Metering tonality...18% or 12%?

I can't resist throwing in my two cents here, Although I know what I am about to say is going to make steam come out of Wilt's ears.

I think grey cards are outmoded and obsolete holdovers from from film days that should go the way of buggy whips. When I use an exposure aid it is a white target, specifically a white terry-cloth washcloth, which represents the brightest textured highlight (Zone VIII). I meter it and add 3 stops because current DSLRs have roughly 3.5 stops above medium grey to clipping in RAW and I think that nailing the highlights is more important than the middle tones. Alternatively, in a landscape I spot meter the sky and add two stops. But in dynamic situations, cards and ETTR are equally inappropriate, so it's back to CW metering and chimping the histogram.

tzalman wrote in post #12101099
I can't resist throwing in my two cents here, Although I know what I am about to say is going to make steam come out of Wilt's ears.

I think grey cards are outmoded and obsolete holdovers from from film days that should go the way of buggy whips. When I use an exposure aid it is a white target, specifically a white terry-cloth washcloth, which represents the brightest textured highlight (Zone VIII). I meter it and add 3 stops because current DSLRs have roughly 3.5 stops above medium grey to clipping in RAW and I think that nailing the highlights is more important than the middle tones. Alternatively, in a landscape I spot meter the sky and add two stops. But in dynamic situations, cards and ETTR are equally inappropriate, so it's back to CW metering and chimping the histogram.

Although I'm a neophyte relative to the posters on this thread, I have to say that I do what tzalman says above, with a caveat. My impression from the reading I've done is that after the RAW converstion, because of how the converter converts, significantly pushing or pulling the exposure back to where it should have been may not get you the same as if the exposure was correct at the time of exposure.

Great thread.

Great thread indeed, I've learned a lot..
Thanks guys!

tzalman wrote in post #12101099
I can't resist throwing in my two cents here, Although I know what I am about to say is going to make steam come out of Wilt's ears.

I think grey cards are outmoded and obsolete holdovers from from film days that should go the way of buggy whips. When I use an exposure aid it is a white target, specifically a white terry-cloth washcloth, which represents the brightest textured highlight (Zone VIII). I meter it and add 3 stops because current DSLRs have roughly 3.5 stops above medium grey to clipping in RAW and I think that nailing the highlights is more important than the middle tones. Alternatively, in a landscape I spot meter the sky and add two stops. But in dynamic situations, cards and ETTR are equally inappropriate, so it's back to CW metering and chimping the histogram.

No steam! When new techniques make sense (like ETTR does for better use of the 4096 levels of brightness within the camera RAW data), I am an advocate. However, ETTR done blindly (i.e., move everything to the right, but then do NOT restore things back in postprocessing to their inherent tonality!) leaves the black cat in the coal mine unnaturally whitish gray!

The technique which you use preserves highlight detail when using a white towel surrogate for the 18% gray card...that in basic concept is no different than using the palm of your hand as a +1EV surrogate for a gray card, or grass or the sky, etc. ( -- other than provide a better standard for preservation of highlight detail.)

But one thing for which a gray card has enduring need is to provide a reference standard for tonality, so that you know when you have counterbalanced ETTR back to the left the correct amount. A MacBeth card does the same thing. Even the brighter-than-18% WhiBal card does the same thing...they ALL provide a visual reference of inherent tonality to restore things to.

tzalman wrote in post #12101099
I can't resist throwing in my two cents here, Although I know what I am about to say is going to make steam come out of Wilt's ears.

I think grey cards are outmoded and obsolete holdovers from from film days that should go the way of buggy whips. When I use an exposure aid it is a white target, specifically a white terry-cloth washcloth, which represents the brightest textured highlight (Zone VIII). I meter it and add 3 stops because current DSLRs have roughly 3.5 stops above medium grey to clipping in RAW and I think that nailing the highlights is more important than the middle tones. Alternatively, in a landscape I spot meter the sky and add two stops. But in dynamic situations, cards and ETTR are equally inappropriate, so it's back to CW metering and chimping the histogram.

Perfectly said. Once you know where you want tones to fall and then figure out what is important to your visual statement then you can make the decision on what you loose, if it is or isn't important to the visual statement which may or may not be the way the scene is in reality (pre-visualization)or you may find you wont loose anyhting. 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. Maybe the shadows are falling below a zone II but the highlights are at zone VIII. The highlights are very important but the shadows really mean little to the image. I'll expose to maintain the highlights instead of bumping up my exposure to preserve the shadows. The shadows are not important. In another situation the opposite might be true.

But lets throw this into the mix. In the old days there were 1/2 stop tolerances at Nikon. So that means there was 1/2 stop in meters, 1/2 stop on lenses and 1/2 stop on shutter speeds. So there could be potentially a full stop or maybe more of two camera/lens combos that were produced on the same assembly line. I think tolerances in todays digital camera world have gotten tighter but they're still there and that would effect everything we are talking about or no?

In large format zone system you'd do the test to see what your proper ISO is with each lens, meter and camera you are using because there are manufacturing tolerances. If this was already discussed in this thread sorry I must have missed it.

tzalman wrote in post #12101099
I can't resist throwing in my two cents here, Although I know what I am about to say is going to make steam come out of Wilt's ears.

I think grey cards are outmoded and obsolete holdovers from from film days that should go the way of buggy whips. When I use an exposure aid it is a white target, specifically a white terry-cloth washcloth, which represents the brightest textured highlight (Zone VIII). I meter it and add 3 stops because current DSLRs have roughly 3.5 stops above medium grey to clipping in RAW and I think that nailing the highlights is more important than the middle tones. Alternatively, in a landscape I spot meter the sky and add two stops. But in dynamic situations, cards and ETTR are equally inappropriate, so it's back to CW metering and chimping the histogram.

Elie, I'm curious about your setting white to +3 EV! I've never thought that as an appropriat amount, I've thought more like +2 EV, but then I've never really tested it out. In fact, my 5DC only goes to +2 on the metering scale. My 1D3 goes to +3, though.

When I get bright highlights I tend to work off the blinkies!

Do others find the same thing (set white at +3 EV)?

Also, I am in the habit of setting a bright blue sky at say +1 1/3 EV, not at +2, so maybe I'm missing something...?

Tony I can tell you that I would do is take a spot meter reading and place the hottest white I needed detail in at a zone VIII. Shoot manual and meter the white and open up 3 stops from that reading. Should hold pretty good detail in whites now those would be some of the whitest areas in the scene not the specular highlights which in some cases I don't mind going off the scale. ANd this isn't an absolute just what I do in some situations depending on what i've determined is important detail in the image.

Not sure if tzalman would do it that same way but I gotta feeling he would.

tonylong wrote in post #12097589
We do need to boil this all down to what a layperson can comprehend and apply to photography!

I'm not sure this started off as something intended to be applied practically. I read it as more of a theoretical rambling (and I mean that in a nice way; I like that sort of stuff) about the coincidence of the alternative metering with the appearance of the gray card at the center of the histogram.

IMO Wayne's contribution was fantastic, explaining that there was no relationship between 18% gray and the center of the histogram that deserves to be enforced purely on its own merit (although it isn't necessarily bad).

Since reflective metering on 18% seems to work so well for so many people for so long, is there really a need to prepare a guidance statement that goes beyond "increase by 1/2 stop if it works for you"?

Working in M, I just move a wheel 9 clicks. Below are two versions of a white cloth with +3 exposure. I shot it with the 5D2 in UniWB for the purpose of this illustration so that DPP's RAW histogram would more closely reflect the RAW data and the second version is after click WB. As you can see, in the first version there is lots of headroom before clipping (when I move the Brightness slider to the right, the clipping warning starts to appear at +0.67 - gotta love that 5D2) and even after WB it is still just short of clipping.

Warning to jpg shooters: When the camera processes a jpg it routinely clips about a stop off the top of the capture data, thus a jpg will blow at roughly 2.5 stops above medium grey.

HOSTED PHOTO
please log in to view hosted photos in full size.

HOSTED PHOTO
please log in to view hosted photos in full size.

tzalman

are you shooting with UniWB as your default?

windpig wrote in post #12109390
tzalman

are you shooting with UniWB as your default?

No, I usually don't use it, but I thought in this case it would help illustrate the point. I don't use it because I realized that when you have a tri-color histogram the green channel, which is hardly affected by WB, closely resembles the Raw data in most cases. If the green isn't clipped, the red and blue won't be either in normal lighting.

tzalman wrote in post #12110657
No, I usually don't use it, but I thought in this case it would help illustrate the point. I don't use it because I realized that when you have a tri-color histogram the green channel, which is hardly affected by WB, closely resembles the Raw data in most cases. If the green isn't clipped, the red and blue won't be either in normal lighting.

Thanks

Good stuff, Elie, and thanks Allen!

Like I said above, I just never have tried to meter off a white target. Maybe a snowy scene a long time ago, but never have nailed it down to a certain level, so knowing the three-stop thing can be helpful. I'll have to play with it. I've pretty much just followed the avoidance of "blinkies".

In further investigation into the thread topic, I decided to photograph the PhotoVision target, to see where the 18% area falls within the histogram, compared to the white area and the black area. Shooting with the 18% area metered with the spot meter, and zeroing all the values in Lightroom, this is the result...18% area spike is just to the left of the center.

tonylong wrote in post #12103539
Do others find the same thing (set white at +3 EV)?

Yep. I shoot almost exclusively with manual exposure and spot meter my important highlights at +3, shooting raw, of course. This will capture sunlit feathers on a swan, or a sunlit snow scene beautifully. This is something that frustrates me with the +/-2 stop meter on the 5D2. The 7D and 1D3 are far quicker to use and adjust on the fly for +3 stop metering. If I see a hint of clipping then I am delighted. I shoot with neutral picture style in order to flatten the contrast and sharpening in order to avoid false positives on clipping.

If you shoot straight to JPEG, or with Nikon, then you need to be less aggressive.

If I have no highlights in the scene - dark brown dog against green grass (which I do shoot quite often) - then I meter off the grass at between 0 and +1. This gives me detail in the dog's fur that would simply disappear if I was to make do with standard grey card metering. If there was brilliant white in the scene, such as that swan again, the green grass would probably need to be metered at -2/3. That is a disaster for shooting my dog. She turns black. I also use my palm at +1.3 as a grey card substitute for standard mid tone metering that will hold bright whites. Again, for the dog I'll boost that by a stop or so.

As an example, here's my dog shot in bright sunshine at 1 stop brighter than "Sunny 16", no edits, then with exposure set to -1 in DPP. I know which version looks closer to reality and which one I'd rather have as a starting point. Quite honestly, mid tone metering, or using a handheld meter is very limited in the real world. In the real world you need to get more sophisticated than that. I believe one should strive to optimise exposure for the scene content. Mid tone metering fails to deliver on that score.

HOSTED PHOTO
please log in to view hosted photos in full size.

HOSTED PHOTO
please log in to view hosted photos in full size.