HDR Efex Pro?

[ShaneS]

I'm officially in the corner now about to pull out the white jacket and hope that the walls are padded. This is way over my head technically and I must agree with Mark that I don't understand the point (I get that shadow recovery is very challenging in PP if that is the point).

Is there a way of saying this in words that are more easily understood by the layman without a degree in physics or an interest in pursuing one?

[pnodrog]

We don't need to add anything. The cloud in that situation is just like the white piece of paper or bride's dress in bright sunlight. I'm assuming when you say that we're "not shooting into the sun itself" that the sun is not being obscured by the clouds (example: the clouds are in front of us and the sun is behind us).

I was going to argue with you until I read your qualifier that the clouds had to be behaving as a reflective surface rather than filtering the sun and therefore a lighting source....

[Mike Buckley]

Is there a way of saying this in words that are more easily understood by the layman without a degree in physics or an interest in pursuing one?

Be patient and wait until Colin returns to the discussion. He always explains this stuff in appropriate terminology. He is carefully leading us question-by-question down a path. So far, it's making sense to me. If it makes sense to me, it can make sense to any similarly challenged knucklehead, not that anyone else participating in the thread is in that camp.

[Donald]

I'm officially in the corner now about to pull out the white jacket and hope that the walls are padded. This is way over my head technically and I must agree with Mark that I don't understand the point (I get that shadow recovery is very challenging in PP if that is the point).

Is there a way of saying this in words that are more easily understood by the layman without a degree in physics or an interest in pursuing one?

Shane

There are times when a thread sort of breaks into two parts. Some people get into a very technical discussion that leaves the rest of us way behind. Like you, I'm totally lost by it. But that's okay. You don't need to buy into that part of the thread. You stick with that part that's about taking it back to basics and learning about dynamic range.

The last thing we want is that you get turned off learning because the discussion takes a direction that's not helpful to you.

[Mike Buckley]

I was going to argue with you until I read your qualifier that the clouds had to be behaving as a reflective surface rather than filtering the sun and therefore a lighting source....

That was actually Colin's implicit qualification when he explained that the shot isn't being taken into the sun itself. I only clarified the point.

[Colin Southern]

Yawn (8:14am here) - 'mornin all!

OK - time to pull all of this together a bit more. To be honest, when I read Shane's initial post - and some of the replies - my first thought was there are several fundamentals that people aren't understanding here, and that we're really going to have to go right back to the beginning before we can move forward -- and I've been trying to do that a bit at a time.

If we're talking HDR - and HDR bracketing - and we know that HDR techniques are used to capture a scene who's dynamic range is too big to be handled by a single exposure - then we really need to know (a) what the dynamic range of the scene actually is, and (b) what the dynamic range capability of the camera is to see if we actually need to use HDR techniques to capture it - or not.

We know that the dynamic range of the camera varies with ISO. Higher ISO settings - in essence - just "under-expose" the shot and then amplify it back up again; a good analogy is to think of a microphone when giving a speech - if my lips are a couple of inches away and the master volume control is set to low then the volume everybody hears is correct - however - if I step back a couple of feet the microphone (sensor) only gets a signal that's (say) 1/2 of what it's capable of handling - so we need to double the master volume control to amplify the signal so that people can hear me at the correct level again. If I keep moving further and further back then one would have to keep increasing the volume control to compensate - but - as one increases the master volume control they also amplify the inherent noise coming from the microphone, and so the further back we go the higher the volume control needs to be for the correct sound output - and the less signal (voice) that the microphone receives. Exactly the same with ISO - the light gathering ability of the sensor doesn't change - we just use less and less of it's potential range - and amplify it more and more - and because the inherent noise (what we call the "noise floor") remains the same, the dynamic range that the sensor can capture at higher ISOs effectively reduces by one stop for each stop we increase the ISO (other factors do come into play - but just ignore those for now).

OK - all of that is a little off-track - all you need to remember from that lot is that if you have a scene with a big dynamic range then you need to keep the ISO as low as possible if you want to capture it in a single exposure. Conversely, if you need to shoot something at a very high ISO then you need to keep in mind that the dynamic range capability of your camera will be compromised - so you then have to look at things like ETTR (eliminating the safety margin that most metering builds in) or adding light - or taking a bracket of exposures.

OK - we know that Shane's camera has a base ISO DR (Dynamic Range) of 13.9 stops (what the heck, lets just call that 14!) - and with a scene like she has, there's no reason not to shoot at base ISO - so we have the full 14 stops available to us. Next step is to read the scene, which is what I've been building up to. Just before we look at that, I need to answer that last question about "how many stops of DR do we need to allow for for the sky/clouds etc). In practice, I've found that it depends; if it's just blue sky then under-exposing a metered scene like Shane has by around 1 stop will generally be enough to make the sky look nice - however - if there are nice / bright cumulus clouds with full sun on them then you really need to allow up to about 3 stops.

OK - so lets use all of that info we've learned in "the questions" to evaluate the DR of the scene Shane took;

- it's a reflective scene - lit from the left - so we're going to need 4 stops right off the bat.

- towards the middle we have some naturally darker areas - these will be reflective objects, but in the shadow of other parts of the scene. Ultimately we'll want SOME detail in this area, but not too much or the image will look flat from not enough contrast - so our extra "3 stops of DR for the shadows" should be sufficient here (so we're now up to 7 stops total)

- looking at the sky, we can see a few clouds; to be honest, it's hard to tell how bright they were initially because the image I'm seeing has already had areas like that manipulated, but lets assume that they were somewhere in-between "bright as bright can be" and "just sky" - so I'll allow 2 stops. Add those 2 stops for the sky to the 7 we have so far, and we get to 9 stops.

OK - 9 stops of DR for that scene (give or take) being captured by a camera with a 14 stop DR capability at base ISO. Does 9 fit into 14? Yep - easily. So in conclusion we can say confidently that this isn't a scene that requires HDR treatment to capture. It'll require some DR compression to DISPLAY (because our monitors can only display around 6 stops) (so it'll need a little fill light slider to reveal detail in the darker bits) (which is a different exercise), but we don't need HDR to capture it.

In my mind, the reason the image doesn't "leap out and grab you" is that you're trying to use a specialist program on an image that doesn't need it. In essence, there's nothing for the program to do (but it doesn't know that) - and so by using the program on a non-suitable image - and the program still tries to do SOMTHING - you end up with a flat "meh" kind of image, which is what you have.

So the very very long answer as to why the image didn't "fly" is simply "it's really the wrong type of scene for that kind of treatment".

On a side note - when you DO need to bracket shots to capture a greater DR, varying them by only 1 stop does approximately nothing. I won't go into technical detail too much, but ultimately it's about noise levels, and the difference in noise levels between any two exposures only 1 stop apart is minimal. Normally you'd start with a minimum of 2 stops.

Just as a contrast, I dug up an image - shot from a single exposure - at the right time of day - with just normal processing - that shows what you should be able to get from that composition; just shoot it at sunrise/sunset - long exposure - and don't blow the highlights, and you should be able to get something like this:



Hope this helps some folks in reading scene dynamic range.

[ajohnw]

I'm officially in the corner now about to pull out the white jacket and hope that the walls are padded. This is way over my head technically and I must agree with Mark that I don't understand the point (I get that shadow recovery is very challenging in PP if that is the point).

Is there a way of saying this in words that are more easily understood by the layman without a degree in physics or an interest in pursuing one?

You have it in one Shane. At some point as shadows for instance get darker they get harder and harder to recover and at some point can't be recovered. Hence HDR for the dark end of things. So no need for the padded cell.

John
-

[Colin Southern]

PS: If you want to REALLY see how good HDR done properly looks, then look no further than the work of my friend and (currently inactive) CiC member Yan Zhang:

http://yanzhangphotography.com/

PPS: You can read more about him here if anyone is interested:

In conversation with .... Yan Zhang

[ShaneS]

Thank you for bringing this back to basics Colin - it is appreciated!

I will grant that this photo will never be a 'wow' image and was really just me out practicing in general with a specific focus on my composition skills. And now venturing further along in my overall education so thank you again for taking the time to share your knowledge.

But back to the lesson, this makes sense

In practice, I've found that it depends; if it's just blue sky then under-exposing a metered scene like Shane has by around 1 stop will generally be enough to make the sky look nice - however - if there are nice / bright cumulus clouds with full sun on them then you really need to allow up to about 3 stops.

But let me rephrase...are you saying that at the time of capture I should have dialed in 3 stops less exposure and lightened the shadows (which would now be darker) or used fill light in post? (assuming no filter - see below)

Next...

looking at the sky, we can see a few clouds; to be honest, it's hard to tell how bright they were initially because the image I'm seeing has already had areas like that manipulated,

I did not manipulate this area at all before posting - what you see above is untouched with the exception of the adjustments you see in the ACR screenshot. Remember I left the curves at linear and no sharpening for the sake of showing the image here. I did however use a Hoya circular polarizer that I believe is causing the 'dulling down' of the sky as you see presented. Is this a poor quality filter, did I use it incorrectly or inappropriately? If I am correct I lost a stop+ just by mounting the filter and if I understand everything above (barely I think) then that would also deepen shadows that already existed in the scene and would make it even harder to recover them in post?

The next question that I have is a bit of a tangent and related to IQ for a scene like this. As shot and for the sake of this discussion, the image seems a bit soft at 100% (again with no adjustments as mentioned above)



I have been using tonal contrast in the NIK collection and usually like the result. In this case I dialed it back quite a bit from the standard presets but it seemed like what ever I did I got artifacts around the edges of the cliff and the image just got really noisy. Is this the filter, the lens, my focus point, wind moving my tripod, the sea spray from the waves or???? Sharpening of any kind (pre or post for web presentation) made it even worse!

The only other thing that I can think of is that a scene like this one with a high dynamic range can create issues with IQ in general? What are the challenges to keep in mind beyond shadows and clipped highlights? Then how to deal with them pretty please

The white jacket is off and I am about to remove the dunce cap (but I will keep it close by just in case )

Thanks to everyone for helping to keep this thread alive and adding your comments (even if I didn't understand some of the really technical stuff).

[pnodrog]

Generally bright scenes help improve "IQ" in that the ISO can be reduced (less noise and better dynamic range), the aperture decreased (better DOF and lens performance) and shutter speed increased reducing the likelihood of camera shake etc.

P.S. The EXIF date indicates that you were focused very close and the f11 and relatively wide angle may have saved the photograph from being completely soft.

[Colin Southern]

Thank you for bringing this back to basics Colin - it is appreciated!

You're welcome.

But let me rephrase...are you saying that at the time of capture I should have dialed in 3 stops less exposure and lightened the shadows (which would now be darker) or used fill light in post?

Basically, the brightest portion of the composition is what dictates the exposure. I'd suggest turning on "blinkies" (to indicate areas of over-exposure) and then review the image (for blinkies) and the histogram (to show the degree of any under-exposure) after you've taken the shot. If you have blinkies in any important areas (indicating possible -> probably loss of detail) then it's time to dial in some negative EC (exposure compensation).

Also - don't be afraid to shoot a bracket of stops (say, 5, 1 stop apart) - then you can choose the best exposure to work with in the comfort of your own home rather than making a crucial decision in the field in the heat of battle.

But to answer the question - on a bright day (which isn't usually the best time of day for quality landscape shooting) you'll normally have to dial in some degree of compensation -- the exact amount depends on a number of factors though. But don't worry about "the factors" - results are all that matter.

I did not manipulate this area at all before posting - what you see above is untouched with the exception of the adjustments you see in the ACR screenshot. Remember I left the curves at linear and no sharpening for the sake of showing the image here.

Has the image not been processed by HDR Effects?

I did however use a Hoya circular polarizer that I believe is causing the 'dulling down' of the sky as you see presented. Is this a poor quality filter, did I use it incorrectly or inappropriately?

In my opinion, CP filters are a waste of time and money; all they'll do in a shot like this is cost you a couple of stops of light, and if your lens is wide enough, give you a very uneven sky. Yes - they can saturate the sky nicely, but so does under-exposing by a stop. I have 2 very expensive Heliopan CP filters ... and neither of them ever sees the light of day. Literally.

If I am correct I lost a stop+ just by mounting the filter

Yes

and if I understand everything above (barely I think) then that would also deepen shadows that already existed in the scene and would make it even harder to recover them in post?

No. It's like putting sunglasses in front of the lens - on one hand it cuts the light entering the camera, but the camera metering will compensate by keeping the shutter open for longer (A mode) or using a wider aperture (S Mode). If you had a fixed aperture AND shutterspeed and used AutoISO then yes it would cost you a couple of stops of DR.

The next question that I have is a bit of a tangent and related to IQ for a scene like this. As shot and for the sake of this discussion, the image seems a bit soft at 100% (again with no adjustments as mentioned above)

What sharpening are you applying? If you're not going through an optimal sharpening workflow then it's normal for an image to look soft. I wrote a little about sharpening here:

Sharpening and Noise Reduction Sequence

I have been using tonal contrast in the NIK collection and usually like the result. In this case I dialed it back quite a bit from the standard presets but it seemed like what ever I did I got artifacts around the edges of the cliff and the image just got really noisy. Is this the filter, the lens, my focus point, wind moving my tripod, the sea spray from the waves or???? Sharpening of any kind (pre or post for web presentation) made it even worse!

Sorry, can't say without seeing original RAW files -- not a big fan of NIK products to be honest -- not sure what they might have done to the image.

The only other thing that I can think of is that a scene like this one with a high dynamic range can create issues with IQ in general? What are the challenges to keep in mind beyond shadows and clipped highlights? Then how to deal with them pretty please

But that's not a scene with a high dynamic range - that's just a normal everyday dynamic range. The camera (with it's 14 stop DR capability) can capture it easily - but - remember that reflective objects (including photographic prints) can only give us around 4 stops of DR. Monitors with their active light source give us about 6 - so a scene with 10 stops captured isn't going to display correctly on a 4 stop DR media or 6 stop DR monitor. The solution is "compress the dynamic range in a non-linear manner". Big & fancy sounding words for "move the fill light and brightness sliders until it looks good"!

The white jacket is off and I am about to remove the dunce cap (but I will keep it close by just in case )

If it's one of those with a propeller on top, can I have it? (always wanted one of those -- I could wear it when flying RC helicopters! ).

Speaking of RC helicopters ...

Alan Szabo is in Hawaii this weekend at some kind of RC helicopter event -- if you can get to it, I promise you won't be disappointed. The guys flying is absolutely insane. Here's a demo:

Sharpening and Noise Reduction Sequence

[ShaneS]

Thank you for the clarifications. I will put the CP away then and concentrate on getting the image right in camera and continue to work on my PP skills.

Has the image not been processed by HDR Effects?

I was referring to the second image that I posted in the thread but no matter as you answered my question.

I will also seek out better light for my practice sessions in the future.

I'm off to go read the sharpening link.

We are having the biggest little airshow in Hawaii this weekend and I think this is the event you are referring to:

http://www.pacificaviationmuseum.org/bigairshow

The only reason that I might be tempted to go is Ford Island (the location) is not usually open to the public as it is a military base but I suspect that they won't let visitors stray too far especially with camera in hand

I certainly would not even attempt to capture images of things that move that fast - I mean look what I can do to a static landscape

[Colin Southern]

That looks like the place. Darn - wish I could go

Go on have a go at shooting them - just use longest lens - shoot wide open - and aim for about 1/250th whilst panning - piece of cake!

I'll give you a professional retouch of a good shot if you do <-- bribery!

[ShaneS]

Piece of cake...yeah right! My longest lens is only 200m - they would be out of focus specks or streaks that no amount of professional retouching could help

[Donald]

Piece of cake...yeah right! My longest lens is only 200m - they would be out of focus specks or streaks that no amount of professional retouching could help

Don't assume so. You're likely to be able to do quite a lot with a 200mm lens at an event like that.

[Colin Southern]

Don't assume so. You're likely to be able to do quite a lot with a 200mm lens at an event like that.

Absolutely. 200mm is all I'd take. If Alan Szabo is there with the T-Rex 800 then that thing has about a 6 foot rotor span. The RC jets will be even bigger.

Not as big as full scale, but they'll be flying a lot closer to compensate.

[ajohnw]

There is an interesting example of an HDR image here part way down the page that gives some idea of the exposure variations that might be used

http://en.wikipedia.org/wiki/High_dynamic_range_imaging

There are a couple of things to note. The exposures give one good one of the main subject and it's well over or under exposed in the others. Depends on the package used but this can help the software get it right. The usual problem is shown by the other buildings in the shot. They show well in 2 exposures and have different lighting levels in each - that often results in the output being what it is here - somewhere between the two.

The next question that I have is a bit of a tangent and related to IQ for a scene like this. As shot and for the sake of this discussion, the image seems a bit soft at 100% (again with no adjustments as mentioned above)



I have been using tonal contrast in the NIK collection and usually like the result. In this case I dialed it back quite a bit from the standard presets but it seemed like what ever I did I got artifacts around the edges of the cliff and the image just got really noisy. Is this the filter, the lens, my focus point, wind moving my tripod, the sea spray from the waves or???? Sharpening of any kind (pre or post for web presentation) made it even worse!

The only other thing that I can think of is that a scene like this one with a high dynamic range can create issues with IQ in general? What are the challenges to keep in mind beyond shadows and clipped highlights? Then how to deal with them pretty please

The edge effects are probably down to a high contrast boundary between the cliff and the sea. Various types of filters make use of blurred layers and this is what can cause the problem. If it's caused by a plugin the best answer is probably to find out how to DIY or use it locally one way or another. I use the GIMP a lot and a number of it's plugins cause the same problem. What I usually do if I must use one is run a blur brush round the problem in this case smearing the sea into the problem on the full sized image before it's reduced. The sky can cause similar problems.

If you ever want to post a nef file this site is useful. Any size, retained for some time and no sign up or anything else

http://filebin.net/

John
-

[Mark von Kanel]

The solution is "compress the dynamic range in a non-linear manner". Big & fancy sounding words for "move the fill light and brightness sliders until it looks good"!

Now there is language that i understand....

[Mike Buckley]

I'm confident Colin was unintentionally exaggerating when he mentioned that in a shot like this the only thing a circular polarizer will do is cost stops of light. In a shot like this a circular polarizer will help make the clouds pop out from the blue background and it will reduce or eliminate unsightly glare on the rocks and water. In other similar photos that include foliage, it will also reduce unsightly glare on the foliage.

Though he doesn't like to use circular polarizers, there are many prize-winning photographers making a living at making landscapes that recommend using them. To each his own.

[Colin Southern]

I'm confident Colin was unintentionally exaggerating when he mentioned that in a shot like this the only thing a circular polarizer will do is cost stops of light. In a shot like this a circular polarizer will help make the clouds pop out from the blue background and it will reduce or eliminate unsightly glare on the rocks and water. In other similar photos that include foliage, it will also reduce unsightly glare on the foliage.

Though he doesn't like to use circular polarizers, there are many prize-winning photographers making a living at making landscapes that recommend using them. To each his own.

I'm afraid that your confidence is somewhat misplaced Mike.

As I see it, CP filters are a tool - like high ISO modes - like flashes - like tripods. And like any tool, one really needs to know when to use it - when NOT to use it - and when it's just a waste of time using it. The issue I have with CP filters is that people are quick to rattle off exactly what you did (clouds popping, reducing/eliminating glare etc) - but that's where they leave it. People pick up on that and think "OK - sounds like a good thing - I'd better put one on". Unfortunately, that's not all there is to it;

- Yes - they can appear to enhance "cloud pop" - but - that effect is more noticeable looking through the viewfinder than it is on the captured image. It's my belief that it comes more from assisting the human eye through the light attenuation -- regardless, in my experience anyway, it never seems to translate through to the captured image. What DOES make a huge difference though is using clarity along with some appropriate hand-dodging (set to midtones or shadows). So a CP for cloud pop doesn't fly in my opinion - I'd rather do that in post where I have more control over the effects.

- Reducing glare. They do - they absolutely do - no doubt about it - but - ONLY when the light is coming from an appropriate direction. If you're wanting to "show the fish that you're feeding in the water" then yes - absolutely it'll help, but we're not trying to see fish underwater here - we're shooting landscape. And if anyone is shooting landscape at a time of day where a CP filter is going to help them then they're shooting it at the wrong time of day (most of the time). Shane's photo is a good example of that; it's shot with a CP filter at the wrong time of day -- shoot it without a CP filter at the right time of day and it'll improve tenfold. Shoot it with a CP filter at the right time of day and it'll improve tenfold too - so long as the loss of 2 stops doesn't complicate anything too much. Translation: "shoot it at the right time of day and the CP won't add anything - only potentially take away)

- What people often fail to mention is that landscape is usually (rightly or wrongly) associated more with wider fields of view - and that CP filters will give a ridiculously uneven sky that effectively ruins the shot.

- What they also fail to mention is that using a CP filter often makes it difficult or impossible to work with other filters at the same time - ones that may be far more significant (eg ND or GND)

So that's my beef with them. People are too quick to parrot off the "advantages" but very slow to point out either the disadvantages or the fact that if people are "needing" them then they're probably shooting at the wrong time of day. In reality I suspect that most people simply regurgitate what they've heard from others without really understanding that the practice is considerably different from their sound-good-in-theory mantras. It's not about whether "prize-winning professionals" recommend them (although for what it's worth, my customers vote with their wallets and I have my work hanging on walls around the country and even in some places overseas) - it's about explaining what the tool is - how it works - when to use it - when NOT to use it. Education is considerably more powerful than someone saying "I'm an award winning professional and I recommend it".