UKC

I was taking some photos at a well known viewpoint in the USA recently at sunrise. Also there was a photography class. Their instructor was getting all excited because the light was so stunning (he obviously went to this spot regularly) and he exclaimed "wow, the light's so good this morning it doesn't even need HDR". I really don't know what to think of this remark. Does most light "need" HDR?

In reply to Robert Durran:

No, unless you're into hideously over processed landscapes that are neither photography nor art.

In reply to galpinos:


Even used in moderation? My compact camera has a built in HDR facility which deals subtly with high contrasts.

In reply to Robert Durran:

My Nexus phone takes HDR photos as standard, especially in low light conditions, cutting out the need for the flash. And to be truthful I can not tell that they are HDR images, they just look like normally composed and exposed images from any standard phone or compact camera.

HDR can be both subtle and garish, but with the progress of a camera's computing power to process an image, maybe HDR is a continued evolvement of point-and-click photography?

In reply to galpinos:

For me it's all about replicating the human eye. SLR's don't have the dynamic range of the retina and HDR can compensate for that in high contrast scenarios. If you like blown out skies or under-exposed landscapes then of course, be my guest.

In reply to galpinos:


Well, that's misuse of tone-mapping, not 'HDR'.

In reply to Robert Durran:

HDR has become a synonym for (mis)use of 'tone mapping', which is a secondary process used to bring the range of luminosity values of an image into the viewable range of the medium (be it print or digital image file). This in itself isn't dependent on the original data being HDR - gamma compression is also a method of tone mapping.

Whether an image 'needs' HDR is dependent on the range of intensity values in the original scene, and what the photographer is trying to achieve. The human eye is pretty good at managing to 'squash' all the intensity values of a given scene into one image, and HDR imaging (or, strictly speaking, HDR image capture followed by a tone-mapping operation) is an attempt to replicate this process. The classic image from the original research was the interior of a cathedral with a stained-glass window, a scene that would typically present problems for conventional 8-bit imaging due to the range of luminosity values (both deep shadows and bright highlights in the same scene).

If anyone's particularly interested in the geeky details ( which of course you all are ), Paul Debevec's 1997 paper kicked off the whole thing. http://www.pauldebevec.com/Research/HDR/debevec-siggraph97.pdf

In reply to purplemonkeyelephant:


Me too, and if HDR can help then great.

Bur what interested me about the instructor's comment was, as I took it, the implication that HDR could compensate for poor light (the word "good"). If he had just said "the light this morning doesn't need HDR", then fine.

Anyway, he was clearly a real enthusiast and even asked how I was getting on with my shots.

In reply to planetmarshall:


Is that what I'm doing when a slide the lttle thing along on my computer to bring up the shadows?

In reply to Robert Durran:


Basically, yes. If you did this with an 8 bit image file, you'd get 'banding' artifacts in the processed image (if you only have 10 greyscale values in the original image, then multiply them all by 2, you still only have 10 values - just with bigger gaps inbetween).

If you have an HDR image to begin with, you have much more data with which to process, so you don't tend to end up with processing artifacts in the final image.

In reply to Robert Durran:


Not really. HDR just gives you a more faithful reproduction of what was in the scene to begin with. It's still Garbage in, Garbage out.

In reply to Robert Durran:

Their instructor was getting all excited because the light was so stunning (he obviously went to this spot regularly) and he exclaimed "wow, the light's so good this morning it doesn't even need HDR".

If the light was really good, then it was probably 'reflecting' on the rocks, etc. Therefore the contrast between the land and sky would be minimal and you wouldn't need such things as HDR or grad filters.
The same would be true if the light was really flat for the same reason.
Alternatively, the bloke might have been a sound byte bullshitter!

In reply to planetmarshall:


Ok, so I could (charitably!) interpret the remark as just meaning the light happened to be such that a single exposure could give enough data to faithfully reproduce what the eye/brain does (rather than that the sunrise that morning for once wasn't a turd that needed polishing)? Though the way he was jumping about excitedly suggested that most other mornings he was there were (coincidentally?!) turd like in comparison with this one!

In reply to FactorXXX:



I'll post up on here a couple of my amateurish (non HDR) shots to show what the light was like!

In reply to purplemonkeyelephant:


I'm not sure that it does replicate the human eye. When you look at a view your eye won't take in all of the detail in one go, it will refocus onto different areas at different times, rather than trying to see the whole lot at once.

In reply to timjones:


Yes, but I think our brains kind of compensate for this in the sense that we scan a view and kind of think of it all at once.

In reply to Robert Durran:

You were lucky to get the clouds! I stood on nearly the same spot years ago wishing for anything but the bland, cloudless start to another, stinking hot day!
The instructor wasn't the Brit photographer Niger Turner, by chance?

In reply to dek:


Yes, it was a tad hot!


No, he was (very) American!

In reply to Robert Durran:

The idea of ETTR is that far more gradations of brightness are available at the bright end of the tonal scale than the dark end. Therefore if you overexpose the sky (but not to the point of burning it out) you can darken it down without introducing noise, and you will have lighter shadows which will withstand more pushing.

But it's really designed as a technique to use on raw files rather than JPEGs.

https://luminous-landscape.com/optimizing-exposure/

In reply to Robert Durran:

I hope you got up to the Dunes, Robert? Shooting in DV is hard work, trying to come up with something different. The dunes at sunrise are magical!

In reply to Robert Durran:

The second one is a stunner.

In reply to Removed User:


Thanks!

In reply to dek:


Yes, I am in love with the dunes......a photographer's dream; I think I am these days as much a desert as a mountain person! I took a lot of my best photos amongst them. I criss crossed the DV area several times in between climbing so saw a lot of it, though some major storms with flash floods had wiped out some roads. The Dumont Dunes (accessible though little known because they are just outside the park) are absolutely stunning and the ones near Amaragosa likewise just outside the north end. The Ibex dunes near the Dumont ones but inside the park are bewitching. The Mesquite dunes right by the main road through the park are ok, but nothing on the more remote ones. I finally managed to drive to the fabulous Eureka dunes on my third attempt to coincide with the full moon and the Venus/Jupiter/ Mars conjunction which was staggeringly sublime.

In reply to John2:



Thanks. But with jpegs (and yes, I know I should switch to using RAW, but the computer stuff intimidates me!) should I in general expose for the highlights, the shadows, the average or the most important bits?

In reply to Robert Durran:

Who can say, it changes from shot to shot.

What is certain is that raw files will take more pushing about.

If you use Lightroom (not a terribly expensive product), processing raw files is no more difficult than processing JPEGs. There's a bit of a learning curve, but it's time well spent.

In reply to John2:


I've got lightroom. It's been sittting on a shelf in its box for several months now taunting me for my technological phobia/inadequacy............ eventually I'll find enough time to psyche up and try uploading it to my computer!

In reply to Robert Durran:

Just get as much as you can inside the histogram and you can't go too far wrong

In reply to Robert Durran:

On my Canon SLR you can shoot what's called an auto exposure bracket. The camera takes 3 shots in a row, one based on standard exposure meter value, and either side at either +/- 1/3 EV or +/- 1/2 EV. Might be an option to try, then pick the best one.

Raw images are generally 16 bit so it captures the light values in 4096 steps rather than 256 steps. Hence why they are better if you are shooting interesting lighting conditions and wish to capture the subtle graduations.

In reply to James Rushforth:


Yes, I do that, but if I have a choice having fitted everything inside the histogram, is a good rule of thumb to go to the right?

In reply to Orgsm:

65536 (2^16), to be pedantic, but yes. Actually that's the maximum, with the practical number of 'discrete' levels determined by your sensor. My Nexus 5 can capture RAW images with about 1020 levels, a decent DSLR will probably be much higher.

This doesn't necessarily give you a higher dynamic range, though. For that you typically need to do exposure bracketing and combine the images using software.

In reply to Robert Durran:

Yes as a general rule but not always: (a few example of when you wouldn't).

1) If you're not using your cameras base ISO (usually 100) - any extra benefits you get from exposing to the right will be lost by the increased ISO - better to drop the ISO and not expose to the right.
2) If you need to quickly snap a shot and risk losing some of the highlights.
3) If the increased exposure dictates a higher ISO (see above) or longer shutter speed where there is a risk of blurring your subject.
4) If your not shooting in RAW there is very little point in the first place.

Ultimately it's often pixel peeping differences at 100% crops anyway with a decent camera and there's a lot of debate about whether it's worth the risk of clipping highlights for such a small improvement.

In my humble opinion far better to worry about the light and what your capturing and leave the tech crap to the scientists...

If you want to do some more reading: https://photographylife.com/exposing-to-the-right-explained

In reply to Robert Durran:


I'd agree with that, sadly my brain struggles when the same info is all presented in a single photo It makes my eyes ache.

I also wonder whether the ability to include everything robs the photographer of the artistic licence to choose a focus for the photo?

You've probably worked out that I'm not really very keen on HDR

In reply to timjones:
You can divide tone mapping approaches into two general forms, "global" and "local".

Global algorithms treat the whole image equally, and are basically equivalent to setting up some curves. A point with brightness X in the original will end up with brightness Y in the result no matter where in the image it is. You generally end up with a natural looking picture that your brain will accept as a photo. The down side is that you lose detail as it gets compressed out. Exactly the same as doing a raw conversion.

Local approaches try to increase local detail by accounting for local brightness and boosting the contrast of smaller features. These are the ones that people are thinking of when they complain about HDR. Big problems there are flattening all the big gradients out to a uniform level, noise and halos.

It really doesn't help that most HDR software turns the settings up to 11.

It's possible to do halo free local tone mapping but hardly anyone does. I suspect it's because HDR haters have already decided they won't use it, and HDR lovers actually like the artifacts so there is no point.

I ended up building my own stuff to try and get sensible results. Here's an example: http://www.spectral3d.co.uk/Personal/pics/2012/ballater_feb_2012/ballater_f...

Not a great photo, but the only one I could find at short notice.

In reply to James Rushforth:


Thanks! Useful advice.


Excellent! Most reassuring coming from you. I absolutely love the process of trying to put myself in the right place at the right time, especially when a bit of luck adds some magic, but I hate the computer stuff.

In reply to Robert Durran:

What James said, plus, it's a good rule of thumb to go with whatever looks right to you. I'm currently learning Capture1 Pro8 (reviews said it rendered RAW better) and it is easy to get blinkered by received wisdom or 'correct procedure.' It didn't take me long to figure out how to get what I wanted (from what I could get; Andrew's "garbage in, garbage out" is spot on) from a shot. It is daunting to begin with but you soon get the hang of getting more from your shots and, dare I say, even get quite into it, and I'll bet I'm at least as big a technophobe as you.

Oh, and Photography Life is a brilliant resource, and home to a certain John Sherman.

In reply to timjones: what Robert Duran said. It replicates what we think we see.

In reply to James Rushforth:
I tend to follow the late, great Galen Rowell's advice, "..always expose for your most important highlight."
Mind you, he also said always ski the fall line and I nearly died on an icy black run doing that.

In reply to Robert Durran:


Not sure I fully understand your point here. In my very limited experiments with HDR I've been astonished at how much colour can be pulled out of of a really badly lit interior on a dull day. In that way I'd say HDR certainly can compensate for poor light. Colours appear richer than a scene appears with the naked eye.

In reply to Robert Durran:

Many things would benefit from both HDR capture and HDR display.

Very very few things benefit from HDR capture luminance compressed to an SDR display. Yuck

In reply to stp:



If the light "needed" HDR, I would take that as meaning HDR would help reflect a scene as seen by the naked eye.

If HDR could compensate for poor light, it would be enhancing what the naked eye sees (ie it makes up for bothering to search out or wait for the magical light which, for some of us, is what landscape photography is all about).


Precisely! Which is what gives HDR a bad name!

In reply to Robert Durran:

I don't think this is a bad thing. In fact you can just take a regular photo in a dark place and with no post processing the camera's autoexposure will make it brighter than it appears to naked eye. You can boost saturation either in camera or in post too. You can even make it black and white , perhaps make the sky jet black which is totally different to the naked eye view.

But there's much more to light than simply saturation. You can't compensate for midday sun, you can't alter the direction of light, the shadows or the softness or hardness of the light using HDR. If you want to shoot a sunset you have to be there at sunset.

There are two possible goals for a photographer. One might be to record an image as faithfully as possible. The other is create a beautiful image. I think most landscapes I see fall into the latter category and even if not HDR they've certainly had some post processing work applied. Even simple RAW editing can make a huge differences to an image. Some photographers these days even say that post processing is 50% of a photographer's work.

I think what gets HDR a bad name stems from the fact that when used in a certain way it leaves a distinctive and recognizable signature. It could be seen like a photo has just had a simple filter run over it. With common use this has become something of a cliched look which is why people don't like it. I try not to let myself get biased in that way because there are some great HDR images around. But like every kind of image there are far more second rate ones and when we look at those we don't get taken in by the amazing image but instead what jumps out is that that predictable HDR signature.

I draw & paint as well as take photos. I've read assertions that paints with optical brighteners or synthetic dyes in them cannot make 'fine art'. Some people just think in absolutes, and they're often people who spend less time doing, and more time paying attention to what others say than actually thinking for themselves.

In reply to stp:


Indeed. And it is never possible to remember exactly what the landscape actually looked like anyway (I sometimes, when post processing, catch myself wishing that I had a photo to remind me........ now wait a minute.......!), and it will always be at least tempting to make the the final picture as beautiful as possible.


Yes, I'd like to think that all post processing is at least "invisible" in the sense that the final picture looks plausibly realistic even if it is not entirely accurate. I like the analogy of a woman wearing make up - if you can tell she is wearing it then it is overdone and looks rubbish.

In reply to Robert Durran:


Personally I wouldn't even go that far. There are certain style or genres of 'photography' that are obviously unreal. I like some of the high end fashion photography that's produced. Often the photographer and retoucher are different people. A good retoucher might spend several hours working on a single image. What comes out looks to me something like a cross between a photo and a painting. The fact that its not real doesn't matter to me. It's a beautiful stunning image and the unreal perfection I see as analogous to the unreal perfection often portrayed in statues.

In landscapes there seems to be something of a trend of night time images with stunning backdrops of the Milky Way. I love some of these shots even though I know full well the Milky Way never looks like that in real life. It's just a function of a long exposure, post work and possibly multiple exposures too.

In reply to Robert Durran: Nothing to add here except that I'd agree that image 2 really is superb.