Light meter

[WayneF]

Camera meters are pretty good (although they are reflected meters), so there are really only two reasons for another hand held meter... For incident metering, or for metering multiple studio flash.

As Gary explained well, reflective meters (camera meters) meter the light reflected from the subjects color, not the light on the subject (except it's around 12% gray though, not 18%). So for example, black or white clothing can affect the portrait metering, we have to know to adjust the meter reading for the subject. But incident meters instead meter the actual light level incident on the subject, regardless of the subjects colors (an incident meter is aimed at the camera, not at the subject). So that's more accurate, little adjustment necessary. But incident is metered AT THE SUBJECT, and not at all automatic, so more awkward to meter than just doing it from the camera.

We can use a reflective meter to meter the light reflected from a known 18% gray card (held at subjects position), which is then like an incident meter (independent of subjects colors is the big plus), except then we have to open about 1/2 stop more to match the reflected meters (the 12% thing, the standard instructions on Kodak gray cards, 20 years ago when Kodak made gray cards). The specs in all Sekonic meter manuals say K=12.5, and Nikon and Canon are the same. Here is a Wikipedia article on these constants

If we have one hot shoe flash, it's easy to tweak its power level by trial and error, until we get a picture with the flash we want. A meter works, but this is usually easier than using a meter.

But if we have several studio lights, it's much harder, so the big deal is that we simply use the meter to set each flash to its proper power, for the ratio we we want. To know exactly what each light is doing, to know each light is exactly how we want it.

For example, maybe set the main lights power level to meter f/8 at the subject. If we want a one stop ratio, we set the fill lights power level so that it meters f/5.6 (at the subject). The light on a colored background is often set to around the same main light level (f/8, at the background), but black or white backgrounds probably are handled differently. The hair light varies with the color of the hair, maybe a stop brighter than main for dark hair, or a stop lighter than main for light hair (at the hair). The background and hair light do not affect the camera exposure, so we meter main and fill together to set the camera aperture (they will add, to be about 1/3 or 2/3 stop more than the brightest one). Anyway, the meter makes the multiple studio lights be quite easy, and quite repeatable next time.

You might see Why would I need a handheld light meter?

But unless we want incident metering, or have multiple studio flash, the camera meter is all most of us use. We learn to deal with the reflected meter.
See How Camera Light Meters Work

 

[mgblunt]

Camera meters are pretty good (although they are reflected meters), so there are really only two reasons for another hand held meter... For incident metering, or for metering multiple studio flash.

As Gary explained well, reflective meters (camera meters) meter the light reflected from the subjects color, not the light on the subject (except it's around 12% gray though, not 18%). So for example, black or white clothing can affect the portrait metering, we have to know to adjust the meter reading for the subject. But incident meters instead meter the actual light level incident on the subject, regardless of the subjects colors (an incident meter is aimed at the camera, not at the subject). So that's more accurate, little adjustment necessary. But incident is metered AT THE SUBJECT, and not at all automatic, so more awkward to meter than just doing it from the camera.

We can use a reflective meter to meter the light reflected from a known 18% gray card (held at subjects position), which is then like an incident meter (independent of subjects colors is the big plus), except then we have to open about 1/2 stop more to match the reflected meters (the 12% thing, the standard instructions on Kodak gray cards, 20 years ago when Kodak made gray cards). The specs in all Sekonic meter manuals say K=12.5, and Nikon and Canon are the same. Here is a Wikipedia article on these constants

If we have one hot shoe flash, it's easy to tweak its power level by trial and error, until we get a picture with the flash we want. A meter works, but this is usually easier than using a meter.

But if we have several studio lights, it's much harder, so the big deal is that we simply use the meter to set each flash to its proper power, for the ratio we we want. To know exactly what each light is doing, to know each light is exactly how we want it.

For example, maybe set the main lights power level to meter f/8 at the subject. If we want a one stop ratio, we set the fill lights power level so that it meters f/5.6 (at the subject). The light on a colored background is often set to around the same main light level (f/8, at the background), but black or white backgrounds probably are handled differently. The hair light varies with the color of the hair, maybe a stop brighter than main for dark hair, or a stop lighter than main for light hair (at the hair). The background and hair light do not affect the camera exposure, so we meter main and fill together to set the camera aperture (they will add, to be about 1/3 or 2/3 stop more than the brightest one). Anyway, the meter makes the multiple studio lights be quite easy, and quite repeatable next time.

You might see Why would I need a handheld light meter?

But unless we want incident metering, or have multiple studio flash, the camera meter is all most of us use. We learn to deal with the reflected meter.
See How Camera Light Meters Work

Thanks Wayne good info I appreciate your time!

 

[desertrattm2r12]

There are light meters that measure the light falling on your subject and there are flash meters that read how much light an electronic flash (strobe) puts out. Some of the former will give you an f-stop/shutter speed combination that will give you the right exposure. Some of the latter let you read the f-stop you need to use in a given situation. The shutter speed is up to the camera shutter, and how high a shutter speed will work with strobe.

 

[unpopular]

I really have fundamental reservations about incident metering. We're not photographing the light, we're photographing the subject. The light, without reflection or transmission contains no information it's intensity and spectrum. The subject is more important than the light itself.

Incident metering is only "more precise" if you're not taking into account dynamic range and tonal compression. Spot metering allows the photographer to decide how the exposure is made according to the subject, rather than being dictated by the available light; not the generic "middle of the road" exposure that you get from incident metering. Spot allows precise visualization of tonal range, allowing the photographer a concise overview of what compromises will be needed, and how those compromises affect editing choices - spot provides a framework which ties together exposure and post-exposure into a continuum.

I suppose ultimately, what I am saying is - the idea of a "correct exposure" is a really poor way of looking at photography. yes, exposure is technical, precise and scientific - but that does not mean it's cold and lifeless, it's not a sort of battle to overcome, to find the one correct combination of camera controls that is "the right one". Exposure is photography, not some technical inconvenience.

 

[WayneF]

I really have fundamental reservations about incident metering. We're not photographing the light, we're photographing the subject. The light, without reflection or transmission contains no information it's intensity and spectrum. The subject is more important than the light itself.

IMO, that is the wrong idea. It is said that photography is about the LIGHT.

Incident metering is only "more precise" if you're not taking into account dynamic range and tonal compression. Spot metering allows the photographer to decide how the exposure is made according to the subject, rather than being dictated by the available light; not the generic "middle of the road" exposure that you get from incident metering. Spot allows precise visualization of tonal range, allowing the photographer a concise overview of what compromises will be needed, and how those compromises affect editing choices - spot provides a framework which ties together exposure and post-exposure into a continuum.

Sure, Spot metering possibly can work, and possibly could help with dynamic range, but ONLY if you are going to take multiple different spot readings, of the highlights, and the shadows, and the important subject areas, and know what it means, and do some calculations, and do considerable thought to put it all together. Most of us are not trained for that. We can only have ONE exposure, one size must fit all. But yes, there are different ways of trying to determine that one exposure.

Since Spot metering is reflective (metering is affected by the scene colors and reflectivity), we do have to consider that a bright/white scene reflects a lot of light and reads too high, and black/dark colored scene reflects little light, and reads too low, so that unless both are compensated, BOTH will come out appearing a middle gray tone, neither of which is correct or useful. The only one advantage of reflective metering is that it can work from the camera position, but it meters differently dependent on whatever we aim it at. We have to KNOW to THINK about what we are aiming at. The only possible goal of reflective metering of the metered subject is that the average tone of the subject comes out as middle gray (which is surprisingly in fact often about right, on average scenes mixing many colors... which rules out Spot metering on only one color. But of course, there are many exceptions too, many scenes are not average.) I would suggest How Camera Light Meters Work

But reflective/spot metering is very dependent on the subjects colors/reflectivity. We learn to deal with that, but the big deal is that incident metering is independent of the subject colors, and is only dependent on the light intensity. The hypothetical picture of the black cat in a coal mine, or a polar bear on the snow, reflective metering gets both wrong, and incident metering gets both right. That's a huge plus for incident.

One problem with spot metering is that beginners don't understand it, and initially imagine if they spot meter on the face, the face will be exposed properly. That is very wrong, it only means the face will come out middle gray tone (which is all that reflective meters know how to do). The experienced know they must increase spot metering exposure perhaps one stop to make the face be natural tone (reflective always underexposes white things). This one stop increase is NOT metered or meterable of course, it's just a guess (and needs tweaking, but one stop is often near ballpark right if spot metering on faces. Of course it ignores all of the scene except the face.)

No reflective metering can be considered "correct". Perhaps metering on the gray card is close, which is merely an attempt to meter the light falling on an average subject (an attempt to meter independent of the subject).. Trying to be incident metering... When Kodak made gray cards (20 years ago), they always told us to open 1/2 stop more if metering on their gray card (because the card is 18%, but reflective meters are 12.5%.) But of course, some random subject is not likely either 18% or 12.5%. But otherwise, if independently of the subjects colors, if we meter to make middle gray be middle gray, then white and black will take care of themselves, in that range.

(My "black or white" and "middle gray" is speaking generically in old school terms for grayscale, and of course bright yellow and dark blue work the same way. All this stuff has been known for more than the 85 years we have had light meters. It is only newbies that don't know it yet. To make progress, they need to learn it.)

The idea of incident metering of the light intensity itself is if we get the light right, then white/bright things naturally come out white/bright, and black/dark colored things naturally come out back or dark colored. The "exposure" is indeed about the "light". If we get the light right, any subject comes out right. The girl in the black dress ought to get the same exposure as the girl in the white dress (specifically so that white will be white and black will be black), unless perhaps the entire idea is to show the detail in the dress, when possibly we might adjust it a little, to extend the range in one direction. But the normal reflective middle gray reflective result for both is hardly any correct idea of metering. Not without a thinking brain to help figure it out.

We have to learn to compensate reflective metering with our thinking brains to help it come out right. When we walk up to a white subject or wall, we already should know reflective metering will underexpose, and we must add more exposure. That requirement is hardly a plus for convenience. And it detracts from the concept of accurate metering. Whereas incident metering meters the actual light intensity, independent of the subjects colors (meter aimed away from the subject, aimed at the camera from the subject), so all the subjects come out about right. It "centers" the range, so to speak, around an average middle gray tone in the middle.

(middle gray does not mean one specific area, or even necessarily gray, it's an old grayscale term meaning a middle tone. It just means the tone of the averaged color of the averaged area of the scene is a middle tone, possibly near gray when all colors are averaged together, but the green forest will of course be a middle green. This is all reflective meters are able to do. They don't know what the subject is, or how it should be, they just see a blob of averaged color. Middle gray means that blob is made to be "not too dark, not too bright, but pretty much a middle range tone. That is all reflective meters can do.)

We can of course also compensate incident metering in the very few cases it might need a bit of help at the extremes of special cases (which we already saw and knew would be needed). But it is rarely needed, and on the whole, incident is the very easy way to bet. It just unfortunately is the more awkward way to meter (walking to the subjects position, except for bright direct sun cases).

 

[unpopular]

Before I start, I want to make it clear (even if it's not clear in our discourse) that we're starting to get into some philosophic differences in process, rather than hard-line technicals.

IMO, that is the wrong idea. It is said that photography is about the LIGHT.

This is one of those photography cliches that someone said once and people picked up without really thinking about how incredibly short-sighted and self-limiting it is. As I've said, vision and photography alike are measurements of interaction between materials and light. If you're going to want to photograph light itself, you're better off with a spectroscope than a camera. Which would make pretty dull pictures, unless you're really into rainbows.

Sure, Spot metering possibly can work, and possibly could help with dynamic range, but ONLY if you are going to take multiple different spot readings

I typically take one reading: the placement for Zone 8, which with a digital camera is around +2.3 to +2.6, depending on how well your camera handles highlights. In extremely contrast or low contrast situations I will also meter the shadows to see if this placement will result in either clipped shadows or extreme compression to the right. I then process the file for shadows. Ideally, the top end of the shadows should be at Zone 4-5.

It's not about getting it "right" because there is no such thing as "right". Sure, your 12 or 18% percent reference can be calibrated to Level 127 on the 8-bit scale, and with that you can say it's "correct". However that does not mean that what you're getting from the camera will match what you saw, or more important, what you perceived.

So really, here you have a choice, you can either throw up your hands and say "oh well, the lighting scenario is too contrasty/flat/whatever to take a good photo" or you can make an exposure that takes into account these conditions such that they can adjusted for in post ...

Now, you might be thinking I am advocating a "fix it it post" sort of attitude. I'm really not. What I am advocating for is that pre- and post-exposure is all *photography*, that it's a continuous process and that no one aspect should be considered without the other.

I think this SOOC attitude comes from one part machismo and two parts silver chrome photography which you really had a very limited range to work with and no real options in processing else everything goes bonkers.

Because "pros" shot chrome, I think this attitude that SOOC is better, no matter what was able to prevail - this despite that Adams' Zone System was significantly about post-exposure, only we didn't get to see the data encoded onto the negative the way we can with Raw files.

Most of us are not trained for that. We can only have ONE exposure, one size must fit all.

I don't think it is a good idea to have the attitude that because people lack the technical skills means that they should be given watered down advise, and I think the best place to start with reflective metering is looking for a single reference, such as green grass or 90-degrees from the sun on a clear day and spot meter from there. This forces people to look for tonal regions and understand how to break up scenes according to brightness - a skill that is the foundation for pre-visualization.

But moreover, the search for the "correct" exposure is really a loosing proposition. There is, of course, the "wrong" exposure, namely images with regions that contain no useable information where useable information should be. There are images that are "too bright" or "too dark" - but these are more aesthetic problems that originate from exposure than "wrong exposure" in itself. But I reject the idea that there is a single "correct" exposure. It really depends on how the scene was understood by the artist making the picture.

This one stop increase is NOT metered or meterable of course, it's just a guess (and needs tweaking, but one stop often near ballpark right).

So I think there are a couple of points you're missing on this one though. First, it is not about where a tonal region "is" it is where it "should be". If you wanted to emphasize paleness or darkness in a skin tone, you wouldn't expose with some sort of photometric precision. The exact measured value of brightness is not important. We don't place little stickers under photographs indicating the number of photons that are reprensented by each level of grey, nor graphs and charts that somehow represent the spectral analysis of the scene in CMYK space with (if such a thing were even possible!).

Just as light has no meaning without subjects, subjects have no meaning without perception. Even Adams, who sought perfect objectivity rendered skies in ways that I find disturbing - but this does not mean Adams was wrong, only wrong to assume it was objectively "right".

So I think it makes more sense to encourage beginners to understand their meter, their camera, it's controls and how those controls affect results. Incident metering just doesn't provide this - as you say, it's a single "set it and forget it" sort of affair.

For the kind of photography we're talking about exposure shouldn't be this big inconvenient step that should be deligated to the least intrusive level possible. Granted, for action photography and photojournalism it is - though I am sure every photojournalist who is a photographer first wishes that they had a level of control over exposure which they cannot afford.

There is this huge temptation in these new fields to squarely place a dichotomy between the technical and the artistic. But this is like saying you can be a painter but not have any sense of how to paint. No matter what medium we're using, we cannot avoid the fact that we have to understand the tools in order to use them to a full extent.

And since Spot metering is reflective (metering is affected by the scene colors and reflectivity), we have to consider that a bright/white scene reflects a lot of light and reads too high, and black/dark colored scene reflects little light, and reads too low, so that unless both are compensated, BOTH will come out appearing a middle gray tone, neither of which is correct or useful. [...] No reflective metering can be considered "correct".

Provided that you are spot metering (and not weighted or average), you only have to compensate for one and adjust for the other. In fact, you can't expose for both, unless you're doing an exposure merge of on or the other. In this setting, I would expose for the snow compensate for the level of detail I want, then meter the shadows to check if they are within dynamic range given the chosen placement of the hilights.

If the shadows are beyond the camera or film's latitude, I can make a decision about how the exposure is to be made, or the image processed. I can determined what regions can be clipped and to what extent. I can determine if multiple exposures are required, and how those exposures should be made.

If fact, every scene has one reference point that is absolutely fixed - the brightest diffuse region. Set this to the upper-most latitude and everything will be at maximum exposure possible. In typical situations, all regions will contain usable data. The image will often appear flat SOOC and will need editing, but the data is good.

I am a huge advocate of referencing to the extremes rather than the middle because those extremes are really where the limitations of the camera are.

I had started to develop an entire system, in fact, to calibrate cameras and make notes about zone placement. The idea was you could calculate exactly what gamma adjustment to apply in order for middle grey to render accurately. But I quickly discovered why Adams did not encourage photographers to think in these terms - it's just not particularly useful.

The only one advantage of reflective metering is that it can work from the camera position, but it meters differently dependent on whatever we aim it at. Reflective/spot metering is very dependent on the subjects colors/reflectivity. We learn to deal with that[...]

It is interesting to me how much we differ in approach here. What you see as a significant limitation, I see as the primary advantage, and what you see as liberating about incident metering, it's subject independence, I see as substantially limiting.

 

[WayneF]

I typically take one reading: the placement for Zone 8, which with a digital camera is around +2.3 to +2.6, depending on how well your camera handles highlights. In extremely contrast or low contrast situations I will also meter the shadows to see if this placement will result in either clipped shadows or extreme compression to the right. I then process the file for shadows. Ideally, the top end of the shadows should be at Zone 4-5.

Sounds to me that you are judging both the scene tone and the result tone by eye, just guessing at what it might be, or ought to be, but not actually metering either it or the light.

It's not about getting it "right" because there is no such thing as "right". Sure, your 12 or 18% percent reference can be calibrated to Level 127 on the 8-bit scale, and with that you can say it's "correct".

It should be mentioned that 18% only comes out "near" 127 (actually about 117 or 46%) only because our histograms are gamma encoded. We only see gamma histograms, and gamma is not related to exposure. In linear data, 18% is 18%. In the gamma data we see, our so-called midpoint of 127 is actually about 186 in our gamma histogram, near 3/4 scale (but cameras also do white balance and contrast and color profiles and saturation, etc, so this number becomes only an approximation, never very exact).

I think this SOOC attitude comes from one part machismo and two parts silver chrome photography which you really had a very limited range to work with and no real options in processing else everything goes bonkers.

Because "pros" shot chrome, I think this attitude that SOOC is better, no matter what was able to prevail - this despite that Adams' Zone System was significantly about post-exposure, only we didn't get to see the data encoded onto the negative the way we can with Raw files.

Do you still shoot film B&W? If so, I can better understand your position.

Digital is not so much different than chromes. Whereas negative film had very wide latitude, almost anything could be saved and processed in the dark room. Until 1960, Kodak rated all ASA film speed at 2x what they were, just for a safety factor of exposure (few had meters until then, Sunny 16 was the only guide, until meters became popular around 1960 when more cameras started adding meters, and made that 2x be embarrassing). The philosophy of using negatives was to "expose for the shadows", which is of course overexposure, but negatives could deal with it, and certainly it helped the shadows. But digital is sort of touchy about clipping at 255, so overexposure is not really a choice. Expose to protect the highlights in chromes and digital.

The Adam's Zone System was about printing in the 1930s. Negatives manipulated in the dark room, to be what they could be made to be. There were very few light meters then, in that Zone System era. First Weston meter in 1932. Adams had one though, no doubt given one of the first. They were expensive, and it was the depth of the depression. The System was about printing.

I don't think it is a good idea to have the attitude that because people lack the technical skills means that they should be given watered down advise,

That is however just the facts of life, but incident meters are very good advice, and IMO, probably universally used in the studio. Few novices listen, but of course, they do still have larger problems.. I've learned reflective meters, and don't bother to use a handheld meter for most outdoor work, but incident seems mandatory in the studio.

So I think there are a couple of points you're missing on this one though. First, it is not about where a tonal region "is" it is where it "should be".

Which again, is about some human guessing judging what it should be, not about metering what is. I am in fact discussing actually metering the scene.

So I think it makes more sense to encourage beginners to understand their meter, their camera, it's controls and how those controls affect results. Incident metering just doesn't provide this - as you say, it's a single "set it and forget it" sort of affair.

I wholeheartedly promote understanding our tools, more should try that. And yes, incident metering is greatly automatic, it does make precise metering be very simple, which really is otherwise much more complex. Reflective meters only show the reflection from the subjects colors, whatever that is... it depends on the subject colors, it's not about metering proper exposure. That includes Spot metering. Yes, beginners should be taught this, because it is how things work. To aid that effort, see How Camera Light Meters Work

If fact, every scene has one reference point that is absolutely fixed - the brightest diffuse region. Set this to the upper-most latitude and everything will be at maximum exposure possible. In typical situations, all regions will contain usable data. The image will often appear flat SOOC and will need editing, but the data is good.

It seems more important to know if that brightest highlight has any importance to the subject. Probably most don't.

Adams had a good rule that maximum dramatic contrast was very important to grayscale that sells, and that all pictures should have some very white and some very black areas to insure that dramatics. His Zone System was to promote that concept of extremes.

Color today is sort of different however, excessive contrast can be a minus, garish, not a plus as in grayscale.

I am a huge advocate of referencing to the extremes rather than the middle because those extremes are really where the limitations of the camera are.

But the middle is normally our subject. I think the time for dramatic grayscale has passed now.

It is interesting to me how much we differ in approach here. What you see as a significant limitation, I see as the primary advantage, and what you see as liberating about incident metering, it's subject independence, I see as substantially limiting.

Yes, we do agree on that point, that we do disagree. Different goals I guess, to each his own.

 

[unpopular]

Sounds to me that you are judging both the scene tone and the result tone by eye, just guessing at what it might be, or ought to be, but not actually metering either it or the light.

Not exactly. I still need to know the local brightness as measured by the meter so that I have a reference point to start from, and from that reference I can compare everything else. If I place one area at Zone 8, which is where I typically meter from (a la ETTR), then I can judge the relative brightness of every other region to determine relationships and estimate noise levels, clipping, gamma corrections, etc.

Do you still shoot film B&W? If so, I can better understand your position.

I come out of a b/w film background, so naturally, yes, I do tend to think this way.

Digital is not so much different than chromes.

Digital is sort of touchy about clipping at 255, so overexposure is not really a choice.

Digital is touchy and the highlight is absolute. There isn't any rolloff before processing. However, there is a lot of tonal compression going on also and raw files that look good SOOC will appear vastly under-exposed without gamma correction. I couldn't tell you where exactly middle grey is according to the camera (it can be calculated), but from my experience it's around Level 160-180 in 8-bit terms; quite a way to the right from 127.

This is the principle of ETTR, and one of my technical goals is to precisely determine how to expose to the right rather than the popular approach of just "slightly over exposing" to some arbitrary level. Without knowing how much light is being reflected off the subject, this isn't possible.

The Adam's Zone System was about printing in the 1930s. Negatives manipulated in the dark room, to be what they could be made to be. There were very few light meters then. First Weston meter in 1932. Adams had one though.

The Negative, the book which covers metering and exposure in conjunction with processing, was published in 1949.

Which again, is about some human guessing judging what it should be, not about metering what is. I am in fact discussing actually metering the scene.

Which again, to circularly return, says nothing about the subject nor the camera.

It seems more important to know if that brightest highlight has any importance to the subject. Probably most don't.

This starts to go into some SNR and ETTR. But basically the way I think of things is the RAW as the unprocessed film, the TIF as the negative and the PSD and the print. So my approach is to record as much data as possible, process the raw file such that the tonal curve is accurate and do any local adjustments in Photoshop.

So for me, I expose for the hilights and develop for the shadows. Like I said, this is really all starting to get into philosophy and process.

Color today is sort of different however, excessive contrast can be a minus, garish, not a plus as in grayscale.

But the middle is normally our subject. I think the time for dramatic grayscale has passed now.

My technique typically results in extremely deep shadow detail that I tend to hoard. Typically, the complaint is my work is appears too flat because I try to preserve everything. Again, this is a result of highlight compression inherent to processing linear data into nonlinear colorspace.

So by metering for the highlight nothing blows because you've set the maximum exposure at the brightest region. However, the RAW file will at first appear washed out because you're reading deeper into the shadows.

 

[WayneF]

I come out of a b/w film background, so naturally, yes, I do tend to think this way.

Yep, been there, done that, for too many years. I know the drill.

The Negative, the book which covers metering and exposure in conjunction with processing, was published in 1949.

Yeah, not sure if very applicable to digital now, but I still have that book and his others here someplace. I was not seriously into dark room until nearly ten years later when I bought the books. I never bothered getting into the Zone System, but I knew the concept about printing the extremes. And I stayed with B&W far too long, but printing color was sort of difficult at home for years. I had a color head, and I tried it now and then all along, but I never got an automatic processor. Cibachrome was my best efforts. Then when digital came along, it was so much easier, and results much better. I'm a big fan of digital, and raw, and incident meters, all are the easy way, which I don't mind at all. (FWIW, I rarely use Incident outdoors, but I did learn how to use reflective metering). Digital is tremendously easier than film ever was - but film was fun too, in its day. I really don't miss film at all. I cannot imagine those limitations now.

So for me, I expose for the hilights and develop for the shadows. Like I said, this is really all starting to get into philosophy and process.

Protecting against digital clipping seems a necessary requirement, not a philosophy.

Regarding clipping, I do worry about those beginners trying to use the cameras single gray histogram (which doesn't show clipping), instead of the necessary three RGB histograms it also furnishes. They can see: Surprises in the Use Of Histograms

 

[unpopular]

My experience on my A700 the histogram is close (not exact). I have never calibrated my X-E1.

As for protecting against clipping, I think spot metering is very capable provided that you know your camera and can accurately meter the upper end of diffuse reflections. Incident metering can as well, but you have no control over anything else. Sure you can address this in post, but you're far more limited since the image is all compressed to the left.

I do think The Negative still applies, at least in principle. I actually find that digital has more in common with b/w than it does with color chrome.

 

[WayneF]

My experience on my A700 the histogram is close (not exact).

You really ought to check to know that the single gray histogram shows luminosity, instead of actual real RGB data.

To make a believer of you, this A700 link

Sony Alpha DSLR-A700 Digital Field Guide

(on that first page shown here, page 94) clearly shows the A700 single gray histogram is luminosity, instead of real actual RGB data. Notice that showing the picture of the red door (a single color makes it be much more clear, much less confusing), Luminosity shows a fake peak not actually in any RGB channel. That is just an artificial math manipulation (called luminosity of grayscale), which cannot show clipping for RGB, for the reasons shown at my link. In Luminosity, each RGB channel is multiplied by a small fraction, incapable of showing any previous higher value it might have been, such as clipping at 255.

Look also at second histogram on second page (page 95). Same thing.. . Luminosity shows a peak NOT in any of the RGB channels. Luminosity is not Real RGB data in the sensor. Luminosity is a "what if" concept of how would its brightness appear if it were grayscale? But it is NOT grayscale.

This is a common thing, not any exception.

We should only watch the three actual RGB channels. No one questions the three RGB histograms... they are real data.

 

[unpopular]

It's been a while since I've used the A700 (I actually am not even sure where it is ATM), I seem to remember there being an RGB historam present; maybe I just used the blinkies? I do know that it will show clipped channels, it has a tendency to clip the blue channel's low values and I know i've seen this happen in a mustard field.

I look more to the histogram for shadow distribution than highlight clipping anyway. Blinkies work better for that since it gives me an idea of where I should have metered from if I missed.

I, like you, prefer no clipping at all. I have known some that don't mind clipping the specular. I like to put the specular at the uppermost edge of the histogram so that rolloff can be better managed. This isn't always possible, of course, but I agree, hilights should ideally never be clipped entirely.

The histogram and preview has pretty limited value though if you're shooting and processing raw as it is derived from a converted JPEG. Raw conversion does all sorts of nasty things to your files. Scary, aweful, nasty things that we best not talk about

 

[unpopular]

Yes, actually, come to think of it I know it did. The color histograms wer stacked up one on top of another. I'm unsure about the XE-1.

But again, I don't really use the histogram that way.