is get the spot meter point(ie. your focus point) on something that represents about 18% grey.
This is why I think Am's calculations of 6 stops difference is incorrect.

ps. I realise that Am is referring to 6 stops of exposure difference in the actual images, but I'm also assuming that he's referring to actual Ev differences in the two images. The problem is that the two images don't seem to have a properly exposed neutral point, which is how you calculate actual Ev requirements for the scene.
(Am can beat me at a date to be determined when I get the chance to get up his way for him to effect just such a beating! )

Anyhow, using spot metering: point the focus/meter point on something that is close to neutral grey(as this is usually easier).
In the sky in this scene, you have grey clouds! .. perfect
In the foreground you have some nice grey railway bed rocks, which we know from experience are grey bluestoney type rocks.

This is how I guessed the actual Ev difference in the two sections of the scene. The grey clouds in the sky are too dark for neutral grey, and the grey rocks in the foreground are (JUST!) a touch too bright for neutral grey too.
So I reckon the actual difference in Ev could be more like about 7-9Ev. it's impossible to say without being there and knowing exactly what processing was done to the image.

That's the technical mumbo jumbo you need to know on how to 'well balance' an exposure. My answer of maybe for the grad filter will also become a factor soon.
But, forget all that for now and lets work with what AM did say about Ev differences shown here.
With the vast majority of cameras, you can generally recover more shadow detail than highlight detail, and the trick to balancing exposure is to take advantage of this.
So, if you needed 1/6s for the exposure of the sky, then if you over expose it by 1-2 Ev, that means you could(or should have) shot at say 1/3s(1Ev) or 1/1.6s(2Ev) slower with your shutter speed.

(note that this is basically the way I shoot landscapes)

There's a common saying with digital photography and landscape exposure correctness:
shoot for the highlights and let the shadows fall where they do
That is, always watch were the highlights are exposed!!! Shadows can be recovered (and/or deliberately processed to black for effect) and the image won't look bad

But with 1-2 Ev brighter image than the top one displayed, recovering the shadows is now only 4stops rather than the almost impossible 6 stops by using that 1/1.6s shutter speed.
You will have to pull back the sky a little at this 1/1.6s image as it'd be much brighter than at 1/6s.
In fact what I'd have shot is a 1/1.6s and a 1/3s shot, get them home and play with each to see how much detail could be recovered in each image.
Having done this for long enough now, I know what my cameras can allow me in PP in terms of recovery, so I shoot according to those limitations.

This is where the grad filter makes a bit of a comeback into the equation to help out .. so the answer here for the grad has changed from maybe to yes.
Use an easy to compensate for grad. Some folks may be inclined to use a 3, 4 or 5 stop grad(if they'd be available) but I think this would be counter productive.
Even a 3 stop grad would give detrimental looking shadowing(grad effect) in parts of the scene, which makes it look unbalanced.
What I'd use is a 1 or 2 stop grad just to assist with some latitude in PP again .. rather than to straight out balance the scene itself.

That is, if I used say a 1 stop grad, just to get that one extra stop of help for the shadows that 1 stop of shadow help in PP could be enough difference to make the image look nice.

So, if this were me and I had a 3stop/ND8 grad .. I'd have not used it in this scene. To try top process out the demarcation line it'd create on the hill on the RHS would be too much(I think).
I'd use a 1 or 2 stop grad tho .. maybe even try one without a grad too.

Good to see the question asked too .. and remember that what you want to really understand is the metering system more so than the actual exposure required.
Once you have a grip of how and what it's doing, exposure will come to you more easily.

ps. I made a reference to grey spot when using spot metering, and this can't always be done.
I've already made this reply wayyy too long, so will minimise this next section, but needless to say you can use other areas/colours to achieve the same end result.
eg. instead of grey, you can also use greens to determine a neutral Ev exposure. But note that the tone of that green has an impact on the accuracy of the exposure similarities to neutral grey.

Hope that helps .. if it doesn't put you to sleep

[nardes]

Not too sure if you want to go down the HDR path, but I copied/pasted both images into CS6 and ran Auto Align and Auto Blend plus a few tweaks to obtain the following results.

Cheers

Dennis

[sanger]

Thanks guys, some good reading and info to sink my teeth into.
At this stage I'm shooting in jpeg and not raw...once I understand more of the basics I'll get some kind of editing software.

To Dennis I like what you achieved.
And John yep we have two Lilac's to keep us on our toes.