D300 - Built-in or SB900 Underexposed

Joined
Jun 13, 2010
Messages
11
Location
Edinburgh
Can anyone help. I've just bought the kit so am still finding my way around.

With both the Built-in flash and the SB900 (TTL) ( used separately) I am getting consistently underexposed pictures. i.e. taking pictures in a normally lit living room.

I've web-searched this one only to find one article effectively saying that it was normal for the D300. I do find this difficult to believe.

Increasing EV value in the camera by 1.0 cures the problem but this is an irritation I can do without.

Any suggestions ?
 
Joined
Apr 3, 2006
Messages
5,614
Location
Texas
It is not "normal" for a D300, but it is not unusual for the TTL system - that up to +1 EV flash compensation might sometimes needed - sometimes. Maybe more so for off camera flash than for hot shoe flash.

You probably will want to use Flash Compensation (button on other side of viewfinder), instead of Exposure Compensation (near shutter button). Flash Compensation only affects flash, but Exposure Compensation affects both flash and ambient light.

This compensation is not a constant for all situations - it is NOT correcting the hardware. It is instead dependent on the scene metered in front of the camera. It will vary in different settings, depending on the scene. It is definitely NOT a "set it one time and forget it" type of thing.

Also note that flash intensity falls off fast with distance. Flash exposure can only be correct at some ONE distance. So if your subject is at say 8 feet, and properly exposed there by flash, then anything at 11 feet will be one stop under exposed, and anything at 16 feet will be two stops underexposed. We must understand and accept that. Except --- Using bounce flash indoors will substantially aid this distance falloff problem (evens things out in a regular size room - but it has limits for greater distances).
 
Joined
Jun 13, 2010
Messages
11
Location
Edinburgh
Scott - This was in an artificially lit living room, no daylight input. This was me in dumm point-and-shoot mode, flash on the hot shoe, turn everything on and shoot. I naturally expected the same result as I might have got from my little point and shoot i.e. a perfectly exposed picture but alas not.

Wayne - I think you're saying that this dumm mode is not an option but I don't think I'm expecting too much. The subject was the contents of a table against a wall at constant distance from the camera, so no variable distance objects to fool things.

twistedlogic - D Lighting was off.

Thanks to all three of you.
 
Joined
Apr 3, 2006
Messages
5,614
Location
Texas
Wayne - I think you're saying that this dumm mode is not an option but I don't think I'm expecting too much. The subject was the contents of a table against a wall at constant distance from the camera, so no variable distance objects to fool things.

So this wall was white? Or similar light wall color? And that wall was probably no doubt a significant part of the subject, which the TTL light meter evaluates?

Here are two quicky flash pictures, both alike:

SB-800 flash, hot shoe, TTL mode, direct flash (I normally use bounce, or an umbrella, at every chance).
D300 camera, manual mode, f/8 at 1/200 second, matrix metering (I normally use Center).
No post processing and no corrective steps.

dse_4203.jpg
Subscribe to see EXIF info for this image (if available)


dse_4204.jpg
Subscribe to see EXIF info for this image (if available)


Both are the picture of an Alienbees flash. Only difference is that one background is a white umbrella, and one background is a black umbrella (back side). This is intentionally to show extreme effects, the more typical scenes are not this bad, but yes, this is simply how it works Jeff.

The umbrella really is white, bright white, and about 18 inches behind the rear of this flash. Camera and flash distance 18 or 20 feet (105 mm lens), so the 18 inches falloff should not have caused much effect.

However, the single goal of every reflective light meter is to make the scene average out to middle tones, on average. And this is often correct... the distant dark mountain, the blue sky, the white clouds, the green trees/grass. It often does average out to a middle tone overall (in many average scenes), so the meter is often right.

But to do that here, this white "wall" caused the TTL meter to cut back on the flash power, seeking middle tones, and so this picture of a AB subject appears badly underexposed. And it is... it needs human help - Flash Compensation.

However, the meter thought it got it just right, by its rules. If you look at the histogram (PhotoMe is a great browser tool), you see the prominent large spike a bit below center. Unfortunately, the white should have been at the right end of histogram. This simply is NOT a typical subject. It was my chosen subject however, and it needed my help.


The black picture simply changed the background, and yes, technically it is a different flash unit now, but all else is close enough to the same, believe me. Now the flash provided quite a bit more light, trying to reach middle ground - and it failed at this distance - Ready LED was blinking insufficient power. It wanted to get the black up more middle ground, and it would have if it had sufficient power. I could have planned that better. :smile: I could have opened the aperture or increased the ISO or moved closer. I didn't.

You can see the different flash power, on the subject of course, and also on the far room walls in the corners. Black pumped the power up, all it could go.


Humans have brains, and can immediately recognize the subject, and the situation, and the result, and they know how bright it should be... humans know what it should look like. Dumb computers have absolutely no clue, not about nuthin'. It does not know if it is white that needs more light, or black that needs less light. All it can do is to try to make every scene average out to middle gray.

The reflective meters goal is simply to make every scene average out to be middle tones, overall average. I like to say middle gray, but it may be blue or green or whatever. And I do NOT mean center of the histogram... normally slightly lower is the idea.

Beginners ALWAYS expect their camera to do everything for them. And it will try hard and it will do a lot, and it can often get close, but photography simply needs a human brain to help the camera a little bit. We call this experience. :smile:

We can bemoan our fate, and cuss the camera, but it works so much better to simply learn the first few things, and pitch in and help a little. :smile: For every TTL flash picture, yes, pay attention to results, and consider necessary Flash Compensation. As required to get the picture you want. You are the photographer.

There are both reflective light meters and incident light meters available. Incident directly meters the light falling on the subject, and never sees the subject, and so the reflectivity of the subject never affects incident meters. They are generally much closer to right in all situations... however, they cannot be built into cameras.... they must meter at the subject.

You might see the light meter link below.

Meanwhile, don't despair. Just pitch in and help it a little, with Flash Compensation. This is not a constant, it will vary in every scene in front of the camera. Yes, it is how it works. It has always been this way. And in fact, it is pretty easy. Simply just do what you see you need to do. Soon, you will just already have a pretty good idea what you will need to do, and can do it first.
 
Last edited by a moderator:
Joined
Jun 13, 2010
Messages
11
Location
Edinburgh
Wayne

Again thanks for your detailed reply.

I understand what you are saying but I don't think this applies in my case. I've posted 3 pictures at

http://jeffedin.cwahi.net/

They are pretty self-explanatory.

First - D300 (EV = 0.0) + SB900 (EV = 0.0) The histogram is skewed well left indicating underexposure.

Second - D300 (EV = 0.0) + SB900 (EV = 1.0) Life improving but still a histogram skewed left

Third - Point and Shoot on its dummest setting (Auto) and perfect, histogram spot on. This is what I'm expecting for the first, or am I missing the plot.

Jeff
 
Joined
Mar 15, 2010
Messages
516
Location
Seattle
Try it in multiple shooting locations to determine if you really think it is the camera.

I had a lens that I had to set EV to +1 to match the results from my other lenses on the same camera (d300), so, maybe your lens is causing underexposed images.

If the exposure seems to be right in other situations, then I'd guess it's just this spot that is tricking the metering for whatever reason.

At any rate, I'd argue this is the point of learning the hobby. It's fun to control the exposure (however one chooses to do that) and get the picture you're looking for.
 
Joined
Apr 3, 2006
Messages
5,614
Location
Texas
This is what I'm expecting for the first, or am I missing the plot.

Maybe you have a magic compact Jeff. Hang on to it. :smile: Any chance it has been imported into PC software that automatically does automatic adjustment on it? (slightly clipped white point suggests it?) It is a little hot, but you can always crank up D300 Flash Compensation to match it.

I do think it is false expectations. Clayton is right, you should try more situations to better understand how it works... All situations are not equal.

The dark colors (black, or any less reflective colors, etc) reflect less of the light that is hitting them.

The light colors (white, or any more reflective colors, etc) reflect more of the light that is hitting them.

White may easily reflect 90% of the light. Black may do well to reflect 10% of it. Stuff in between is in between. Just how life is.

Then when the same light hits white, it reflects more brightly than when the same light hits black. The light meter naturally reads higher then, and reduces the flash power accordingly (underexposure of white, but just right exposure of a middle gray average). As photographers, we are expected to understand how this works, and react accordingly too.

Reflective meters simply read this reflected light. They have no advantage of human "smarts", they have no clue what it is they are seeing, or what it should look like. If you are expecting miracles from that, you may be expecting too much.

I am trying to say... what the reflective meter reads depends entirely on how much light the subject reflects. Different subjects and scenes are simply different (in this metered respect). Yes, also depends on how much light is hitting that subject, but all that really counts is how much is reflected, from subject to meter. What the meter sees. It is all the meter knows.


These below are all automatic flash pictures, of same subject in exactly same situation.... just different cameras.
No automatic import adjustments.


Here is a D300 with internal TTL flash picture of two stacked sheets of paper, black paper and white craft paper - to show that they really are white and black.

dse_4292.jpg
Subscribe to see EXIF info for this image (if available)


And to show that the two together more nearly ought to average middle gray, so they come out better, since the meter is going to make the average be middle gray. If the average is made to be middle gray, the white things will be brighter, and the black things will be darker.

My guess is that you expect the white paper to come out white, and the black paper to come out black, because any human can see that they simply are... But here is what happens:

Here is just the white paper (D300) - expecting underexposure.

dse_4290.jpg
Subscribe to see EXIF info for this image (if available)


Here is just the black paper (expecting overexposure).

dse_4291.jpg
Subscribe to see EXIF info for this image (if available)



Both alone become more middle gray, because that is what reflective meters do. (a more average scene might average out about right, much of the time, but all scenes are not average).

You can do these tests too, to become a believer.


Here is a little Canon point&shoot, model SD770 IS, a very decent little compact for the wife, it works well for her. This is in green Auto mode: (same white, and then black, paper as above)

img_2236.jpg
Subscribe to see EXIF info for this image (if available)


img_2237.jpg
Subscribe to see EXIF info for this image (if available)



Note both from compact camera are middle gray... As expected... That is how reflected light meters work.



Here is an Apple iPhone 4, same white and then black paper: (it is the least capable camera of course)

img_0074.jpg
Subscribe to see EXIF info for this image (if available)


img_0075.jpg
Subscribe to see EXIF info for this image (if available)



A different shade of gray, an LED flash, but both are gray, neither is white or black.

This is how reflective light meters work - and it is good to learn that too. :smile: We know what to expect, and can stand ready to apply Flash Compensation, on cameras that provide it.


These were very quickly done, first pass. The only wonderment above is why the D300 did not more nearly give two middle gray exposures, more nearly equal. My previous tries always did. Not sure what I did wrong here. :smile:


The Exif data is in these, if you want to check.


You should repeat these tests, to believe it, to make the knowledge yours. Then you will have a very strong start on understanding what the camera light meter is telling you, in the different situations you will encounter. That is a big part of what photography is.
 
Last edited by a moderator:
Joined
Jan 11, 2009
Messages
3,409
Location
Tokyo, Japan
Hi Jeff,

Quick question: Is your computer monitor calibrated?

I can't see the histograms, but from what I see of your three images on my calibrated LCD monitor, the first image looks about right (maybe a bit underexposed), the second looks slightly overexposed, and the last taken with the P&S looks overexposed.

FWIW,
Isamu

Wayne

Again thanks for your detailed reply.

I understand what you are saying but I don't think this applies in my case. I've posted 3 pictures at

http://jeffedin.cwahi.net/

They are pretty self-explanatory.

First - D300 (EV = 0.0) + SB900 (EV = 0.0) The histogram is skewed well left indicating underexposure.

Second - D300 (EV = 0.0) + SB900 (EV = 1.0) Life improving but still a histogram skewed left

Third - Point and Shoot on its dummest setting (Auto) and perfect, histogram spot on. This is what I'm expecting for the first, or am I missing the plot.

Jeff
 
Joined
Apr 27, 2008
Messages
186
Location
London UK
What a great thread with some excellent advice. Another thing to note is that in a small way we're comparing apples to oranges. A D300 is a perfectly serviceable professional camera (I use it as such on a regular basis) and it's designed to be used by photographers who generally wish to retain control over their shots i.e. we're not that keen on the camera running off and making too many decisions on it's own. Hence why the D300 doesn't have 'scene modes' such as 'landscape' and 'sport' etc. So it's probably producing a faithful representation of the scene (bearing in mind WayneF's excellent explanation about light meter evaluations).

You're point and shoot is most likely designed for a photographer that isn't looking to be that involved in the creative process and is happy for the camera to start making decisions about what the best shot is going to be. The point and shoot might well be doing plenty of it's own post processing on-board. Of course I'm making an assumption about your point and shoot that may be off the mark but it's maybe worth considering.

As IsamuM mentions it might be down to a calibrated monitor as on mine the first image looks correct while two and three do look over exposed.
 
Joined
Jun 13, 2010
Messages
11
Location
Edinburgh
Again thanks to everyone for all your excellent contributions, especially to Wayne for his ever-complete explanations.

A huge thank you to Isamu whose mind must be on more pressing matters. Our hearts go out to you and your country at this difficult tragic time.

I've added the associated histograms. Greg's drift is an interesting one. Unlike all of you I know the ambient light levels that the pictures were taken with. They were quite dim with only two wall lights on. The camera has actually delivered what my eyes saw. Picture 1 is a bit underexposed (still grumping) and picture 2 is a little over. However the reality of the scene has not been tampered with.

Now the point and shoot. Wayne says -
"Maybe you have a magic compact Jeff. Hang on to it. Any chance it has been imported into PC software that automatically does automatic adjustment on it?"

I think that that is close to the mark. It delivers what average Joe Consumer wants - a nice bright picture as if it had been taken under a Meditteranean sun. However it's not what's sitting in front of you. You can achieve that with retouching or ramping up the flash output. But don't automatically do it because I might not want that.

I wonder if I have now got this right ??
 
Joined
Apr 3, 2006
Messages
5,614
Location
Texas
Well, TTL exposure is just going to vary sometimes. Often is pretty close, but sometimes is not, and my notion of the trick is that we simply stand ready with Flash Compensation to get the picture we want. Which is vastly easier to accomplish in digital than back in the film days. :smile:

The reflective meter reading depends on the scene in front of the camera. The meter simply exposes according the amount of light reflected from the subject. Different subjects simply vary in their reflectivity. This fools the reflective meter, instead of helping it. The meter is a dumb computer with no clue what is, or what it should look like. It simply has zero human recognition. The only goal of a reflective meter is to try to make all pictures come out to the same middle gray average (specifically, the metered area of the picture). My white and black paper is simply an intentional extreme example of this, but all scenes do this in lesser degree.

You cannot judge this by just the one first picture, you need to see more situations, for more experience with it.

FWIW, my mistake before when the D300 internal flash did not give strong enough (expected) "same middle gray" result from white and black, was because the internal flash is always in TTL BL mode (I am more used to hot shoe flash in TTL mode).

Here is a comparison using Spot Metering to force the internal flash into TTL mode.

internal3.jpg
Subscribe to see EXIF info for this image (if available)



This was in a room dim for even indoor standards... daytime, but very cloudy outside. So the TTL BL difference surprised me here, the ambient was several stops down, and I assumed it was insignificant. Perhaps it was the lens distance info, but it did bias the result. This same thing happens in TTL BL mode on the hot shoe flash too. That is a different subject.

My cynical philosophy, but automation like TTL BL or Matrix metering is sometimes helpful, but is not very understandable to predict... some dumb computer doing unknown things in there that neither we or it understands. :smile:

Which ever mode is used, if we want white to appear white, and black to appear black, depending on how dominant this is, we need to be ready to use Flash Compensation.

Again, this was indeed genuine white and black paper... the setup looked like this:

scene.jpg
Subscribe to see EXIF info for this image (if available)


The dark black background is velvet, it does not reflect light like black paper does. Also harder for limited flash power to light it up to middle gray.

To make the continued point about "this is simply how it works", note this last picture above is a better mix of light and dark areas (overall), so the actual scene's overall average is closer to middle gray (also true of most average or typical scenes), so when the reflective meter forces its average to middle gray, it comes out just about right. This is simply how reflective meters work.

By forcing the average to middle gray, I mean something like this...

scene.gif
Subscribe to see EXIF info for this image (if available)



This is the same above tabletop picture, all blurred together to create one average middle tone (RGB 133, 125, 122), using Photoshop Blur filter "Average". It creates one average tone. This method however averaged the entire picture area, where the camera meter only averages the specific metered area, depending on the rules of that metering system. So only a similar concept, but the point is, the reflective meter simply tries to average all pictures to a similar middle tone. Every time. Unless you compensate it otherwise.

This reflected exposure seems to be harder for flash than for the ambient, but in the camera, both are reflected meters, and this is how reflective meters work. Therefore, it seems extremely useful information to know.

Still, indoor hot shoe TTL and bounce will be pretty close much of the time, but we always watch the exceptions, and stand ready to do what we see we need to do. It becomes much easier pretty soon. Honest, it is not difficult.

There is lots available to read about how it works:

Search Google for reflective metering

Here is one good link from Sekonic:

http://www.sekonic.com/classroom/classroom_21.asp

See the short section on reflective, every word. Believe it, it is simply how it works. There is no concept of a precisely accurate meter reading. Reflective depends on the subject.
 
Joined
Feb 22, 2010
Messages
711
Location
Denmark
This thread should be made a "sticky" as a text book example of what Nikon Cafe is about. It was a joy to read through, thank you very much, gentlemen!

Wayne, as always very very informative posting with a lot of quotable remarks; today's favourite (almost Zen-like):

Stuff in between is in between. Just how life is.
 
Joined
Jun 13, 2010
Messages
11
Location
Edinburgh
I think all this boils down to Clayton G's comment "people have different opinions on what is properly exposed". If you add to that a consistent exposure measuring mechanism which can be tweaked in a consistent way to achieve your idea of properly exposed then that is all that is required.

In all of this discussion I have kept coming back to my Pentax KM which served me from 1976 to 2003. A simple mechanism of 2 strategically placed CdS cells, battery, variable resistors to represent different aperture & shutter speeds plus a moving coil meter. This was ground breaking stuff in the 60's. It worked though and it was consistent in measuring average exposure. Jeff always needed an extra half stop to get proper exposure. Sometimes 1 and occasionally 1.5 extra. It was very much an assistant and not the master as one might get lulled into thinking with talk of 3D this and that, databases of images etc. As Wayne has eloquently said 'machines are dumm'. They can never replace higher human thought processes so very necessary in any form of artistic expression - thank goodness.

In the end we are still artists however clever and complex the tools at our disposal.

I'm a good few steps ahead after this discussion, so a huge thanks to all.
 
Links on this page may be to our affiliates. Sales through affiliate links may benefit this site.
Nikon Cafe is a fan site and not associated with Nikon Corporation.
Forum post reactions by Twemoji: https://github.com/twitter/twemoji
Forum GIFs powered by GIPHY: https://giphy.com/
Copyright © Amin Forums, LLC
Top Bottom