Why use a light meter?

[Manfred M]

2. Disclamer:
I would like to state that I do not have any experience with external lightmeters and do not deny their usefulness for applications I may not use or understand. I just wanted to correct some wrong statements made here. I hope this to be useful to the community.

Excuse me for my heartfelt laughter; the statements I've made are not "wrong" but I simply suggested the advantages of an incident meter over a reflective meter for certain types of shooting. As I mentioned elsewhere in this thread, most people that run down incident metering have never used one. By your own admission, you appear to fall into this category.

I probably use an incident meter for less than 5% of my work, so I am well aware of the advantages and limitations of reflective meters, and unlike most people, I not only relying on my in-camera reflective meter; I also use a 1° reflective spot meter head with my hand-held light meter. I probably understand metering as well as, if not better than most...

The important thing to understand with any reflective metering there is that it is based on the assumption that the parts of the image measured, converted into black and white, could be represented by a middle gray value (I’m not going to get into how middle gray should be defined; as there are all kinds of differing views on that as well). All of the various algorithms use different ways and / or points to measure, but the ultimate assumption still resolves around the underlying middle gray value giving a proper exposure. If the collection of measured points result in values that are not close to middle gray, the recommended exposure will not be correct. That is true for any metering mode, spot, matrix, etc.

Again; an incident meter is only useful when shooting subjects that are relatively close to your camera; especially when one is using multiple light sources to light the scene. This is where I tend to use an incident meter; building up the lighting for a scene. It’s far faster and easier than trial and error, and I do keep notes, so doing a reshoot later on is really accurate and easy, because it can be validated quite quickly.

Your second point, I use a similar technique when doing a pano landscape. I try to understand the lighting and light distribution across the scene so that I can determine the “best” exposure across the multiple frames I will have to blend while trying to protect against blown highlights (and to a lesser degree, blocked up shadow detail).

[Colin Southern]

Disadvantages of reflective metering - in a nutshell - with examples:

http://www.sekonic.com/classroom/met...reflected.aspx

[Inkanyezi]

Disadvantages of reflective metering - in a nutshell - with examples:

http://www.sekonic.com/classroom/met...reflected.aspx

The Sekonic paper shows the simple way this topic is often covered in their marketing, but it is in a very profound way disregarding how a skilled photographer might actually work.

The key is understanding exposure. Any light meter, hand held or the one in your camera, can be used to determine exposure, and the incident reading takes out some of the guesswork that must be done when taking reflected readings.

Incident light meters like the Weston Master were invented to facilitate work for movie shooting, where the mood of the shot must be nailed at the time of shooting. At that time, there was no way to immediately check what you got.

The zone system was conceived in a different approach to the same problem, taking into account the actual dynamic range of the scene and the medium's ability to cope with it. It used techniques unavailable to the movie shooter, as different development time for different scenes, an also incorporated darkroom work, that would not likely be done to the movie. Hence both approaches have strong and weak points, that may apply or not to our digital photography. The incident readings were easily applicable to shooting on slide film, while the zone system would be more difficult to fathom, although once mastered the idea behind would also work well. The incident readings however gave a "correct" reading faster, something that was important to professional movie shooters. To use the zone system, its range of zones must be adjusted to that of the medium. Once our digital cameras had a very narrow range, similar to slide film, but that was in the past.

We are in a way back to making zone readings useful, as digital sensors now match B&W film in dynamic range, the best ones recording up to 14 full stops. Hence it is possible to bridge the enormous gap in brightnesses in just about any scene. But as our presentation medium has a much narrower brightness range, we need trickery to overcome the inherent problems, just as did Ansel Adams with his zone system. He dodged and burned, and we must do that too if we want to accomplish anything similar.

There has been suggestions here along the ideas of the zone system, as spot reading on the brightest tones and compensating +3 (in my cameras +2), which will place those areas in the desired zone (VIII). A particular problem here is colour. It is said that the meter is colour blind, but we don't know exactly how, and those blinkies we get from the camera's highlight warning won't kick in if only one colour is blown - mostly red. We won't see that until our RAW processing software tells us that red is clipping. That aside, proper understanding is still the key.

Yellow is a tricky colour. It does not exist in a digital image. We have green, red and blue in our sensors and the receptors of our eyes, but no yellow. Yellow is a construct, tickling our red and green receptors. Colour is not a physical entity, but a psychological one caused by our brain. The filters in the camera physically strips unwanted light wavelengths from hitting the sensels. Green won't ever clip in the camera. Red will. However when yellow clips, it is so bright that we won't see that green is still there. It will still look like yellow where its red is clipped.

When doing macro work, the incident light meter is often not the easiest tool to use. As the lens is often elongated when shooting close, the actual f-stop is mostly smaller than the one indicated on the scale, and readings have to be adjusted. The reflected reading does off with that problem when done in the light path. It only has to be corrected just like any other shot for the general brightness of the scene. Still understanding exposure is the key.

My experience is that when shooting flowers, which often have very saturated colours, highlight blinkies fail, although the separate colour histograms will mostly tell you whether you get full tonality or not. Mostly it is the red histogram that has to be used to determine exposure, but sometimes the blue. A real time histogram is a great help in determining exposure before shooting, an advantage with mirror-free systems.

PP tweaking of the image mostly is needed when exposing to the right, as the images otherwise might look rather dark. The on screen jpeg won't tell you much about this, as even a dark shot comes out rather bright on the screen. The simpler approach is to accept clipped highlights, which are often the result of using the "matrix" readings without any compensation. When only the red channel clips, saturated red tones appear "flat", something that often occurs when shooting red flowers as roses or poppies. If you google "poppy" and look at the images, you'll see that well over 90 % are clipped.

So, many of us still use hand held incident light meters, and they still do the job they are intended to do, and do it well. I have used it occasionally, to check whether my old way of doing things would be easier than the new ways with the built-in meter. In that way I learned in what circumstances the camera would be prone to give incorrect readings and when it may be trusted to do a good job on its own.

I never really became friends with what's called "matrix" metering, simply because I cannot understand what it does, what decisions it makes. Still the key is to understand exposure. I can use the incident meter, I can also use the spot readings, and I do use the "matrix" metering when I see no problems with it. However in the simplest of situations It often fails. When it gives a reading that is different from "Sunny 16", it is probably wrong and burns out highlights that I want to retain. Neither matrix nor spot will give a correct reading for a poppy - incident however will, unless it is backlit.

But the hand held meter is not needed. The one built into the camera does a decent job, once you learn to use it properly. It gives a reading for 18 % (or 16 % or 12.5 % or whatever) reflectance, but it is up to the photographer to evaluate the scene brightness and tweak the reading for proper exposure. Here's where understanding the zones and how they relate to your meter readings comes into play. And that yellow flower mostly should not get into the very brightest zone, neither should the red one. But colour is different entity from the simple grey values of B&W, and it is not easy to see when a red colour clips. The histogram and statistics for clipping is the only help there. A clipping red channel would still look like not as bright as a bright yellow or white, but would still lose tonality because it is clipping. The only visual clue is that it looks "flat" in the area where it is clipped.

So I think that the light meter might be a help in learning about exposure and get to grips with it, but it is not necessary. Learning the zone system in all its aspects is not necessary, but the concept of zones and brightness range is essential to understanding exposure. Also learning how to use the histogram is a great help, particularly where your readings as well as your eyes may be fooled by a saturated colour. When shooting flowers the red channel histogram often will give a clue.

[calexe]

Excellent input explaining in detail many things. Including the marketing confusion induced probably voluntary by some lightmeter manufacturers. See also link provided by Colin (thanks).

Again: Yes, camera has reflective lightmeter. Yes, incident lightmetering is better (and unfortunately rarely possible). But camera "wrong" exposure in matrix metering mode is heavily caused by the metering algorithms of the camera, not necessarily by the fact that the camera meters reflected light. And camera "wrong" exposure indication with spot metering is usually caused by the fact that the method is not used properly, I mean not according to the Zone System.

...
There has been suggestions here along the ideas of the zone system, as spot reading on the brightest tones and compensating +3 (in my cameras +2), which will place those areas in the desired zone (VIII). A particular problem here is colour. It is said that the meter is colour blind, but we don't know exactly how, and those blinkies we get from the camera's highlight warning won't kick in if only one colour is blown - mostly red. We won't see that until our RAW processing software tells us that red is clipping. That aside, proper understanding is still the key.
...

Yep. Right. Somehow additionally to what was said: When exposing to the right the final image may have some problems with the colors. I mean, even if no channel would be clipped, when pulling the exposure back in postprocessing, colors might not land where you really want. Some tweaks may be necessary to bring them back where they should be. ETTR is very effective in preserving details and low noise in dark zones. But if one would not hardly need ETTR, then an exposure closer to WYSIWYG may be better. That is: expose for a zone / object with neutral luminance (set spotmetering in the camera and adjust needle of meter to 0) and check for highlights in zones of interest in the image. This is called "Meter for Midtones, Check for Highlights". A standard grey card may help here, but not everybody would carry one with her/him and not every photographying situation allows you to use one. In such a case learning some "templates" is very effective. You may know, skintone = +2/3, northen daylight sky = +1, white plumage = +1 2/3 ... +2, etc.. Thus you get an image very close to WYSIWYG and effort in post will be minimal.

[Colin Southern]

Again: Yes, camera has reflective lightmeter. Yes, incident lightmetering is better (and unfortunately rarely possible). But camera "wrong" exposure in matrix metering mode is heavily caused by the metering algorithms of the camera, not necessarily by the fact that the camera meters reflected light.

It's somewhat of a moot point though because if the net result is sub-optimal exposure then whether it was caused by "this" or "that" it's still a sub-optimal exposure. USUALLY, matrix / evaluative will do a pretty reasonable job with average scenes containing only reflected light; the "fun" starts when there's an incident light component ("back lighting") or a scene that varies significantly from "average" ("black cat on a black rug" etc). Incident light metering works well for the latter example. Nothing really works well for the former example (usually they're situations where fill flash is needed more than some mystical "ideal" exposure).

Yep. Right. Somehow additionally to what was said: When exposing to the right the final image may have some problems with the colors.

ETTR is becoming as controversial as UV filters for front element protection, Canon -v- Nikon, and Mac -v- PC. At higher ISO settings it's a must when the decreasing DR capability of the camera becomes a tighter fit with the DR of the scene, but at low (base) ISOs I contend that usually it's a waste of time because (a) the camera is already capturing (11 or 12 or more) far more stops than we can print (4) or display (6). What they DON'T tell you is that pushing things too far can put some channels into the sensor's non-linear response region where you end up with weird colour shifts that end up being "close" but still "not quite right" - that you just can't fix in post production. Assuming that it's not some high DR scene that requires every stop that the sensor can record, I've personally never seen the logic in over-exposing something by a few stops only to reduce the exposure back to where it was in the first place in post-processing to protect shadow detail that we're probably going to clip to increase the image contrast anyway.

In summary - IMO - ETTR has a place - it's just not needed anywhere nearly as often as people think. Some think "why not use it if there's no downside" - but to that I reply "often there IS a downside". Personally I've also found that often its difficult to get the gamma right after ETTR as well. For me, I'd have to say that more often than not it creates more issues than it solves.

In terms of exposure, more often than not I'm simply using the histogram to indicate any degree of under-exposure and the blinkies to alert me to areas of potential over-exposure; yes - of course - it's true that the in-camera histogram is based on the in-camera JPEG conversion -- but on the flip-side, that in-camera JPEG conversion usually has some REALLY NICE safety margins associated with it. These days I can normally use the fill light slider to reveal all the shadow detail I need - a FAR bigger issue however is trying to recover sufficient info in highlight regions - esp in sunset landscapes.

[Inkanyezi]

I'd like to add that the histogram has its own learning curve. Mostly we can disregard everything in the histogram but the extreme right, and ETTR is to let the histogram end exactly in the lower right corner. However the histogram resolution often won't show us any highlight that we might want to retain. So still, judgement is crucial to the process.

Then when we look at the colour blindness of the exposure meter, if we assume that it reads different colours in the same way, it is easy to understand that each one of them, red green and blue, caters for 1/3 of the reading of a white or neutral grey shade. This will mean, that a clipped saturated red where no green or blue is present, will only show up on your meter reading as one third of the reading for a zone X. To tell the light meter that it would clip, all three channels should read full. So for saturated colours, the spot meter will not give you a "correct" reading corresponding to reading a neutral area. There's where taking a reading from a grey card may help, or the dome of an incident light meter.

The incident or grey card reading would disregard the colour of your subject and only read the light falling upon it. The poppy will not be 90+ % reflective, but only 30+ %, even though it is bright enough to clip the red channel if you use a reflected light reading without taking that into account. As it won't clip blue or green channels, the colour blind reflected light meter is happy with telling you it is only a third as bright as it actually is. The histogram later will tell you the truth. The combined histogram will show that it has hit the right wall, but not really how much. The red channel histogram will be more tell-tale.

You cannot learn this only by theory. Just as swimming is best learned in the water, you learn photography by taking pictures. But the theory may be a help in understanding what happens and why things sometimes go wrong or work a bit different from what you might first have assumed. I think it is a Good Thing to understand what "saturated" means, and why it gives a lower reading on the reflected light meter. A saturated colour lacks the other colours.

[ajohnw]

I've never really proposed actually using the zone system as is only reading the wiki to get an understanding of what exposure compensation is about - what a metering system has to do. With mid grey metering system, spot or hand held something very similar is needed. There are also other similarities when the image gets displayed on a PC screen or printed.

An additional fact is that iIt's easy to talk about cameras having 14 stop capabillities but the limitation is the final image. The 14 stops aren't so simple as they sound either. Counting digitally they go .0,1,2,4,8,16,32. The problem with the lower numbers is that they can't really show gradation that our eyes would recognise and relate to the original image we could actually see when the shot was taken. Being generous in that respect to illustrate what I mean the first real stop might be 32. In that case stops go 0,32,64,128,256 etc. That illustrates another aspect of the problem. Sensors are linear so there is sufficient gradation 0 to 32 .and way too much at 128 to 256. Raw converters and camera jpg engines get up to all sorts of things to try and get round that but the net effect is that finally we find 5 maybe 6 stops come out in an acceptable manner and that PP can bring some of the dark end up into that tone range or increase that aspect at the expense of the light end. A lot depends on the shot. Some shots only have a 5 - 6 stop range in them. I'm tempted to mention the golden hour and light conditions here but people should be able to work that out for themselves.

Colin's link like most similar ones doesn't even mention variations on matrix metering but should mention that the readings assume average reflectance. Try a car or add a highly polished stainless steel plate to the photo shown and it wont work out. The reflected light shots ignore matrix type metering and show exactly what a mid grey meter should do. One problem here is that mid grey has been both 18 and 12% at times and may still be. The reason that it isn't 50% relates to stops ie how we see. An incident meter should also give a perfect exposure of a mid grey card. It will have average reflectance.

Matrix metering uses reflective light intensity and colour to make it's decisions augmented by area relationships in what it senses.Where it mostly runs into real problems are when light level variations are more than the camera can record and also the tendancy to concentrate on what the camera is centred on - think faces for a simple example of small items in a larger scene within reason or some large object taking up a lot of the view.. There are other side effects such as the flower photo posted. Here it can sense a significant amount of back in the scene that has light variation - so it captures some of it at the expense of slight highlight clipping. It wouldn't matter what colour the staemen were if they were of the same "brightness". As I see it this is an operator problem - under exposing the black card isn't a problem over exposing the rather small area of the brighter staemen. There is a need to understand the characteristics of the metering that is being used which ever one is used - some are easier to master than others. I've been as far as spot metering a scene and calculating an exposure to fit what I want so I know what I think. That sort of approach quickly shows that it's oh so easy to miss a bright spot in the shot. Another problem with matrix metering - it doesn't measure the entire scene only samples it.

Urban if you think camera metering software and the other bits and pieces associated with it such as area recognition you should go to the camera manufacturers and sort it out for them. Seems they have wasted many years of effort and improvements. I suppose people look at the size of firmware updates and relate that to windows and it's sotware. No such comparison is possible. I have worked on some rather sophisticate set up that use no more than 4kb of code and even included diagnostics. That's assembler. C is generally used now and even more sophistication but even 100K can do rather a lot. Actually there is an interview with Nikon's senior engineer on the web some where, wish I had kept the link. He feels that it can't be improved further in any significant fashion so the next step is to cure the problems with the results. Hence D light, autogradation and auto HDR etc variations cropping up all over the place. So in the near future cameras will produce perfect jpg's in situations like this one. On this camera due to the light range the tone curve the camera used has kept the clouds at the expense of the dark end. It's an interesting example of an E-PL1's jpg engine plus metering and too much tone variation so what to do? In this case it's decided the clouds are more important and compressed the dark end as a result. In other situations it will decide some item in the scene is more important - might be a face and then go on to tell the focus software to focus on an eye in more recent cameras. There are all sorts of things going on in metering decisions.



That is a typical example of what matrix metering will do with too much light range and a lot of sky. Less sky and the rest would be have better exposure. Get this result with a spot meter or a hand held meter - wish you luck. With this particular camera it was often best to make the camera expose in this direction to retain cloud detail - or take shots with less tonal range.

John
-

[Colin Southern]

That is a typical example of what matrix metering will do with too much light range and a lot of sky. Less sky and the rest would be have better exposure. Get this result with a spot meter or a hand held meter - wish you luck. With this particular camera it was often best to make the camera expose in this direction to retain cloud detail - or take shots with less tonal range.

John
-

I would argue that the exposure metering for that shot was just fine; it's not a metering/exposure issue - it's a dynamic range issue (we can't display the full DR). A little DR compression and it's perfectly fine; enough information was captured. Any evaluative / matrix metering will generally protect large areas of highlight.

[ajohnw]

I would argue that the exposure metering for that shot was just fine; it's not a metering/exposure issue - it's a dynamic range issue (we can't display the full DR). A little DR compression and it's perfectly fine; enough information was captured. Any evaluative / matrix metering will generally protect large areas of highlight.

I'd agree Colin in fact I reckon it's as perfect exposure as it's likely to be for pp with a camera standard tone curve etc. That's why I posted it. I just used the words tonal range rather than dynamic. There will be some compression in the highlights and even more at the dark end but the detail in that is there and could be bought out - more so from raw. This is where Nikon are heading with D-Light to automate the recovery. In this case If I had pointed the camera down more reducing the amount of sky it would exposed more for the rest of the shot.

I may not have been clear about what I was getting at as had to go out for a while. In real terms given several attempts at photo's in various situations of various objects it soon becomes pretty clear how matrix metering on a particular camera behaves and make it do what the photographer wants with a fair degree of precision. In my view it's pretty unique in that respect to other methods but still needs compensation.. I feel that the preview is the final word in situations where some aspect is critical - what ever metering is used. It's a case of how many shots to get there. Actually given 4-5 stop range in a shot I would expect it to get it spot on. Much more than that and there will be some degree of a problem in actually viewing it without PP. In the case of jpg's the camera will compress all or some of the dynamic range to make it fit. The all seems to be a recent facility on some cameras, other just compress more. However it's done contrast is lost some where in the tone range. Raw converters do exactly the same thing. Actually I get the impression that jpg engines are slowly but surely catching up with working from raw and have already passed the basic levels of pp in that area.

John
-

[purplehaze]

This has been a very informative thread for me and has given me some goals for my weekend shooting. Thanks Louise for asking the question and everyone who has contributed. Thanks especially to William, for sharing your idea about spot metering different colours; the timing couldn't be better for me.

[wlou]

Looks good to me though the stamens[?] look a little soft perhaps becuase of slight over exposure [ clipping?]
The color plus the black background makes for a luxurious image

If the mixing of light sources gives you a peculiar color mix then turn it into a B&W photo, or just ignore the error as I have done a couple of times when I add a tungstem anglepoise to daylight when shape and form are the key aspects of the shot.

John, the colour was really like it shows on my screen. As soon as the first flower blossomed I knew I wanted to take pictures of it.

[wlou]

I couldn't find this coming up already:
Leaving flash aside, there is a principal difference between spotmeters which read reflected light and such meters which read the light which falls on the surface of the motive. In the latter case, you have to read in direction of the light source (whereas you hold a spotmeter on the motive). The benefit of reading the intensity of the light source is that the reading then is independent from the color of the motive (a light meter is color blind), and independent from the relative brightness of the motive. Portrait photographers often like to use this method.

Lukas

Lukas, to read the light source of a flask, you would have to read the actual "light" or firing of the flash which is different than the ambient light? Is this correct?

[Inkanyezi]

Flash meters are a different matter but incident light meters often aren't suitable for many shooting situations. One instance where they can be of some use is out door portraiture as they will measure the light falling on simple subjects like that or even the reflected light off the subject which is what the camera is really interested in. In real terms this is little different to what the camera can measure if it has a large spot mode and has the same problems - metering to mid grey.

I didn't jump on this one at first, because I didn't want to mess up the thread more than necessary. But the above is close to the worst hogwash I have seen on this board. It contains significant disinformation that shouldn't pass without comment.

Incident meters did not fall into disuse because of any drawback with them. They were never used much other than by professionals and serious amateurs, or perhaps real nerds. Incident meters are still used in a few professional fields, where studio flash setups might be the most significant one. Fact is that from about mid sixties and on, cameras have increasingly had a meter built in, and no hand-held meter has seemed necessary for most people.

Flash meters usually are used for measuring incident light, but the TTL system in the camera reads light off the subject for obvious reasons. The in-camera meter captures the light in the same path as the sensor, the image itself. Thus it shares the downsides of any reflected light meter, although it is possible to write algorithms that take different aspects into account, such as contrast, highlights, shadows and saturated colour. I do not know whether there is any camera manufacturer that has put much effort into this other than Nikon, with their Active D-Lighting, that seems not to take colour into account, but makes a general assessment of highlights and shadows and adjusts exposure accordingly. To my knowledge it is the closest so far we have got to an automatic system that "exposes to the right".

The camera is not interested in anything, neither your visions of what to accomplish nor the light reflected off the subject. It is a piece of junk put together by rather competent magicians with the purpose to capture images. The person behind the camera is what's more important to image making; the camera is but a tool, even if the crucial one for the job to be done.

No assumptions at all are done about reflectance when taking incident readings, and it is exactly the strong point with incident metering. As no assumption is done, black will register as black and white as white, and all the tones between will fall in their right place - provided that they fit within the dynamic range of the medium for registering. Even the red poppies will come out correctly exposed and not as red blobs. It is only the reflected light meter that is built upon an assumption of reflectance. The incident measuring method completely disregards it. No assumptions, only measuring the light falling onto the subject.

Exposing black or white or any low or high colour tones to mid grey wouldn't be of much use to any one. The idea of an exposure is to finish up with dark areas being dark and light areas being light and all the tones in between being at the correct level so any metering system just based on mid tone levels is very likely to need some correction.

The above is exactly why any light meter reading light off the subject often fails to correctly assess exposure for the subject, as it makes an assumption, that can be wrong and often is. Evaluative metering was intended to take out one of the inherent problems, that of unlinear response of the light meter. Bright areas will carry too much weight on the reading, but with an evaluative system built upon evaluation of various fields can figure out what is highlight and what is shadow and adjust exposure accordingly. Whether or not this is actually done could be up to debate.

Again, all problems pointed out here are inherent to measuring light off the subject (as TTL metering) and do not apply to incident readings.

I have never used a flash meter but did know some one who used one daily along with several studio heads with modelling lights. The modelling lights were fine for getting the eventual lighting correct. What to do with the flash meter reading when applying it to the camera came down to experience.

I also knew Lassie.
But in fact, I have used incident light meters since the sixties, and I still have my Sekonic Studio De Luxe and sometimes use it.

Early in his career, on film, he did try using a light meter on subjects such as cars, trucks, vans and anything else that some one wanted to sell that was out doors. Customers were happy but he wasn't. He eventually found that the exposure guide that used to come on the side of a packet of film or in it gave the best results. He still waved the light meter about using it more or less as a prop to impress the customers. I suspect a lot of that sort of thing went on. Might still do so too.

The table in the film box is based on the Sunny 16 rule. It always works when applicable, and I use Sunny 16 as a litmus for how the "matrix" metering performs under known conditions. In all cameras I have used - several brands - the "evaluative" system burns highlights in plain ordinary outdoor lighting, mostly by one stop. Obviously I use compensation to cope with it, as it is easier than changing exposure manually for varying light. I guess that this is because of an assumption by the camera maker, which differs from how I want my images exposed.

Personally I feel that the best option on modern cameras is to use the fancy metering mode. Usually called matrix or something like that. It forms a fairly firm basis for when and where to apply compensation which people need to learn. I used to use centre weighted on film but on a Nikon. Unlike current centre weighted it used an oblong shape that went from the bottom of the frame to about 2/3 of the way up. The exposures were also heavily weighted to this area and on film at least this gave very predictable results probably because the scene mix it obtained was often a good approximation to mid grey. Modern cameras in my experience are not too good in this area and virtually every shot will need significant compensation.

Now I would want to know exactly what it means that the exposure is "heavily weighted to a particular area", and why predictable results on film would become unpredictable when shooting digital, and why the "scene mix" was a good approximation to mid grey in the past, but not nowadays. And exactly when in time did this threshold appear, that changed the way we were assessing our photographed subjects? Why would, suddenly, a modern camera become worse in a field that is so relatively simple as measuring light off the subject, and what was it that changed the reflectance values so much that compensation became needed now, but was not before?

Hint: The reason why I am not friends with neither "evaluative" (matrix) or centre-weighted metering is that I cannot understand what on Earth it is supposed to do, or what it does. If anyone here can explain it, I am happy to listen, but from my 50+ years of photography experience, I still haven't got a clue. I do know however how to use both incident and off the subject readings with any type of meter, and I can do the guesswork to overcome the colour blindness of meters to expose my poppies correctly also with a spot meter. I can do that, because when I measure with the incident meter or the reflected meter or the spot meter, I know what it does and what to expect. I know that the red poppy is saturated and that if I want it to all but reach the value 255 for red, I can use the spot reading with a compensation of +0.3 as it is "black" for the colours green and blue. I haven't figured out exactly how much each of the three colours will contribute to the reading, but somewhere between ¼ if the Bayer pattern is the decisive factor and 1/3 if it's just the colour. And as anyone can see, googling "poppy", most of the poppy images have the red blown out completely. I would say it is a sign that "modern cameras are not too good in that area". Thus I second the statement that significant compensation is often needed.

It all boils down to a learning curve. Any artist will have to learn using the tools of the trade, and there is no shortcut. There is one great advantage with digital in this respect, and it is that we can waste as many exposures on a single subject as needed, and it doesn't cost a penny.

[ajohnw]

Hogwash Urban - well we have been here before haven't we.

I suppose I have never used a camera without a meter - wrong
I suppose that didn't teach me the problems associated with using them - wrong

Worst of all in the general area is what various web pages etc show to prove there point. Contrived situations. The even worse aspect is that cameras of any kind only record reflected light and pretending that does not have it's implications is an insult to peoples intelligence.

I should have added that if some one wants to use one that's fine by me. I have no strong feelings on the subject. It's just that I have been there, done that and find I prefer the ones built into the camera.

John
-

[Polar01]

Been following this thread with interest, I do not have an incident light meter however been thinking of getting one as it can help in some situations. Now I was watching a web show by Joe Brady, which showed a number of things that have been talked about on this thread. He was shooting a like skin model with black hair and top, it was out side against a building with light shade. He shot one image, matrix metering, one spot metering, and the final one with a incident light meter, want to guess which image was the better one. Matrix blew out the skin, spot skin better but image overall to dark, incident light meter great skin and background was still there.
I look at the metering as done by the camera either matrix or spot is good enough for most of us. A lot of pro's never use a hand held incident light meter let alone how to use one as what they are getting is good enough. That said for pro's at the high end of the scale, the camera light meter is good, however it is not good enough for them. There are times when the camera's meter is good enough, but depending on the type of images being done and with a client paying for them then you want to make sure they are as close to perfect exposure as possible.
Compared to 40+ years ago the in camera metering has come a long way, as I said, in camera meter for some including myself is good enough most of the time for others it is not good enough all the time.

Cheers: Allan

[lukaswerth]

Lukas, to read the light source of a flask, you would have to read the actual "light" or firing of the flash which is different than the ambient light? Is this correct?

Yes, I very much think so. Typically, however, light meters which measure the light source do so through a diffuser. But as far as I know, this method doesn't make much sense with very concentrated light source, like direct sun or direct flash. Too harsh, not wanted for portraits or products (as was mentioned above), unless you want to set some additional spots for your flask.

In my analog work, I have used mostly spotmeters in order to evaluate the dynamic range. This makes principally also sense with digital, but one may use an inbuilt spotmeter or just evaluate highlights and shadows through trial exposures (no film to waste, results can be instantly checked). For me at least, with my DSLR a lightmeter, also for measuring the light source, seems an overkill. If I would want to be a bit derisive, I'd say it's something Leica photographers often use to give themselves some style.
-Not speaking for commercial portrait or product photography here, of course.

Lukas

[ajohnw]

Yes, I very much think so. Typically, however, light meters which measure the light source do so through a diffuser. But as far as I know, this method doesn't make much sense with very concentrated light source, like direct sun or direct flash. Too harsh, not wanted for portraits or products (as was mentioned above), unless you want to set some additional spots for your flask.

In my analog work, I have used mostly spotmeters in order to evaluate the dynamic range. This makes principally also sense with digital, but one may use an inbuilt spotmeter or just evaluate highlights and shadows through trial exposures (no film to waste, results can be instantly checked). For me at least, with my DSLR a lightmeter, also for measuring the light source, seems an overkill. If I would want to be a bit derisive, I'd say it's something Leica photographers often use to give themselves some style.
-Not speaking for commercial portrait or product photography here, of course.

Lukas

One thing light meters remind me of is a friend of mine that was a successful professional photographer on medium format film. He started with weddings and social work etc even schools via a company that puts out work to people near schools that want them. He used a light meter. His aim was to get into product work. He did eventually rather a lot of it. One of his early jobs came from a company that had bought a large number of ex military jeep like things and they needed a photo of every one for sales purposes. This is where he 1st came to grief with the light meter. The customer was happy with the shots but basically he wasn't. Something more lucrative might turn up where higher quality images would be needed. One aspect of a job like this is going out and shooting when ever the shot is needed. They might stick one under cover if it was raining but that's about it. Lighting and weather was as is and this is where the meter fell down. I don't think his solution would work out on digital but it does on film. Packets used to contain exposure guides, bright sunlight, shade, partial cloud etc etc. As strange as it might sound he obtained more consistent results that way and always used it out of doors colour or black and white.

Flash metering was a different. His studio was 2 floors of a wing of an old mill building. Automated processing equipement taking up about 1/3 of one floor. Flash heads were on overhead gantries and products were placed on a large table. I'm only mentioning details to point out that he was a very serious professional photographer. I know for a fact he compensated some readings before taking shots but suspect that the biggest difference in this area is that it's a controlled consistent light source.

Leica - Yes and he was known by me at least to wave an incident light meter about for effect. Pro's need to do that sort of thing just as they needed to shoot MF film to set them apart from 35mm part timers etc. Once established his social work was extremely expensive and some people you might say ideal people from a business point of view still wanted it and were also prepared to pay for it.

John
-

[Inkanyezi]

Again the same anecdote of the successful photographer that didn't learn to use his light meter. I cannot help but feel a bit sympathy for the bloke.

Any tool of the trade has to be learned. A light meter is no different from other tools in this respect. But personally, I would not assume that failure to achieve consistent results would be a shortcoming of the tool, unless something was seriously wrong with it. It's a bit funny that this anecdote is repeatedly used as proof that a light meter is useless, and that proper exposure could not be assessed by gauging the light falling onto the subject.

So we have one anecdotal photographer that never learned to use a light meter, and I have used one for half a century, and our opinions on its usefulness are different.

[pnodrog]

Again the same anecdote of the successful photographer that didn't learn to use his light meter. I cannot help but feel a bit sympathy for the bloke.

Any tool of the trade has to be learned. A light meter is no different from other tools in this respect. But personally, I would not assume that failure to achieve consistent results would be a shortcoming of the tool, unless something was seriously wrong with it. It's a bit funny that this anecdote is repeatedly used as proof that a light meter is useless, and that proper exposure could not be assessed by gauging the light falling onto the subject.

So we have one anecdotal photographer that never learned to use a light meter, and I have used one for half a century, and our opinions on its usefulness are different.

Light meters are fairly easy to use. Saying someone does not know how to use a light meter is a bit like saying a poor driver does not know how to read a speedometer. The skill that is lacking is the understanding of the tonal range in the scene and how to evaluate and apply or modify the exposure to capture the scene as you wish using whatever method or tools you prefer.

For the photographic work I frequently do a light meter would be fine(and look good) but not offer any advantage and simply slow me down.


P.S. Urban I have just read your very informative first post on this thread and realise that overall we agree completely on the main issues.

[ajohnw]

Light meters are easy to use. Saying someone does not know how to use a light meter is a bit like saying a poor driver does not know how to read a speedometer. The skill that is lacking is the understanding of the tonal range in the scene and how to apply or modify the exposure to capture the scene as you wish using whatever method or tools you prefer.

For the photographic work I frequently do a light meter would be fine but not offer any advantage and simply slow me down.

Thanks I didn't feel like replying. No point.

John
-