Subject too dark...
- Thread starter J-see
- Start date
It is not really about WB. You have the materials to do the easy perfect experiment to see the absolute truth about how meters work.
Do a metered picture of each the cards (white, black, and gray). Three pictures, close enough to include only the one card in each view (to rule out other colors, other effects).
See? That is how reflective meters work. (incident meters are different)
(focus on a blank card is difficult, but it need not be exact focus for this... Just use manual focus).
Do a metered picture of each the cards (white, black, and gray). Three pictures, close enough to include only the one card in each view (to rule out other colors, other effects).
See? That is how reflective meters work. (incident meters are different)
(focus on a blank card is difficult, but it need not be exact focus for this... Just use manual focus).
Last edited:
I did and posted them in the reply before. I used the macro to rule out any accidental metering of something else. The first I shot both next to each other, the second and third both separate. White and black become reasonable similar. They're slightly different but there's probably a margin of error. I didn't have grey but if black and white become grey, it showed what I needed to know.
If I want to adjust exposure to make them both match pure black and white again, I have to overexpose white 1.60 in LR and underexpose black 3.60. Only then they merge into the pure version. That's the shot including both.
The separate shots of white and black need around +5 and -5 adjustment to become pure again.
If I want to adjust exposure to make them both match pure black and white again, I have to overexpose white 1.60 in LR and underexpose black 3.60. Only then they merge into the pure version. That's the shot including both.
The separate shots of white and black need around +5 and -5 adjustment to become pure again.
Last edited:
OK. I think you may still be resisting, but you'll probably get it soon. 
I would disagree, IMO the first picture is obviously better, but I'd give it a bit more exposure, not less. More white. Look at your histograms... white ought to be towards the right end.
The idea for any photographer is to figure out how their meter works, and then when they first walk up to a scene and subject with a white wall background, or a bright sky background, or one like this black bag, the alarm goes off in their head. They simply just realize what is going to happen, and what to do about it (compensation is what we do, first try before the first shot). It is called experience. It is not just about white or black, it is any more reflective or less reflective colors. The meter only sees the light that is reflected, it meters what it gets. We can sort of judge the degree of how much it can reflect, and know what to expect.
If we don't learn that, we should still know to watch our exposure results (camera rear LCD) and compensate it as we see is needed.
This always dismays beginners, who want to assume the camera ought to always get it right. But life simply ain't like that. Never has been.
Many or most scenes are average scenes, meaning they do contain a wide mix of colors, which does average out near middle gray. Your first bag picture almost does, white boxes and background. Then when the meter puts this average in the middle, it is often about right. But photographers do have to deal with the exceptions too. It is why the cameras have compensation buttons.
I would disagree, IMO the first picture is obviously better, but I'd give it a bit more exposure, not less. More white. Look at your histograms... white ought to be towards the right end.
The idea for any photographer is to figure out how their meter works, and then when they first walk up to a scene and subject with a white wall background, or a bright sky background, or one like this black bag, the alarm goes off in their head. They simply just realize what is going to happen, and what to do about it (compensation is what we do, first try before the first shot). It is called experience. It is not just about white or black, it is any more reflective or less reflective colors. The meter only sees the light that is reflected, it meters what it gets. We can sort of judge the degree of how much it can reflect, and know what to expect.
If we don't learn that, we should still know to watch our exposure results (camera rear LCD) and compensate it as we see is needed.
This always dismays beginners, who want to assume the camera ought to always get it right. But life simply ain't like that. Never has been.
Many or most scenes are average scenes, meaning they do contain a wide mix of colors, which does average out near middle gray. Your first bag picture almost does, white boxes and background. Then when the meter puts this average in the middle, it is often about right. But photographers do have to deal with the exceptions too. It is why the cameras have compensation buttons.
Maybe I have weird logic or too high expectations but when I take a shot of something, I prefer that shot to look as what is in front of me. I don't understand what's so hard to accomplish that. Technically that is. Individual visual differences aside, it's just a measurement of incoming light. Light in front of dark or dark in front of light should not matter. We could still under or overexpose as we see fit but at least we'd know what we be adjusting. Now the cam is first deciding how it is going to look and then we have to adjust that again.
Maybe what I prefer is too simple.
Maybe what I prefer is too simple.
Horoscope Fish
Senior Member
I think it's safe to say that's a fairly common sentiment... Unfortunately, and for reasons I am not qualified to explain, that's simply NOT how it is. I have a feeling it has a lot to do with our brains, a whole boatload of "data processing" it performs without us even being aware of it and that digital cameras simply can not duplicate at this point in time. That being the case we have to learn to "see", or at least understand, how the camera meter interprets the world and adjust accordingly.
.....
There's something fundamentally illogical about it. I understand this approach was practical in the days of SLR since technology wasn't what it is today. Those were the days we had to carve our mail into clay tablets but in this day and age it seems very stone-age.
Even when it uses 18% gray, or 12 or whatever, it still has to calculate the individual differences of incoming light. That does imply it can distinguish else we'd end up with not so correct shots. When throwing the whole grey thing out, it should be able to perfectly calculate and represent the scene in front of us and that scene should match what we see since we do look through that lens too. We receive the same signals and once you digitally replicate that conversion, what is out there will always match what is in your shot.
Horoscope Fish
Senior Member
If it was possible to do so a price-point that would be acceptable to the "average schmo" I think it safe to say it would have been done. I'm neither engineer, nor physicist, so I can't explain the particulars but I have a distinct feeling this whole thing is far more complicated than you're making it out be. If it weren't, I feel confident we'd have had better reflective light meters some time ago.
....
Well, sorry, but I can promise that it ain't gonna be that way.
It never has been that way. Might as well get used to it. This is no new subject. It may surprise beginners, but all photographers since light meters were invented (maybe 90 years?) have had to learn this (well, those that do learn things did). Some get it, and some never do.It is NOT a measurement of incoming light (that would be an incident meter). Reflected meters only measure how much of that incoming light these specific subject colors can reflect to the camera. This is a huge difference. Depends on the subject, what you aim the camera at.
Both your white and your black card ought to individually properly meter about the same middle gray (determined by meter - unless influenced by other things in the scene). This applies to both together too, that should average about middle gray.
Like it or not, that's how it is. Might as well learn to deal with it. It is still extremely useful.
The meter (at best) is just a dumb computer. It has no human brain or experience, it cannot recognize your black bag or white wall from your Aunt Martha. It sees some light which it can measure, but it cannot tell the difference in a brightly lighted black scene, or a dimly lighted white scene. Black scenes don't reflect much light, and are seen as dark by the meter, which boosts them to the middle. White scenes reflect a lot of light, and are seen bright by the meter, which drops them back to the middle. The meter has no clue what they were, or how they should be. Middle gray seems like the only workable solution. It's some light, and it is put in the middle. And it is a good guide for the photographer, but we need to pay attention.
But no matter what, the meter is going to put the average of what it sees at middle gray. Your picture of both the white and black card ought to average about middle. Your first bag picture almost did. Mountains and sky and trees pretty much do. Beach scenes and water and palm trees and sunbathers pretty much do. Portrait faces and hair and eyes pretty much do. Snow scenes or bright sky or white walls don't. It can frequently work about OK, but we do have to pay attention, not all scenes are average.
Spot metering works the same, it merely sees a smaller tiny spot, which it puts at middle gray, regardless of what it is, or what we may have expected.
The way a spot meter might be used is to meter on a face surrounded by dark or bright surroundings, and the photographer knows faces are not middle gray, so he opens one stop to compensate what he knows will happen. He just learns this. One stop on faces is a general rule of thumb we all learn. Spot meters are NOT a beginners point&shoot tool.
This is of course speaking of reflected meters, like are in cameras. The meter is "accurate", does what it should, but which may not be doing what you imagination thinks.
Incident meters are very different, and are much better and actually "accurate", more like "expected". Incident meters turn their back on the subject, and stand at the subject, and aim the light at the light source (at camera technically). Independent of the subject, they do not see the subjects colors, and are not influenced by how well subject reflects light. With the incoming light source metered accurately, then white things tend to come out white, and black things come out black. However, these meters cannot be located in the camera.
Last edited:
PaulPosition
Senior Member
Your eyes can see grey and yet your brain interprets it as black. Or as white. Or as grey, of course. All depending on context (it's night. It's in shadows. It *should* be white. Etc.)
Your camera only register photons.
Checkershadow Illusion
Your camera only register photons.
Checkershadow Illusion
Horoscope Fish
Senior Member
I think, @J-see, you might want to look into something like the Sekonic L-308s or another incident light meter. I think this, or another meter like it, would perform much more like what you're wanting.
....
....
Last edited:
Thanks, I'll check that out. In the past light metering as is was needed because photography was a chemical process. Some uniformity was required because of that. This 18% grey uniformity was a way to get around the problem of exposure time needed for that chemical process to occur without changing the time usage of the cam to something relative.
But these days it is a digital conversion done by each camera itself. As such, light-metering of the past seems very outdated. If they still use it, there must be a good reason for that. Either there's a problem they can't get around without this uniformity or someone's very very lazy.
Edit: I'm trying to search an explanation why this out-dated method is still used but alas, there's next to nothing to be found. I however came across a pretty interesting article while doing so:
Optimizing Exposure
Some might have read it.
Last edited:
But these days it is a digital conversion done by each camera itself. As such, light-metering of the past seems very outdated. If they still use it, there must be a good reason for that. Either there's a problem they can't get around without this uniformity or someone's very very lazy.
Edit: I'm trying to search an explanation why this out-dated method is still used but alas, there's next to nothing to be found. I however came across a pretty interesting article while doing so:
Optimizing Exposure
Some might have read it.
Well, things have changed quite a bit. Most modern cameras (except pro models) have "scene modes". When you use the "Snow scene" for example, you are telling the camera that it should expect to over-expose the 18% grey a bit. When you use normal A-S-P-M modes, the camera has NO WAY of knowing what the actual scene looks like, it's just trying to have detail in the medium range.
A point that the article says also is that this exposure compensation is more important and applies when you shoot in raw where you have the latitude to correct the exposure in post processing.
So, in the end, it's the photographer's job to do the processing (correct exposure and post-processing) to reproduce what he/she saw or wants to produce. The cameras and light meters are just artistic tools that can't do anything by themselves.
I'm indeed always correcting in post since my shots always are a bit bright. Which is why I started wondering about the metering and exposure, especially after doing early evening shots. If I take a shot during daytime, I need to lower one stop to match what I see.
I also was testing the exposure compensation and quickly discovered underexposing was a big no-no because of almost instantaneous noise increase. Overexposing I could bring down in post without any quality loss but not so much with underexposing.
I'm currently testing the whole exposure of the cam including compensation. If I understand the article linked, it wouldn't harm to always overexpose a degree. That's what I'm going to check and see if it makes a difference since increasing exposure costs in macro while underexposing brings gain. Sadly it doesn't bring gain quality wise.
I also was testing the exposure compensation and quickly discovered underexposing was a big no-no because of almost instantaneous noise increase. Overexposing I could bring down in post without any quality loss but not so much with underexposing.
I'm currently testing the whole exposure of the cam including compensation. If I understand the article linked, it wouldn't harm to always overexpose a degree. That's what I'm going to check and see if it makes a difference since increasing exposure costs in macro while underexposing brings gain. Sadly it doesn't bring gain quality wise.
Last edited:
Unfortunately, over exposing has it's drawback too... You'll have to use either a slower shutter speed, higher iso, or larger aperture. All these variables could decrease the quality you're after in your quest for perfect exposure correction. There are always limitations we must live with...
That's indeed what it is all about; finding the best compromise. I personally prefer one stop of shutter speed above some overexposure since it makes more difference shooting handheld with a 200mm. Better quality is useless when the shot suffers shake.
But I'll have to test the whole range first to see what I gain or lose and what can be fixed or not. Only when I know that, I can find the best compromise for any given situation.
The camera is exposing correctly, that is it is ensuring that sufficient exposure is there to light up the scene. That the "normal" scene as observed by you was dark, has nothing to do with exposure.I was bored so I did some testing.
Don't check the quality of shots, it's about the exposure.
.
.
.
The first shot is normal exposure. The second is compensated to -2EV. The second is accurate (as reality) which implies when light is low, the D3300 tends to overexpose by two stops. It isn't even that bright when I'm in the room and the bag is black, not gray.
It's the same using center or matrix metering.
Tomorrow I'll check if there's a difference during daylight. I somehow hope it does the same when plenty of light.
I agree with you, that the camera is "brightening" up the dark scenes, which as you rightly say is not the reality. Your logic that dark scene should be exposed dark, will create a lot problems for those of us who shoot in bad light, but expect the images to be bright and crisp. Photometrically you may be right, but photographically I prefer what the camera does in low light. Darkening light scenes is an easy step in any software, but brightening the scene with low noise is not.
You can take overexposure to a limit only. Beyond that the highlights will be blown, with no possibility to recover. I have found that in my D3300 overexposure to +1EV can be recovered, beyond that no. Of course there are times when I am shooting flowers, and the sky get blown, but that does not matter to me as I am after the flower not the sky, especially if I meter for sky and get noise in the flower.
Today I carried out at test to check the noise at all the ISO setting - 100 to Hi1 (25,000). Upto ISO 1600 the correctly exposed areas were quite noise free, 3200 barely some noise at 6400 quite bad. In shadows there is noise after ISO 400, and in deep shadows even at ISO 100. That means that shadows upto -5EV or -6EV have low noise and after that noise starts showing.
The good thing is that at high ISO I can use higher speed for fast moving birds or animals and get a shot, which I would miss at lower ISO, a good enough compromise.