5d Mark III - Metering Underexposed

[rjs1981]

So first off, I apologize as I think I've seen a similar thread here but I can't find it anymore!

I purchased a 5d Mark III as an upgrade from my 7d a couple of weeks ago and, although I am happy, there is one main issue that is bothering me.

When I meter a "perfect" exposure, the camera is delivering about a full stop UNDER that.

My normal settings are; 

Shoot in M

Center focus point

Spot metering

AI Servo.

I generally shoot 2/3 over anyway, and find that I'm needing to now shoot at least 1 2/3 over to get the same result as my 7d. In addition to that, I'm getting a noticeable vignette in lower (but even/consistent) light situations.

I have tried all the different metering options, including variations on the focus points. I have done a full factory reset and the issue is consistent across all my lenses. 

Any ideas?

[RobertTheFat]

---

@TCampbell wrote:

---

@RobertTheFat wrote:

What isn't explained in that reference is how the 7D and the 60D map their 63-zone metering system onto a sensor with only 19 and 9 autofocus points, respectively. Incidentally, the URL you provided includes a trailing comma that must be removed in order to allow the link to be followed.

---

The AF sensors are located on the floor of the mirror body (just below the reflex mirror).  The metering sensors are up in the viewfinder above (also, this is why we are cautioned to cover the viewfinder to get more accurate metering if our eye isn't blocking the viewfinder from stray light.)

Here's a diagram that may help (note: this is from a 30D):  http://www.imaging-resource.com/PRODS/D30/D30A3.HTM

---

Canon's overview, cited by cale_kat in a reference to their Web site, seems to say that in each of the 63 metering zones one of the autofocus points contributes to the analysis of the exposure required. It's straightforward to see how that might work on a 5D3, which has 63 autofocus points. It's less obvious how they'd handle it on a 7D or a 60D, both of which have far fewer than 63 autofocus points. A diagram of a 30D isn't likely to clarify the point, inasmuch as the 63-zone metering system didn't exist on the 30D.

[TCampbell]

---

@RobertTheFat wrote:

Canon's overview, cited by cale_kat in a reference to their Web site, seems to say that in each of the 63 metering zones one of the autofocus points contributes to the analysis of the exposure required. It's straightforward to see how that might work on a 5D3, which has 63 autofocus points. It's less obvious how they'd handle it on a 7D or a 60D, both of which have far fewer than 63 autofocus points. A diagram of a 30D isn't likely to clarify the point, inasmuch as the 63-zone metering system didn't exist on the 30D.

---

They only need a metering point which is close.  No need for a one-to-one mapping.

Even on the 5D III where there's a 61 point AF system and a 63 zone metering system, the AF points don't map directly to metering points.  

See:   http://cpn.canon-europe.com/content/education/technical/inside_canon_eos_5d_mark_iii.do  

On this page, expand the "Metering & Exposure Control" section to view the diagram.

The entire top and bottom row of metering points have no AF points in them at all.

[ebiggs1]

"They only need a metering point which is close.  No need for a one-to-one mapping."

Exactly.  The point I was trying, poorly, to make.

[Hannahgivas]

So my question is how do I master the metering on 5d mkiii, I'm having a heck of a time, trying all modes, trying metering off of center point, set exposure and recompose scene, but still having issues

[cb7]

Im having the same under exposure issues with my 5d mk3. Have you or anyone else found any solid answers yet?

[ricardo474]

I also just purchased a 5D lll and put away the 7D, which did a terrible focus job. But...it did a far better job of metering. I have heard alot of technical talk about the cameras etc..but the FACT remains that the 5D lll does not produce the metering desired. I have used the 24-70 2.8 L  70-2002.8 LIS and the 400 L and the camera normally produces underexposed results. Don't start with all this comparison stuff...It underexposes a significant amount of the time. If Canon has an answer let them give a solution, but don't argue with the people who have used other digitals for thousands of frames and talk about comparing to other cameras. So... yes I join hands with those that are experiencing this underexposure problem with the 5D lll and await an answer from CANON.....

[Waddizzle]

Don't wait for an answer from Canon on these forums.  Contact them directly at Customer Support, 1-800-OK-CANON.

I think the metering mode, and the subject matter may have a lot to do with it.  I have frequently taken entire series of shots, only to later discover that, according to the histogram, were underexposed by nearly 1 EV on average. 

It doesn't seem to matter which camera I use, either.  On one occasion when I was trying to take a shot with a lot of dynamic range, I mounted the camera on a tripod and took different shots with different exposure modes.  I suggest that you try it before you call Canon.  Try taking the same shot, but with different exposure modes.  I tried them all, and discovered that they do what they say that they do.

[RobertTheFat]

---

@ricardo474 wrote:

I also just purchased a 5D lll and put away the 7D, which did a terrible focus job. But...it did a far better job of metering. I have heard alot of technical talk about the cameras etc..but the FACT remains that the 5D lll does not produce the metering desired. I have used the 24-70 2.8 L  70-2002.8 LIS and the 400 L and the camera normally produces underexposed results. Don't start with all this comparison stuff...It underexposes a significant amount of the time. If Canon has an answer let them give a solution, but don't argue with the people who have used other digitals for thousands of frames and talk about comparing to other cameras. So... yes I join hands with those that are experiencing this underexposure problem with the 5D lll and await an answer from CANON.....

---

The convenient thing about that attitude is that it's a lot less work than actually trying to analyze what, if anything, you're doing wrong. I have a 5D3 and two 7D's, and I've probably taken as many thousands of frames as you have. And my last camera that consistently and unexplainably underexposed was a Rebel XTi. (My wife also had one, and it didn't; so there may actually have ben a hardware problem.) But by all means have it your way. We'll stop bothering you with "all this comparison stuff", and you can feel free to wait as long as necessary for Canon to fix your problem.

[ebiggs1]

"...  and await an answer from CANON....."

Since this is a public forum, I believe you will be waiting for Canon to answer for a very long time.  If you feel the solutions offered here are worthless, do call Canon Support.  They will respond quickly and probably much better than any of us.

[TCampbell]

I don't believe there are any endemic issues with metering on the 5D III.  The 5D III, incidentally, actually has the SAME dual-layer Canon iFCL (intelligent Focus Color Luminance) metering system as the 7D.

I cannot say for certain whether or not there is an issue with your specific Canon 5D III, but I can say for certain that my Canon 5D III is bang-on accurate (and I know this... I've carefully tested it and built a callibration profile using software to check it.)

But in fairness, I must also point out that the difference between "incident" metering and "reflected" metering as well as things such as metering modes (which zones it uses and how it uses those zones), and the use (even if unintentional use) of exposure compensation features in the camera (and depending on the camera mode... selecting the wrong focusing screen type.  That doesn't apply to the 5D III since it doesn't have swappable focusing screens) can all factor in to someone assuming there's a problem ... even if the problem is user knowledge (the camera was only doing what it was told to do.)

I have EVEN seen exposure issues blamed on the camera when really the issue was a dim LCD screen (always use the histogram to check the exposure and know how a histogram works.)

I have a Sekonic L-758DR hand-held light meter.  It's the flagship meter by a company who pretty much only makes light meters -- that's their entire business.  So their meters tend to be pretty good (and they're a top-rated name in light meters.)  I normally advise caution when comparing the meter reading of an "incident" meter to the reading of a camera's built-in "reflected" meter because of the differences in how the metering systems work.  But in this case, the Sekonic L-758 happens to also include a built-in reflected meter that includes 1º spot metering (an extremely tight angle of view -- much tighter than the camera's spot metering mode.)  

I've tested the 5D III against the Sekonic and find that when both are in spot-metering mode and I carefully select the target, they are bang-on accurate in that they both precisely agree on the meter reading.  But note that as I sweep-around the subject area with the spot, the meter reading will change (and you don't have to move far to get it ot change).   That means you do need to be careful to ensure you really are metering the same subject with both meters.

If you've dialed in exposure compensation (even if by accident) that will show up in your EXIF data (you should see a row for "exposure compensation" and it should read "0".  If it reads any other value (on the negative side) then the underexposure is the photographer's fault (they told the camera to underexpose the shot... it was doing as it was told.)

If the photographer is metering a largely "white" scene (or spot metering on a "white" -- or near-white part of a scene) then the underexposure is the photographer's fault.

I obviosly cannot know if your camera is working problem or if the issues are caused by your own mistakes... I can only offer mistakes that are commonly made so that you can check to make sure "you" aren't the cause of the underexposure.

You can use a gray card to check your meter.  Just be careful because nearly all gray cards are designed to provide 18% gray and most cameras are callibrated to be closer to 12% gray.  That's a 1/2 stop difference.

This means that if you point your camera at the gray card and move in close to get the gray card to "fill the frame" (that's usually not necessary as long as the camera's metering coverage area is all on the surface of the gray card), take that shot, and then inspect the histogram.  You SHOULD notice the peak on the histogram is left of the center line (if it were a 12% gray card it would be "on" the center line of the histogram).  But since the difference between an 18% gray and 12% gray card is only about 1/2 stop, if you deliberately over-expose by 1/2 stop above whatever the meter suggests (or dial in +1/2 stop worth of exposure compensation) then the histogram peak sould be "on" the centerline of the histogram. Of course the camera typically allows exposure compensation in 1/3rd stop increments, not 1/2 stop increments, but if I do this with my camera it does nail the exposure on my 18% gray card (and it also perfectly agrees with my Sekonic L-758 meter.)

Notice I mentioned comparing the camera to a known-accurate meter (an independent source to validate my results) and also included the use of a gray-card (although it's an 18% gray-card and not a 12% gray-card but you can compensate for that if you understand how gray-cards work)?  This is because this provides a reasonable and, more importantly, a specifically measurable basis upon which to form a conclusion.

I should re-assert something I mention in threads from time to time:  If you want to TEST your camera's accuracy (in this case for metering accuracy), then do it scientifically.  YOU MUST CONTROL THE CIRCUMSTANCES OF THE TEST!!!  I cannot stress this enough.  What you absolutely MUST NOT do is take a photo of a typical subject and use that to conclude there is a problem with the "camera" (anytime someone does this, I immediately recognize that the real problem is the 12" behind the camera.)  You can use your normal photographic results to "suspect" a problem with the camera, but you must be scientifically measureable and accurate methods to "validate" that suspicion.