Metering tonality...18% or 12%?

Next, I tried to expose per the white towel metering +3EV which was mentioned...

Here's an example where (from memory) I'm pretty sure I spot metered the sunlit grass at +1. Clearly it is a bright, sunny day, and this exposure is +1.7 stops brighter than "Sunny 16". I'm sure you can appreciate how unsuited a "correct" exposure would be for this scene.

HTTP response: 404 | MIME changed to 'image/png' | Byte size: ZERO

And if we set the Brightness control back to 50, with Exposure set to -0.38, we once again have a white towel but we see that whites are still a bit clipped, but the center spike is well to the right of center.

Wilt wrote in post #12111519
Next, I tried to expose per the white towel metering +3EV which was mentioned...

...but this clips whites. It is possible to reduced clipping, by reducing the Exposure by -0.29EV in Lightroom as seen in this second screen shot...

But to completely eliminate clipping (when using white towel +3EV technique) requires reduction of Exposure by -0.38EV in Lightroom as seen in this third screen shot...

But by metering bright highlights at +3, when shooting raw, the raw data is not clipped. It is recoverable. It is only the tone curve, colour space mapping, camera profile and other default adjustments that create the illusion of clipping. If you use Rawnalyze to look at the actual raw data you will find it not to be clipped.

This is the same sort of processing that turns JPEGs into toast when you push just a little too hard. That's why you don't push to +3 for shooting JPEG. You back off and go only to +2.7.

Here's my own example of spot metering whites at +3. There is a tiny amount of apparent clipping, which is of little importance and easily remedied. ETTR doesn't get any better than this, IMHO.

HTTP response: 404 | MIME changed to 'image/png' | Byte size: ZERO

HTTP response: 404 | MIME changed to 'image/png' | Byte size: ZERO

HTTP response: 404 | MIME changed to 'image/png' | Byte size: ZERO

Yes indeed. And LR (or other converters) obfuscates (love that word) the matter even more because the scales under the sliders are entirely arbitrary and essentially meaningless. Does 0 Brightness mean no brightening? But all gamma correction is brightening of midtones (click the Linear box in DPP to see what an uncorrected image looks like). Or does 0 mean gamma 2.2 and -150 means linear? And what is 0 (no) contrast? Logically it should be undifferentiated middle grey, but not even -50 Contrast is that (although the concept of negative contrast is beyond my ability to conceive). And the question of questions, why is Brightness 50 the default?
Bottom line, you can't draw any conclusions from LR.

tdodd wrote in post #12111603
But by metering bright highlights at +3, when shooting raw, the raw data is not clipped. It is recoverable. It is only the tone curve, colour space mapping, camera profile and other default adjustments that create the illusion of clipping. If you use Rawnalyze to look at the actual raw data you will find it not to be clipped.

This is the same sort of processing that turns JPEGs into toast when you push just a little too hard. That's why you don't push to +3 for shooting JPEG. You back off and go only to +2.7.

Here's my own example of spot metering whites at +3. There is a tiny amount of apparent clipping, which is of little importance and easily remedied. ETTR doesn't get any better than this, IMHO.


and the same raw file displayed in DPP with highlight and shadow alerts enabled. No clipping at all....



Now let's look at the power of Lightroom to pull back a file where the highlights were metered at +4 and then the exposure reduced by -1.5 stops. Not bad for the wedding shooter.


Metering whites/brights/highlig​hts at +3 not only gives you great ETTR, but a little safety buffer too.

So a summary statement might be, "If you insist upon shooting JPG, you lose some headroom and highlight data, but if you take the time to use RAW you have more headroom and ability to preserve highlight data."

Wilt wrote in post #12112320
So a summary statement might be, "If you insist upon shooting JPG, you lose some headroom and highlight data, but if you take the time to use RAW you have more headroom and ability to preserve highlight data."

There is a phenomenon known as "false color" that can occur when converting RAW highlights if just one channel is clipped. Imagine a theoretical overexposed skin tone which should have the relationships between the channels of 280/250/230 but would have to be written 255/250/230. The camera and DPP avoid this problem by simply chopping off the data well below the top. LR/ACR, C1 and Bibble (and maybe others) have a recovery feature that reconstructs the clipped channel and thus maintains all the DR.

Wilt wrote in post #12112320
So a summary statement might be, "If you insist upon shooting JPG, you lose some headroom and highlight data, but if you take the time to use RAW you have more headroom and ability to preserve highlight data."

I find that raw offers a whole lot more benefit than that, although without question the ability to capture a larger dynamic range and then figure out how you want to squeeze it into your limited JPEG format it is useful.

So, our conclusions are...?

Well, gosh, I dunno, I guess what we've all been looking to do, shoot Raw and ETTR (just clicking the Linear button in DPP is pretty scary)...

My conclusions are that I don't care whether my camera is metering for 12% grey, 18% grey or anything else. It's more useful to me to know how my camera performs in real world shooting and what I need to do to optimise my raw data capture. Shooting to raw is all about capturing data. It has little to do with recording a photographic image. Creating the image comes later on.

Given the opportunity to meter highlights (not always possible) I'd far rather be in charge of where I clip those than worry about my mid tones.

In the absence of important highlights I can afford to pay more attention to the shadow end of the scale. I'm not going to waste valuable shadow detail simply in order to place mid tones "where they belong". I'll happily shift the whole tonal range up the scale so that I can better see what matters to the image. I'd rather keep the shadow tones above the noise floor if I can.

Well, it's interesting -- like I've said before, I don't often set a metering "standard" since I shoot in all kinds of conditions -- I think my most common is to set a bright blue sky to maybe +1 1/3 EV but I haven't gotten many blue skies around here lately! And, I've found that the blue sky setting yields a higher exposure than Sunny 16 -- what's up with that?

Anyway, this makes for two lively discussions this week! Whoo!

tonylong wrote in post #12086374
Linearly, one stop lower than white is half the light, and so a linear scale would show that as "half way" to black.

I thought dSLRs were more sensitive to the highlights (hence HAMSTER or ETTR). If more bandwidth was allocated to the hightlights then the histogram could not be linear. Other than that most statements made here went way over my head. I do like the pretty pictures though.

AbPho wrote in post #12115329
I thought dSLRs were more sensitive to the highlights (hence HAMSTER or ETTR). If more bandwidth was allocated to the hightlights then the histogram could not be linear. Other than that most statements made here went way over my head. I do like the pretty pictures though.

Heh! Yeah, that's where this thread went, over all our heads!

I am still working on reading all of this. Thanks guys. I prefer this over the usual stuff I read on POTN. I broadens my understanding of the equipment I am using and makes me better at it.

airfrogusmc wrote in post #12102293
I usually do several spot meter readings, find where my shadows are where my highlights are falling on the scale then make a determination of what is important to the visual statement.

I do exactly the same thing. I do not like using other metering methods because I do not trust them. I have my camera set to spot meter. When I go to photograph a scene I search around to make sure I have no blown hightlights or lost shadows. But I put emphasis on the subject. So, some things might get lost. I believe that every picture does not have to be (or cannot be) perfectly exposed all of the time.

What are these zones people speak of? Zone VIII? Zone II? Etc? I would like to know.

And what is UniWB? Are you talking about AWB?

I shoot quite a bit of BIF and very commonly I will meter off the sky, especially when it is very bright (not blue). If I know the sky will be the brightest part of the scene then I meter it +3. If I expect my subject to be white and even brighter then I build in a safety margin. I'll also verify my reading against my own palm and see if it makes logical sense. The acid test is to chimp a test shot.

Here are a couple of examples where I was pretty sure the sky would be the brightest part of the scene. I metered it at +3 in order to capture as much detail in my subject as possible, while still holding whatever detail/colour existed in the sky to be fine tuned.

HTTP response: 404 | MIME changed to 'image/png' | Byte size: ZERO

HTTP response: 404 | MIME changed to 'image/png' | Byte size: ZERO

HTTP response: 404 | MIME changed to 'image/png' | Byte size: ZERO

HTTP response: 404 | MIME changed to 'image/png' | Byte size: ZERO

HTTP response: 404 | MIME changed to 'image/png' | Byte size: ZERO