An idea regarding exposure metering

[61Dynamic]

How would it know the difference between a field of snow and a field of coal?

[wolfnowl]

Take your gray card out early in the morning and look at it.  Take it back outside in mid-afternoon and look at it again.  Take it out again at night and look at it again.  The same 'scene', each time.  How can a camera compare the three?  It just measures what's there.

Mike.

[Digi-T]

In addition to the issues already raised, there's the additional complication of scenes with more dynamic range than the camera can capture. Deciding what's important and what can be thrown away is beyond the capability of any computer at this time; it's the responsibility of the photographer and not the camera in any case.

Well maybe my idea is not possible but I don't understand what the issue of dynamic range has to do with it. It is already an issue now but cameras still determine an average reading based on the 18% gray measurement. It seems to me it would still be the same if the camera was able to determine if a scene was 7% or 40% gray. Also, when you say it is the responsibility of the photographer and not the camera in any case, what do you mean? If such a feature were possible it wouldn't have to replace any current method but could be an additional feature. None of the features now remove any responsibilty from the photographer. It would simply be a feature that you could either choose to use or not.

T

[Jonathan Wienke]

Well maybe my idea is not possible but I don't understand what the issue of dynamic range has to do with it. It is already an issue now but cameras still determine an average reading based on the 18% gray measurement. It seems to me it would still be the same if the camera was able to determine if a scene was 7% or 40% gray. Also, when you say it is the responsibility of the photographer and not the camera in any case, what do you mean?

Think for a moment about what you're saying, and how ignorant it sounds. You have a 10-stop scene you're trying to photograph. Your camera can only capture 7 stops, maybe 8 if you're really good at noise processing. What do you keep, and what do you throw away? The answer to that is a compositional and artistic judgment call that no computer can calculate; it must be decided on a case-by-case basis. Here's an example where I made the decision to let the shadows clip to black, and retain highlight detail:



And here's an example of the exact opposite:



In the first photo, the most important element of the image is the subject of the portrait, and if the background is OOF and clips to black, that actually strengthens the image, as it focuses attention  of the viewer on the subject of the portrait. In the second case, the primary subject is found in the darker tones, and the best option is let the sky blow out and retain some tonal range and detail in the carved faces. Computers are totally incapable of making meaningful value judgments in situations like these; any auto metering program is going to screw up one or both of these situations. It's YOUR responsibility as a photographer to decide what's most compositionally important, and choose where you need to keep detail and tonal range, and where to let things clip to black or white.

If you're not capable of making those sorts of decisions about your images, then you should forget about photography and take up needlepoint or something.

[Digi-T]

Thanks Mark,

To try to clarify my discussion had nothing to do with dynamic range or exposure latitude, not directly anyway, and I do have pretty good grasp on these issues as well as the limitations they have on digital image sensors and film. I was trying to discuss the possiblities of sensors being able to determine if a given scene might need to be over or under exposed to give a more accurate and desirable exposure instead of the standard 18% gray value it wants to expose for. It is possible for exposure latitude to be a problem with any scene of any overall brightness level. That is a given with todays digital sensors and film. Jonathon, your examples are very good and they do illustrate my point as well because they are examples of images where it was necessary to over or under expose the images to some degree to match more closely to what you were actually seeing as opposed to the exposure that the camera would automatically choose. I guess what I am asking is this: if a sensor can anylize a scene and then determine an exposure of 18% gray then shouldn't it have to start somewhere to make those calculations? Is there no way that the sensor can tell if it is looking at a brighter or darker scene? I admit I am not an expert with these matters and maybe this topic is absurd but so what, I'm curious. If I was an expert I wouldn't be posing these questions here would I? These questions also have nothing to do with my knowledge of cameras and getting proper exposures with them. I am not a pro like you (not sarcasm, I really do respect your work and knowledge of photography) but I am not a beginner either. I am simply more artistic than I am technical. Sometimes I like to try to think outside of the box and envision new ways of doing things regardless if they seem impossible. More people should try this from time to time.

T

[Jack Flesher]

Yeah but... his idea could work if the camera had an incident meter built into say the top of the prism housing in addition to the advanced reflected metering system it currently has...

However, it would require very careful and precise use -- probably more than I'd be willing to deal with   

In the end it seems MUCH simpler to learn some intermediate exposure techniques.  In this fashion, you as the photographer can determine if the gray you are looking at is 7%, and  set your exposure accordingly.  

Of course, you can also determine the gray you are looking at is 7%, but decide you want it rendered as 11% gray for the image you are creating. (This is pretty much the concept Jonathan attempted to illustrate above.)  Now you as the photographer have taken personal control of how the image is going to look, and I think we can call that something like "artistry" ...

Cheers,
Jack

[Mark D Segal]

especially if she had the courage to make the move from needlepoint to a digicam (sorry - I couldn't help it - we need a few laughs at ourselves from time to time, don't we?)  :D  :D

[Jonathan Wienke]

I think we can all agree that understanding what is happening when the shutter button is pressed, and knowing how to tell the camera what we want the image to look like (via exposure compensation or manual settings or whatever) is the best way to go. I like to think of it as job security for professional photographers (though there are many highly knowledgeable "serious" amateurs out there and not a few ignorant "pros").

[Jonathan Wienke]

3) The camera could have a means of selecting the general brightness meteringt. Simply press a button and scroll the wheel to select from a range of say five or six general luminosity presets. The camera would then meter for that 7% or 24% grey instead of the typical 18%. Combine that info with the area in the frame the AF locked on to. For example, if the subject is white (say a bird at mid-day) the camera would automatically try to keep highlights from blowing out in the portion of the image that the camera has focused on. Conversly, if the subject is black, the photog could select that setting and the camera would prevent the subject from being exposed too much.

I think the current system of exposure compensation handles this just fine, without having to explain the difference between 18% gray metering and 45% and 5% to the noobs. There's all kinds of automation that can be done, but the bottom line is that much of it is of limited practical value in real-life shooting situations. The most capable photographic computer (and also the worst) is the one that lies between the photographer's ears.

[DiaAzul]

The problem still remains of what it is photographing. It just ain't smart enough for that.

What Digi-T is requested is not all that far from evaluative metering in current digital cameras. From what I can read in these posts there is a presupposition that the cameras metering system is 'stupid' and that it is not able to make some rational decisions based upon the composition of the scene.

However, if you read the marketing blurb from Nikon they put a lot of effort and emphasis on their 1005 point metering system and the database of lighting situations which it uses to determine accurate exposure. Comparing the snow versus coal analogue, there is no reason that an evaluative system can't automatically build in exposure compensation depending upon the EV measured for the scene. Canon, and many other digital manufacturers use similar techniques to accurately nail exposure for the user. We are a long way passed the dumb metering system.

For those photographers that do not wish to have such a high degree of automation in the camera then there is always the opportunity to over-ride the system and use average or spot metering.

However, the upshot is that Digi-T made a perfectly reasonable suggestion which is, in effect, an integral part of the modern day digital SLR.

As to JWs pretty harsh remarks against Digi-T. I will make the remark that these are public forums. As such anybody's conduct upon them is open to public scrutiny. As a commercial photogapher I would anticipate that JW would wish to maintain a good reputation in the market place in order to ensure that future work is not put in danger. Based upon the current tone of the posts I would neither recommend or contract him for any photographic projects.

[Jack Flesher]

We are a long way passed the dumb metering system.

They're still pretty stupid.

Amen...  Just try shooting a ptarmigan on a snow-field, or a black lab standing on an asphalt road...  They'll both come out dingy gray if you rely only on the meter
-- ANY meter.

Cheers,
Jack

[Jack Flesher]

and there are situations where using a "dumb" metering method like center-weighted average gives more consistent results than evaluative. Evaluative metering may be "smart", but it is certainly severely lacking in common sense.

AMEN!

ANY meter, even Canon's and Nikon's most sophisticated evaluative or matrix designs, will get fooled by a frame full of a single tone, UNLESS that tone happens to be equivalent 18% gray (or 12% -- I don't want to debate who uses what standard).

I USUALLY use my spot metering pattern with my camera in manual mode -- and yes, I even have PF 3 set to "spot."  When shooting landscapes (or indoor sports), I meter with it just like I do when shooting 4x5 using my Pentax Digital spot meter: That is, I select a highlight I want detail maintained in and put it in zone 7, then check the shadows and make sure the ones with I want detail in don't dip below zone 2-1/2 or 3.  Sometimes I'll even simply meter the clear blue sky -- a near perfect 18% gray if there is no haze -- and use that reading directly.  You'd be surprized how accurate that is for any broadly-lit landscape  ::

Cheers,
Jack

[Digi-T]

I've been thinking about the way cameras determine exposures and the problems involved with scenes that are real dark or real light.  Wouldn't it be possible for a camera meter to first analyze a scene and detemine its current exposure (for example: 7% gray for a snowy scene) then compare that reading to the standard 18% measurement to achieve the proper exposure of something like +1 1/3 (just a guess to make my point) overexposure.  I very well could be missing something here but it seems to me that as long as the camea has a baseline exposure measurement based on the 18% gray than it should be able to use this information to determine if your scene needs to be over or under exposed for an accurate exposure.  If I am wrong about this please explain why. Thanks.

T

[Digi-T]

It wouldn't have to know what the contents of the scene are but couldn't it know that it was either looking at a bright scene or a dark scene? It just seems to me that if I can look through an optical viewfinder and tell that I am looking at a bright snowy scene then a sensor should be able to tell as well. It would just need to be calibrated in some way. It's pretty confusing to me so I am likely overlooking some major thing. It was just a thought.

T

[Jonathan Wienke]

In addition to the issues already raised, there's the additional complication of scenes with more dynamic range than the camera can capture. Deciding what's important and what can be thrown away is beyond the capability of any computer at this time; it's the responsibility of the photographer and not the camera in any case.

[61Dynamic]

The point you seem to be missing is that the camera does not know the difference between a 7% or a 40% grey scene.

The camera uses reflected light to meter a scene. It has no idea how much light is coming from the light source; only what is reflected off the subject. It also has no idea what it is you are photographing.

Dynamic Range comes into play when a scene has a broader range of light than what's capturable by the camera. If the camera is not able to know what it is you are shooting, how is it susposed to know if you want hightlights to clip, shadows to clip or a little of each? How would it know it it's clipping a specular hightlight or someones white t-shirt?

The camera is increddibly stupid at being accurate. All it knows is 18% grey and can do a very good job of achieving that based off it's current metering mode. You need to know if you need the scene to be brighter or darker than that.

[Digi-T]

Think for a moment about what you're saying, and how ignorant it sounds.

If you're not capable of making those sorts of decisions about your images, then you should forget about photography and take up needlepoint or something.

Forget I even said anything. Didn't meen to sounds so ignorant to you. I was just trying to think out loud regarding some types of photography that often cause a lot of problems, especially to new photographers. It's no wonder that a lot of people don't like to ask questions.

T

[Mark D Segal]

Jonathan,

You are a very knowledgeable guy, and a skilled photographer, and I like the portrait you displayed above - really good work. But you should learn to mind your bedside manners. Not everyone is at your professional standard, because they are most probably not professional photographers, but they still deserve help and respect. That is one of the main reasons for this Forum to exist. And don't belittle needlepoint - the gazillion needlepointers of America will stick your effigy - for reason!

It would be more constructive and useful to simply inform "digi-T" more about dynamic range and its limitations, but there are several techniques to expand dynamic range if one had artistic or other reasons for not wanting to shed levels in one place or another (the perfectly valid solution you mention above for numerous situations). Examples (i) make two different exposures of exactly the same subject and blend them in Photoshop, (ii) if the problem is not too utterly severe, use two differently exposed variants of the same RAW captureand blend them in Photoshop.

[Jonathan Wienke]

Forget I even said anything. Didn't meen to sounds so ignorant to you. I was just trying to think out loud regarding some types of photography that often cause a lot of problems, especially to new photographers. It's no wonder that a lot of people don't like to ask questions.

When you ask the same question again after having the answer explained in a reasonably clear fashion by 3 different people, you really need to take a step back and think about what you're saying and whether it makes any sense. The point of my previous post was not to tell tou to give up photography, but to explain why you need to take responsibility for decisions you're trying to foist off on the camera. And you did ask for clarification on that point.

If you stick a sheet of black paper in front of a camera lens at noon, and a white sheet of paper in front of the same camera and lens at dusk, the light levels as measured by the camera meter are going to be roughly similar. But in spite of the fact that the light levels entering the camera are similar, in one case the camera is expected to render what it sees as black, and and in the other case it is expected to render the subject white. All the camera "knows" is what it sees through the lens, and it has no way to know what the subject is "supposed to" look like. The decision to take that identical light level and adjust exposure to render it white or black is the responsibility of the photographer; all the camera can do on its own is is put it in the middle gray tones.

[Digi-T]

The photographer is always responsible for their photographs and have total control over them no matter what tools they use. That should be a given if you are at all serious about photography. I mean most of us use auto focus but don't we, at the same time, try to confirm the focus ourselves. However, many new photographers and especially casual snapshooters don't consider this and expect the camera to know everything. The results are almost always going to be gray snow scenes and gray, and blurry, night scenes. One possible use of the feature that I am talking about could warn the user of the potential problem and suggest that it fixes it or to continue shooting in the normal mode. I know that most pros or serious enthusiasts may not use such a feature but many features are most appropriate for casual shooters. I think that anything that can help grandma to take a better picture of a precious moment is a good thing.

T