topic Re: Simple Exposure Question in EOS DSLR & Mirrorless Cameras

Simple Exposure Question

germanduder — Tue, 17 Apr 2018 20:27:12 GMT

Ive scoured the web for an answer too this question, not having any luck. I have an EOS 1200D (Rebel T5).

when using an incident light meter, and even with a gray card, in indoor situations and some others i am finding the histogram does not reach 255 (Rightmost side of histogram), sometimes it even ends a full stop or two before 255... I am a little confused, because based on incident readings and gray card readings, the exposure is correct... Even though the histogram does not touch the right end, are my exposures still correct? When importing to lightroom while preserving the picture style (usually camera standard) I still find the histogram does not hit the rightmost side, and end up needing to increase the exposure by a full stop and stretch the whites out to make the photo look well exposed...

Any ideas? I can consistently reproduce this issue. And yes, sometimes the material im photographing does not contain pure white, but underexposure is indicated by not hitting 255 (or within 1/2 stop near it) on histogram, correct?

This issue is causing me to have to do heavy post processing to make the photos acceptable, adding significant noise in the photo.

Re: Simple Exposure Question

ScottyP — Tue, 17 Apr 2018 21:18:44 GMT

Not all pictures have that much dynamic range. More often than not the histogram won’t stretch all the way between black and white. In processing in Lightroom you can choose to manually increase whites until the brightest white areas are as bright as they can be, without going over and blowing out highlights. Same thing with stretching the blacks left until they almost touch zero.

Re: Simple Exposure Question

germanduder — Tue, 17 Apr 2018 23:21:51 GMT

ScottyP
This I understand. The biggest example of this occurance was earlier this morning, shot a photo of a persons face (posing) under diffused sunlight indoors. I used evaluation metering since the light was mostly uniform, and set focus for the eyes. The result was a histogram with data only stretching out too the middle of the histogram. I found, after using the levels adjustment brush, that the face that was centered and metered for, fell in the left side of the histogram, and the parts of the image that were lighter were expressed as middle gray. Too me, this is underexposure.. I was under the impression that most skin tones are above or near middle gray, given this assumption, I am lead to believe it is underexposed, could my meter be at fault? I will need to take some test photos and post them in comparison with meter readings/ histograms to further explain my concern. I find even with large dynamic range scenes like outdoor landscapes, the Canon meter frequently leaves this gap in the histogram. The only method i have found that yeilds a fair histogram is by exposing to the right, where in post the image is displayed rather faithfully. I am coming from the days of film, and have been trying to apply the zone system to the histogram. It is hard to do this, when the meter seems to underexpose. Apologies if i sound like a complete noob, and/or ignorant.

Re: Simple Exposure Question

kvbarkley — Wed, 18 Apr 2018 01:42:00 GMT

Try the same shot with evaluative exposure and spot exposure on the face.

Re: Simple Exposure Question

ScottyP — Wed, 18 Apr 2018 02:45:44 GMT

I do ETTR about half the time. I also tend to use a narrower metering setting than Evaluative when shooting people if they are coming out underexposed.

Re: Simple Exposure Question

germanduder — Wed, 25 Apr 2018 14:03:56 GMT

Okay. I have done a bit of testing, and I'm finding while using a gray card or matrix, or spot metering using zone system, there is usually a full stop left unused in the right side of histogram still. I understand not all photos have that kind of dynamic range, but my photos are mostly landscapes without sun, or side lit from sun, all i am asking now, is if leaving this gap is considered underexposure. If I bring up the whites, where the sky meets the land the sky will be white, which it never was. Same with increasing exposure by a stop. Might I add my Canon T5 is a refurbished unit

Re: Simple Exposure Question

RobertTheFat — Wed, 25 Apr 2018 14:29:18 GMT

@germanduderwrote:
Okay. I have done a bit of testing, and I'm finding while using a gray card or matrix, or spot metering using zone system, there is usually a full stop left unused in the right side of histogram still. I understand not all photos have that kind of dynamic range, but my photos are mostly landscapes without sun, or side lit from sun, all i am asking now, is if leaving this gap is considered underexposure. If I bring up the whites, where the sky meets the land the sky will be white, which it never was. Same with increasing exposure by a stop. Might I add my Canon T5 is a refurbished unit

A camera sees things differently from the human eye. And ultimately what matters is what the human eye thinks it sees. I believe you're making a mistake by letting the histogram overrule your eyes. If a picture doesn't look right, the histogram may help you diagnose the cause. But if a picture looks right, trying to rectify its histogram at the expense of the picture's visual effect is the tail wagging the dog.

Re: Simple Exposure Question

Peter — Wed, 25 Apr 2018 15:59:53 GMT

ETTR in raw when the dynamic range of the camera is bigger than the dynamic range of the motive. This will bring the shadows out of the noise. Reduce the entire exposure in post.

If the dynamic range of the motive is bigger than the dynamic range of the camera, then you have to sacrifice the shadows or the highlights, or both. Or use a speedlite, HDR etc...

Re: Simple Exposure Question

germanduder — Wed, 25 Apr 2018 17:02:00 GMT

Thanks Robert. I see your point, i do print a lot of my work on a PRO 100, and i am more worried that my prints will not show adequate brightness with this gap in the histogram. You are right, even though there is this gap, the picture does look right with it. I guess i just have it stuck in my head that if the histogram is not fully used(populated with data from 0-255), than neither is the dynamic range of my camera, which seems wrong when im shooting scenes with a large dynamic range. It also seems the canon picture styles tend too boost the exposure a little bit in comparison with the adobe RAW conversion using their profile.

If only canon would implement a RAW histogram, and spot metering in my poor old T5 😞

Re: Simple Exposure Question

ebiggs1 — Wed, 25 Apr 2018 17:46:54 GMT

A histogram is a graphical representation of the pixels exposed in your image. The left side of the graph represents the blacks or shadows. The right side represents the highlights or bright areas. The middle section is mid-tones. That is your 18% grey. How high the graph reads is the number of pixels in that particular tone.

Gaps on either end indicate you are missing information. This means your exposure can be shifted without losing detail. It does not indicate an under or over exposure per say. If you shoot Raw you will have greater leeway to make adjustments in post. It is still a good idea to get it pretty close from the get go.

Re: Simple Exposure Question

ebiggs1 — Wed, 25 Apr 2018 17:50:42 GMT

" It also seems the canon picture styles tend too boost the exposure a little bit in comparison with the adobe RAW conversion using their profile. "

I don't quite understand what you are saying. In Raw there are no Picture Styles. P Styles are software additions to a Raw file in post when it is converted. You are shooting Raw mode?

Re: Simple Exposure Question

Waddizzle — Mon, 30 Apr 2018 22:41:38 GMT

Shot with a T5, and saved as RAW. Rokinon 14mm T3.1

Processed by Adobe LR6, because DPP does not process third party lenses.

The more photos you take, the better you will become.

Re: Simple Exposure Question

germanduder — Thu, 26 Apr 2018 13:07:50 GMT

egbiggs1, I'm taking about camera calibration, last panel on Adobe light room, options for camera standard, Adobe standard, etc. Yes i am shooting raw. In my case when i am utilizing that gap in the histogram as you say too do, my skies loose a fair amount of color. When trying to bring it back in post i will have less detailed clouds or gradients in sky, and if i use the highlight slider to bring some color/detail back i subsequently remove the highlight detail from things like wet grass reflections, etc. It is most noticeable during sunset hours when shooting away from the sun trying to capture the color gradient where the sky meets land, if exposing per your recommendation, this area becomes more white than colorful, which is not what I'm seeing in reality. When the weather permits, I will take two photos and post them too demonstrate this, one with the gap one without, and hopefully it will demonstrate the difference in highlights rendering i am experiencing

Re: Simple Exposure Question

RobertTheFat — Thu, 26 Apr 2018 13:36:48 GMT

@germanduderwrote:
egbiggs1, I'm taking about camera calibration, last panel on Adobe light room, options for camera standard, Adobe standard, etc. Yes i am shooting raw. In my case when i am utilizing that gap in the histogram as you say too do, my skies loose a fair amount of color. When trying to bring it back in post i will have less detailed clouds or gradients in sky, and if i use the highlight slider to bring some color/detail back i subsequently remove the highlight detail from things like wet grass reflections, etc. It is most noticeable during sunset hours when shooting away from the sun trying to capture the color gradient where the sky meets land, if exposing per your recommendation, this area becomes more white than colorful, which is not what I'm seeing in reality. When the weather permits, I will take two photos and post them too demonstrate this, one with the gap one without, and hopefully it will demonstrate the difference in highlights rendering i am experiencing

You are definitely overthinking this. The histogram is there as a tool to help solve exposure problems, not as a means of creating them.

With sunsets, you're bound to see bizarre histograms, because the lighting conditions in a sunset are themselves bizarre.

If you like to shoot away from a sunset, read up on "Alpenglühen" and "Gegendämmerung". The best articles I've found are in German; but if you can cope with that, there's some interesting information there.

Re: Simple Exposure Question

ebiggs1 — Thu, 26 Apr 2018 13:47:47 GMT

"...which is not what I'm seeing in reality."

As someone already said, Bob I think, what you see and what your camera sees can be and usually is different. We don't "see" like a camera sensor.

Also when you look at your histogram the lines can be very short. Make sure you don't delete something that might be hiding there.

"...my skies loose a fair amount of color."

Yes that can happen. All cameras have a limit to what they can do. You may have exceeded the DR of the sensor.

"...last panel on Adobe light room, options for camera standard, Adobe standard..."

I almost always, 99.5%, leave this set to Adobe Standard. I don't even know where it is set in my 1DX. It doesn't matter with Raw mode.

"... I will take two photos and post them too demonstrate this, one with the gap one without, ..."

Great love to see them. Pictures say more than words!

Re: Simple Exposure Question

TCampbell — Thu, 26 Apr 2018 14:54:59 GMT

The histogram shows the distribution of tonality. It isn't necessarily an indication of over or under-exposure (but it could be). If you see data touching the left or right edge of the histogram, then it indicates something in the image was "clipped" (beyond the range of what the sensor could handle). If the data is jammed to one side, but not the other, then it usually means you had an incorrect exposure (but this makes assumptions about what you intended to expose).

When I do concert shots in very very low light, the background is going to drop to black and the histogram will read that as under-exposure. Except... what I really care about is the performer and if THEY are correctly exposed (we don't care about the blackness of the backgrounds and frankly are happier that they are black).

One of the advantages of an incident light meter over a reflected meter is that the incident meter only reads the light "falling" on the sensor. It doesn't care what you are trying to shoot. Whereas a reflected meter is all about reading the light reflecting off your metering subject.

If you put a piece of white card stock and black card stock side by side and get in close enough to the "white" card so that it fills the frame and note that meter reading... then do the same filling the frame with the black card stock, you'll get a different reading even though both cards are sitting in the same light. The incident meter will give you the real reading (because it doesn't know it's sitting next to a white or black card and it doesn't need to know.)

For certain types of photos, you would expect ot see the hisotgram data favoring one side ... or the other. It wont necessarily just be in the middle and spread across from left to right.

If the histogram data doesn't take advantage of the full width that just means your camera is capable of more dynamic range than the scene required (better to have that ... then the opposite). You can always stretch the data in post processing ... but if you have the opposite problem you can't do anything to recover data that was clipped.

Re: Simple Exposure Question

BurnUnit — Mon, 30 Apr 2018 20:11:40 GMT

@TCampbellwrote:

One of the advantages of an incident light meter over a reflected meter is that the incident meter only reads the light "falling" on the sensor. It doesn't care what you are trying to shoot. Whereas a reflected meter is all about reading the light reflecting off your metering subject.

This does a nice job of clearing up some of the basic mysteries of reading a histogram and light metering. But I wonder if it might be worth making a small change to your description of the differences between reflective and incident metering.

Would it be more correct to say "that the incident meter only reads the light falling on the meter"? The way it's written in your reply might lead somebody to believe that you're referring to the camera's image sensor.

Too pedantic?

Re: Simple Exposure Question

TCampbell — Mon, 30 Apr 2018 21:07:59 GMT

@BurnUnitwrote:
@TCampbellwrote:

One of the advantages of an incident light meter over a reflected meter is that the incident meter only reads the light "falling" on the sensor. It doesn't care what you are trying to shoot. Whereas a reflected meter is all about reading the light reflecting off your metering subject.

This does a nice job of clearing up some of the basic mysteries of reading a histogram and light metering. But I wonder if it might be worth making a small change to your description of the differences between reflective and incident metering.

Would it be more correct to say "that the incident meter only reads the light falling on the meter"? The way it's written in your reply might lead somebody to believe that you're referring to the camera's image sensor.

Too pedantic?

That's certainly a fair statement. My term about the incident meter measuring the light falling on the sensor (of the incident meter... not the camera sensor). To rephrase that it's measureing the light landing on the incident meter would prevent confusion.

There are some nuances as well. (Some light meters have amazing capabilities ... they can read multiple data-points and provide the best exposure as a mean of all the sampled data-points; they can tell you when a scene will exceed your camera's dynamic range; they can meter flash; they can calculate "flash contribution" as a percentage of light relative to the ambient light; many also have integrated reflected meters as well as incident meters; etc.)

Reflected metering (some hand-held light meters that can be used for incident metering, also include a reflected sensor. If I'm shooting a landscape and want to know what the meter reading is on some snow-capped mountains in the background, it would be impractical to have to hike all the way up the mount to take a reading and then hike all the way back to my camera) refers to any light that first has to reflect off a subject before being picked up by the metering sensor.

When you use a Canon camera with "evaluative" metering mode (the default), the camera is sampling lots of points across the frame and trying to find the exposure that would best capture as many points as possible without over or under-exposure ... but it's willing to split the difference.

This is why, when shooting a tricky scene where there's a large part that is likely to clip (but isn't important), the camera wont know that. It can end up adjusting the exposure to try to re-claim some of the parts that are going to clip... and that can end up having an undersireable effect on the part of the scene you DID care about.

This is why the camera also has center or spot metering mode as well as "center-weighted" metering mode, etc. in combination with the AE-Lock feature. This allows you to meter the thing you care about, lock it, then re-compose and take the shot knowing that the camera will use the best exposure for your metered subject and ignore everything else.