HDR Efex Pro?

[Colin Southern]

sitting here recovering a crew members screwed HDD, (no backup naturally) so im on the edge of my chair.... doesnt dynamic range alter with iso?

Yes - DR is more or less inversely proportional to ISO.

[Colin Southern]

I had to do some internet research on this one...which unto itself is confusing but I am getting 12stops from this link:

http://www.imaging-resource.com/PROD...000IMATEST.HTM

It's actually 13.9 according to DxOMark (who are pretty much the gold standard these days), but 12 or 13.9 - close enough.

I'm guessing that you will ask me what the dynamic range of this scene is? Based on the histogram it looks 'normal' to me so 10 stops because nothing is clipped on either end? Please educate me.

Patience grasshopper

OK - what do you think the dynamic range of a normal reflective scene is? - ie something white next to something black - both in the sun (Think "Bride in white dress & groom in black suit").

[ShaneS]

I have no idea Colin...she pulls out dunce cap and spins wheely thing on top while heading to the corner...

[Donald]

she pulls out dunce cap and spins wheely thing on top while heading to the corner...

Don't do that. The corner's already too crowded with people who don't know the answer to that one. But stick with this. This is the sort of tutorial that you could pay a lot of money for in a lot of places.

[IzzieK]

I have no idea Colin...she pulls out dunce cap and spins wheely thing on top while heading to the corner...

hahahaha...Shane, I am learning from this post too so keep asking. I do not use HDR because most of the images that I had seen of them are too much yucky looking, but from the examples of some others in this forum on HDR processing, it is not as bad as I have seen them on others. I am waiting with bated breathe on how this post will turn out. Colin's image with the house he used to remember when he was young is a good example of processing I like. I am in the thought that he is guiding you through this. We will all learn from his technique...despite your mistake in other areas of how you took this shot. Love you, Shane, for taking this up.

[Colin Southern]

I have no idea Colin...she pulls out dunce cap and spins wheely thing on top while heading to the corner...

Know worries (as we say)

It's 4 stops.

OK - next question: Behind the Groom - in his shadow - there's a garden with an assortment of plants surrounded by bark. Since this garden is in shadow (and thus not getting the same amount of incident light as the groom), about how many more stops of dynamic range (over and above the 4 we've put aside for the bride and groom) might we need? (tougher question, but others can chip in here too) (hint: I've covered this several times in the past )

[george013]

Collin,
That difference between the dynamic range the sensor can handle and the dynamic range of a normal reflective scene, what causes that? Is that a practical value or because white doesn't reflect 100% and black doesn't absorb 100%?
George

[ShaneS]

She spins wheely thing again...and, regretfully, has to go to bed as it's getting late in Hawaii and I need to get my beauty rest before I head for work tomorrow.

I will check back first thing in the AM to see the answer and the following question(s)...

I hope that others do feel free to chime in

[Colin Southern]

Collin,
That difference between the dynamic range the sensor can handle and the dynamic range of a normal reflective scene, what causes that? Is that a practical value or because white doesn't reflect 100% and black doesn't absorb 100%?
George

Georrge,

What the sensor can handle really doesn't enter into it for a purely reflective scene (at low and medium ISOs anyway) - most sensors are in the 12 to 14 stop range at base ISO now - and obviously 4 fits into 12-14 with the greatest of ease - even with big safety margins.

The black and white reflectance thing is quite interesting though - especially black. We often refer to things as "being black" (eg "black cat"), but in reality, if they really were black (ie they absorbed 100.000% of light hitting them) then we wouldn't be able to see them - only an outline where the things they are in front of disappear. So from a photographic point, most "black" things actually reflect quite a bit of light - a couple of "case in points" ...

- in the studio, it's quite common to have a white background, but in reality, that's actually a function only of how much light reflects off it relative to what's illuminating the subject. With white paper and the same amount of illumination that's hitting the subject (note: same AMOUNT of light - not from the same light) (remember inverse square law for light fall-off) you'll get a white background. Providing that there's no cross contamination from key & fill lights, if you drop the background illuminating lights by 2 stops you'll have a middle gray background (2 stops down is 1/2 of the 4 stops between "black" and "white"). Drop it another 2 stops from the original "white producing" and you'll - in essence - have a black background. BUT - the interesting thing is ... it actually works the other way too ...

... in my studio I have one heck of a lot of "fire power" (somewhere around 6,000,000 watts instantaneous) - and if I point a couple of high-power strobes at a black background - and hit it with 16 times (4 stops) the power that I'd use for a white background ... I get a white background. In practice it's a bit silly to do it that way, but in reality, if I need to raise a black backdrop to a deep gray background then adding a bit of light is exactly what I do (one could go the other way and take 3 to 3 1/2 stops off a white background, but eliminating contamination from key and fill lights becomes more problematic).

- When printing. When profiling paper - on a scale of 0 to 100 (LAB colour) the white point of canvas is typically around 93 to 95 whilst the black point (with the wrong ink) is typically around 23 to 25 (dropping to around 17 after applying the overspray). Paper on the other hand gets down into the 4 to 8 region.

So having said all that, the 4 stops thing is a solid rule of thumb in nature; you won't get naturally occurring things reflecting 100% of light, and you won't get naturally occurring things absorbing close to 100% either. Where things DO get closer to absorbing 100% (eg black velvet & matt inks) it doesn't really enter into the equation photographically because generally we don't WANT to capture detail from those types of objects. In other words - in nature - SO LONG AS IT'S ALL RECEIVING THE SAME INCIDENT LIGHT (ie not in the shadow of another object) (or not receiving incident light) then 4 stops is about all the variation you get. I think that surprises a lot of people.

[Colin Southern]

She spins wheely thing again...and, regretfully, has to go to bed as it's getting late in Hawaii and I need to get my beauty rest before I head for work tomorrow.

I will check back first thing in the AM to see the answer and the following question(s)...

I hope that others do feel free to chime in

Nah - they won't (they're all chickens!). In reality, it's around about 3 stops (it varies, but that's a pretty typical amount). OK - next question: "if a reflective scene is being lit by light reflecting off clouds and coming from the sky in general - but not shooting into the sun itself - about how many stops do we need to allow over and above the exposure for our reflective scene so that things like the sky and clouds don't blow out"?

[pnodrog]

Nah - they won't (they're all chickens!). In reality, it's around about 3 stops (it varies, but that's a pretty typical amount). OK - next question: "if a reflective scene is being lit by light reflecting off clouds and coming from the sky in general - but not shooting into the sun itself - about how many stops do we need to allow over and above the exposure for our reflective scene so that things like the sky and clouds don't blow out"?

A chicken might say that provided the brightest reflecting surface was greater than 50% of the incident light then another stop would probably do....

[Mark von Kanel]

I have no idea Colin...she pulls out dunce cap and spins wheely thing on top while heading to the corner...

+ 1 i thought you were a he

[Mike Buckley]

"if a reflective scene is being lit by light reflecting off clouds and coming from the sky in general - but not shooting into the sun itself - about how many stops do we need to allow over and above the exposure for our reflective scene so that things like the sky and clouds don't blow out"?

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).

[ajohnw]

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'd go for a comment on a shot I posted once - more than the usual 7 stop total - as I assume it was unusual.

John
-

[Mark von Kanel]

Nah - they won't (they're all chickens!). In reality, it's around about 3 stops (it varies, but that's a pretty typical amount). OK - next question: "if a reflective scene is being lit by light reflecting off clouds and coming from the sky in general - but not shooting into the sun itself - about how many stops do we need to allow over and above the exposure for our reflective scene so that things like the sky and clouds don't blow out"?

well if youve got a dynamic range of 12 stops then why would you need any extra? or am i missing something here? (spinning my round disk, in the corner with Shane)

[ajohnw]

well if youve got a dynamic range of 12 stops then why would you need any extra? or am i missing something here? (spinning my round disk, in the corner with Shane)

There isn't a straight forwards answer but look at what the 12 stops comprise of. They go 0,1,2,4,16,32 etc light levels up to 4095 if 12. There is clearly not much chance of seeing any gradation between 0 and 1, more between 4 and 16 and even more between 16 and 32. That has implications when a shot is PP'd. Talking loosely think what happens if the 0-1 is made brighter. It had no gradation. If 4-16 is brightened it does have gradation so will have some contrast.

The other problem is when it's viewed - 8 stops on the same basis but usually reckoned to be maybe 6 in practice on a display due to their limitations. PP is aimed at getting the input to display well within this dynamic range. A straight standard raw conversion will generally leave darker areas much more dark than they appeared to be when the shot was taken. This usually means that shadows will be far too dense. If on the other hand the dynamic range of a scene fits nicely into the brighter region of the cameras all is well

We see equal changes in light levels in a similar way. Each equal change has double the intensity of the previous one but need certain levels and gradation rates to see them. Not so bad when looking around a scene as our eyes adapt but we don't do that when we view a shot at we see it in it's entirety which can make that aspect worse.

I'm afraid I am one of those people who think stinking fish when some states that a camera has a real dynamic range of 12 or 14 EV when we actually get to look at it. The numbers are too small at the dark end which is why they can only be brightened up so much before artefacts appear. This is where HDR techniques come in. Simply put they move areas of a shot from the dark end into the well lit tonal range. Another shot may be used to move high lights in much the same way if they would be clipped in other exposures.

John
-

[george013]

........
So having said all that, the 4 stops thing is a solid rule of thumb in nature; you won't get naturally occurring things reflecting 100% of light, and you won't get naturally occurring things absorbing close to 100% either. Where things DO get closer to absorbing 100% (eg black velvet & matt inks) it doesn't really enter into the equation photographically because generally we don't WANT to capture detail from those types of objects. In other words - in nature - SO LONG AS IT'S ALL RECEIVING THE SAME INCIDENT LIGHT (ie not in the shadow of another object) (or not receiving incident light) then 4 stops is about all the variation you get. I think that surprises a lot of people.

Given a white and a black body. Let's say white is reflecting 90% en black is reflecting 10% of the light. So the difference in reflected light is 80% and under normal condition that will cover 4 stops. I don't know if the numbers are realistic.

George

[Mark von Kanel]

why do we have to analise it to bits? isnt the question, if we expose for the dress how many stops do we need to compensate to stop clipping the sky? if so i dont think the sky would clip and it could be easily recovered from the one exposure..... (still spinning the disk in the corner)

[ShaneS]

Count me as surprised

I think that surprises a lot of people.

This is all quite fascinating and I am starting to see more clearly the impact that some of the exposure choices that we make have on our ability to PP an image to "what we saw".

The incident light thing is confusing me though...will search Google when I have a moment at work today and will continue to watch and participate in this thread as the day progresses.

[ajohnw]

why do we have to analise it to bits? isnt the question, if we expose for the dress how many stops do we need to compensate to stop clipping the sky? if so i dont think the sky would clip and it could be easily recovered from the one exposure..... (still spinning the disk in the corner)

Because the light levels are recorded in bits in the way I described. and have the problem that this isn't logarithmic which is how we actually see things. It doesn't really matter how these bits relate to real EV's - the main point is that as things get darker the gradation that can be achieved gets coarser and that has implications during PP.

Personally I feel this type of discussion is only of use in terms of getting a grasp on what it going on and doesn't offer a fix for the problems. It's very easy as it turns out to go forth with camera and shoot for unclipped highlight especially when there is a nice sunny cloudy sky and see for yourself what happens at the dark end when the scene includes shadows of varying density. Then it's also easy to see what can be recovered to give a more natural look to a shot with PP - providing the techniques that can be used are understood. The problem there is more easily understood using bits. Say the 4 to 8 stop is brightened to a 16-32 level. If done litterallt the 4-8 runs from 16-24 and not 32 so needs a contrast increase. A very very simple way of looking at it again. Perhaps this might help. You can compare the 0->1 and 1->2 stops on a screen. Smile as you are unlikely to see any difference. This aspect is another example of the same problem. Don't take it to much to heart though as colour variation would make a large difference.



John
-