1D Mk3 Exposure - Normal or not?

[colinsouthern]

Hi All,

 

I'm spending more and more time getting to know my 1D Mk3. Last week I did my

first serious landscape work with it - and I was amazed at how beautifully it

exposed the shot - the histogram was perfectly spread-out, nicely touching both

ends - couldn't have asked for more. However ...

 

... last night I was sitting on the couch, feet up on the coffee table testing

a few things. Room lights were on, but I had my EF24-70mm F2.8L USM mounted,

and zoomed to 70mm, taking shots of my foot wearing a "neutral grey" sock (I

know, not the world's most exciting subject) - the room lights weren't anywhere

near being in the shot, and nothing particularly bright in the background.

 

On every occasion the far end of the histogram sat about 2 full stops down from

the right hand side - I tried all combinations of ISO & shutterspeed - and

every metering mode, but only cranking up the EC by 2 stops sorted the issue

out. Anyone got any ideas as to why it would appear to underexpose by so much?

Am I missing something obvious here?

 

I did more tests later using a flash - same kind of thing (had to use FEC=+2 to

sort out), with the exception of Autoflash, which only seemed to be about 1

stop out.

 

Would appreciate any thoughts &/or suggestions.

 

Many thanks,

 

Colin

[dcheung]

um... as far as I know that is exactly what it is supposed to do.

Auto exposure will assume the metered area is neutral 18% grey (or whatever percent). When exposed properly, a white object should be on the far right side of the histogram and a black object should be on the far left side of the histogram. A white object should be exposed with EC +2 compared to a grey object. i.e., a white object is about 2 stops of light brighter than a grey object and should be on the right side of the histogram.

 

Given that, this means that if you are metering off the sock (and say that was the brightest subject in the frame), then the final picture should be exposed such that the grey sock would be two stops darker than what a white sock would look like. A white sock would be far to the right on the histogram (two stops higher than the grey sock). Your auto exposure is doing everything correctly.

[keith_lubow]

My suggestion is to read "The Negative".

 

Keith

[greg_moss2]

If you have an extra foot, try using a black sock, a grey sock and a white sock. Meter off the grey sock and see if white and black fall into proper alighment. You may have discoverd a new test target!

[PuppyDigs]

Indeed, if it recorded your neutral gray sock as a highlight I would be worried. Nevertheless, your camera meter is smart enough to know a neutral gray sock should reproduce as a mid-tone and fall in the middle of the histogram. A white sock should be about 2 stops above it.

 

If you want your gray sock to look like a highlight, overexpose it by 2 stops.

 

You can only create a "classic" textbook histogram if your scene contains a full range of tones from black to brilliant white.

[colinsouthern]

Hmmm ...

 

Good theories, but I'm still not getting anywhere.

 

I've upped the ISO to 1600 so I get some sensible shutterspeeds - if I meter off the grey sock I get 1/200th and an under-exposure of 2 stops - so in theory if I replace the grey socks with white ones it'll expose up to the histogram?

 

Well, I don't have any white socks, so I used a white card, carefully angled so as to not produce direct reflections from the overhead lights. With spot metering it meters at 1/800th, and exposes 2 & 1/2 stops down from the limit.

 

Seems like I need to keep a permanent EC=+2 to get it looking right?

 

Still confused ...

[colinsouthern]

OK - learning curve starting to flatten out a bit ...

 

I've got a coloured card from the back of one of my photography books, with 4 colours on it (Black, White, Midtone, and a light grey suitable for white balance corrections).

 

Reading up on exposure - if I've got the hang of it - I should be able to spot-meter the midtone - lock the exposure - shoot the entire card - and it should fill the histogram?

 

Surprise - it does.

 

Metering off any white object still ends up 2 stops underexposed - I think I now understand why (did a quick google on metering & exposure).

 

So if I've got this right, any scene with lots and lots of bright white (eg closeup of a brides dress), one needs EC=+2, and any scene with lots and lots of dark material will more than likely need EC=-2. Sound about right?

 

... I could have sworn that I've never had to be this proactive on the EC on the 20D.

[keith_lubow]

I would suggest that it might make it a bit more clear, as part of your photographic thought process, to not think of things in terms of EC, but in terms of tones on a tone scale.

 

Some thing like the following is, in my opinion, the most important piece of technical knowledge to have in photography:

 

1. Light meters are devices that measure light, and based on the measurement tell you how to expose in a way that will give you a middle grey tonality. Do not ever ever ever ever think otherwise, as this is all that they do. Tattoo it on the inside of your eyelids.

 

Therefore:

 

A. You fill the frame with a grey card, expose as your meter tells you, and you will have a grey card in the picture.

 

B. You fill the frame with a black card, expose as your meter tells you, and you will have a grey card in the picture.

 

C. You fill the frame with a white card, expose as your meter tells you, and you will have a grey card in the picture.

 

D. You fill the frame exactly halfway with a white card and exactly halfway with a black card, expose as your meter tells you, and you will have a black card and a white card in the picture.

 

This understanding is independent of what camera you are using. This is an issue of understanding your meter. This has been the nature of reflected light meters since their invention.

 

This is why a hand-held incident meter is the much preferred meter whenever convenience allows, UNLESS you are doing precise tone and contrast control. You take a reading of the light that exists at the scene, not the light that is hitting and reflecting off of objects. Thus, all the tones fall into place very close to accurately if you expose according to the meter's suggestion.

 

This, and more other information than you or I would probably ever use, is in the book I suggested. Yes, I realize it is about film, but that is petty irrelevant, as the theory is the same with digital.

 

Keith

 

Keith

[keith_lubow]

Also, I forgot to go into the whole histogram thing.

 

A histogram is nothing but a graphic display of all the pixels in an image that correspond to certain tones on a tone scale. It is nothing but a tally, really.

 

With digital, we are taught to expose in such a way that will give us the raw material we need to achieve the final product that we envisioned. (This is actually very similar to the way we are often taught to expose negatives.)

 

However, if you were to graph a histogram from an "ideal" transparency, it would not look like the histogram you are told to shoot for with digital. The ideal transparency would be extremely close to, if not exactly like, the desired final product. With digital, on the other hand, we are usually taught to expose in a manner that moves the lumps of the histogram as far toward the higher tones (rightward) as possible, without clipping the whites, and to do the tuning later. This is because the higher tones hold more information with digital, thus give us smoother and more "beautiful" tonal gradations. Less specific visualization is required with this approach, as you are only trying to pack in the raw material. This is where the suggestions of an "ideal" histogram come from.

 

While this makes sense to me in a purely digital and technical sense, I still expose digital as closely as possible to the final product, as I find that it makes the processing steps so much more smooth, and "reminds" me of what I originally envisioned, rather than just plopping down a bunch of raw info in front of myself and having endless options to tweak and manipulate it. I think that too many options for fiddling actually stunts creativity and blocks and/or distracts from the true nature of how I personally approach photography. Therefore, exposing based on the "ideal" histogram goes out the window when I am shooting.

 

I mention this so that you might consider experimenting by straying away from the "ideal" and basing your shooting on more than a histogram. It's easy to get stuck in a rut if you are always working for the "ideal" product.

 

Keith

[colinsouthern]

Thanks everyone - I've got it sussed now :)

[michaelschrag]

What I find remarkable is that he does not own a pair of white socks.

[colinsouthern]

Michael Schrag wrote:

 

"What I find remarkable is that he does not own a pair of white socks."

 

In all seriousness, a pair wouldn't help. I learned a lot about metering today - difficult to understand why I haven't come across this before - probably because I don't shoot a lot of socks on feet, and what I DO shoot usually has enough contrast in the scene to push the histogram both sides of -2 for a midtone.

 

Looking back it would be nice to be able to select a CF that automatically adjusts the exposure to the point where it aligns with the right-hand side of the histogram - even if it meant taking 2 shots in quick succession, with the camera simply looking at values & adjusting exposure after the first.

 

Cheers,

 

Colin