What went wrong with the metering

<p>jeremy r- in my view, the problem is not metering, no matter what kind. it is the scene, in the first scene you have a pure white wedding dress along with a distant very dark background consisting of people and church. in the second pic you again have a white wedding dress along with the guys dressed in the black tuxes. in either case you are very probably exceeeding the dr range of the sensor. so no matter what metereing and what settings you shoot with either the white or the darks will not be correctly exposed.<br>

the only solution is to shoot for the whites and somehow get more light into the general scene to up the light levels for the dark areas. the result will be a dr range that the sensor can take. possibly a flash firing into a reflector will be the answer, this would up the general light level. the problem might be the church shot, if flash is not permitted. in the first shot any upping of the exposure would very probably begin to blow the white highligjhts in the wedding dress. so just adding more flash power would not be the answer. the general illumination of the scene would have to be increased. as others have stated, see if you can use a slower shutter speed, that would let more ambient light in. but too slow and you would be geting subject movement. if me, i woukd try a 1/60sec, and let the flash determine the amount of the flash duration. this shoud let more light in for the background but if it will be enough i do not know. i suspect that in this scene the background will still be dimmish.</p>

<p>There are a lot of good answers aready, so I will only reapeat my expereinces. Obviously you have a base knowledge. I would use Matrix metering or Manual. If you use spot all that white at a wedding will give you huge swings in your Histo.<br>

But I shoot in AP a lot, and adjust the Exp comp up or down depending on what is needed. That is easily done with a few flicks of the fingers and doesnt require you to take your camera off the subject.<br>

Not sure what mode you are using for your flash either, you can get more abient light but slow syncing too. Maybe give that a shot.<br>

One last thing, I think I saw someone else above say it, but your white balance is off in that second picture. Those whites should be white, not pink.<br>

my two cents...</p>

<p>To correct something Peter Kervarec said above: Most of the dark tones are in the background and not as relevant to the final image as the white dress of the bride. Your "typical problem" with Nikon TTL is that it avoids blowing out highlights. Blown highlights are harder to recover in post-processing than slight underexposure.<br /> <br /> When used in TTL/BL mode with the flash head pointing directly at the subject, the Nikon flash system uses focus distance information it receives from compatible lenses. If you miss and get the focus on something far in the background, the system thinks it needs to put out more power and you end up with an out of focus subject with blown-out highlights. The camera must also be set to matrix metering for this to work. This distance capability is what Nikon means when they use the term "3D Matrix Metering". (This also means that if you miss the focus, you can end up with a badly overexposed foreground.)<br /> <br /> I don't know how well all of this works with non-Nikon lenses and camera bodies. Jeremy is using a Fuji S5 which is based on a Nikon D200. The S5 uses the same focusing system but I don't know if Fuji changed the way metering works. It's pretty easy to duplicate the test below with find out.<br /> <br /> I've never done a test but when bouncing the flash the focus distance probably isn't be used for the flash output calculations as there's no way for the camera to know how far the flash is traveling or the reflectivity of the bounce surface.<br /> <br /> The SB-600 is capable. I'd get an f/2.8 zoom before another flash. The larger aperture makes it easier to fill in the dark background more. And the SB-800 isn't available new anymore anyway and doesn't have so much more output than the SB-600 that it will make a huge difference. The biggest difference between the two is that the SB-800 can be used in Commander mode. And Jeremy will have to pony up $450 for an SB-900.<br /> <br /> The images below were all shot at 1/30 @ f/2, ISO 200. Note that this is manual exposure on the camera body only. The flash is going to do whatever it darn well pleases, but the results work for me. At an actual event indoors, I'll set the camera to give the desired background exposure and let the flash decide the foreground exposure by running it in TTL/BL.</p><div></div>

<p>A little post-processing fixes the altar shot very nicely. The D300/D700/D3 series can do this when shooting in-camera jpegs. (See Active D-Lighting on the shooting menu.)</p>

<p>I prefer to shoot stuff like this in RAW and as far as I can tell, the D-Lighting setting does not affect the RAW capture. Plenty of discussion about this in the Nikon forums.</p>

<p>The Fuji S5 has a "funny" sensor that's supposed to capture a high dynamic range with two sets of pixels, one tuned for shadows and another for highlights, but the appropriate file recording mode needs to be selected.</p>

<div></div>

<p>jeremy r- my try at fixing the existing image.<br />used pse7 then auot levels, auto contrast, shadows/highlights(shadow only), remove color cast, noise ninja, focus magic. time about 1 minute. this would work better if the file size was larger.</p><div></div>

<p>I'll take a different angle at this - it's not clear to me that this isn't simply a case of lack of flash power. The 1/160 shutter speed would largely explain why the background is dark as well, but by the looks of it, its possible that you're just too far away for the flash to reach at that f/stop.</p>

<p>At ISO 400, the SB-600 has a guide number of 190 ft with the flash head at a 35mm zoom. With the aperture setting of f/4.2, that's enough reach to get decent exposure past 40 feet. (The SB-800 only gives an extra 10 feet , by the way.)</p>

<p>The issue is the SB-600 cutting off because of the large expanse of white, not that it doesn't have enough power for this shot.</p>

<p>Gary's post processing maintains the detail altar cloth at the left edge of the photo. But that comes at the expense of creating a gray bridal gown. Not a good trade-off. (And though Focus Magic and Noise Ninja are useful software packages, they aren't really pertinent for this particular discussion.)</p>

 

<p>tom l-there is a choice of what the pper wishes to put at WHITE-the gown or the alter cloth. i know that the alter cloth is WHITE, but i cannot say the same about the wedding dress. there are such things as off white dresses. the only person that knows is jeremy r.<br>

as for including in decription of my pp process i simply included the noise ninja and focus magic to give a complete list. so that others including jeremy r would know what was done during the pp.</p>

<p>OK, I'm still very, very confused about i-TTL and whether compensating the ambient exposure in an automated mode such as aperture or shutter priority also compensates (somehow) the flash exposure. I have heard or read it does. Also, whether an ambient metering pattern, such as spot metering, actually turns your flash into a spot flash meter (I kind of doubt it--don't see much in the way of metering differences in Tom's examples). I was reading a Tom Hogan article about how plus flash compensation, used with automated modes, can sometimes result in more underexposure (?)</p>

<p>In any case, the question posed by Jeremy is about what metering pattern (actually ambient metering pattern)--he calls it a 'metering system'--he should use to not have underexposed images. He knows that the white value has something to do with flash metering underexposing, but has identified the thing to correct the underexposure incorrectly--center weighted, spot and 3D matrix are ambient metering patterns, not systems.</p>

<p>Compounding the problem in answering the question is the fact that he is using an automated ambient metering mode (does this alter the compensation of the flash or not?). And, the flash has it's own metering patterns (does this change according to camera mode?) and 'systems' (are there only two--regular TTL and TTL-BL?)</p>

<p>Also, anytime there is a lot of white in an image (for both ambient and flash metering) and you don't compensate the exposure, you will get underexposure, because the meter always wants to see everything as middle gray. This is regardless of brand--Canon, Nikon, Sony, etc., and whether ambient or flash metering.</p>

<p>And, as I mentioned, you also have inverse square law entering the arena. The left alter cloth in the first image and the table cloth in the second image are going to receive the metering 'attention' by the flash because they both are closer to the camera/flash than the subjects. This would be the case no matter what the color of the closer item was--the problem is just compounded (flash fall off) because the closer items are also white.</p>

<p>Based on the above, I would offer a simplified answer.</p>

<p>1) Use TTL if the flash is primary, TTL-BL if the flash is for fill only.</p>

<p>2) Plus compensate the flash (in either mode) if there is a lot of white or light colors in the image. Do some casual tests at home using subjects against light backgrounds and against light backgrounds. You should soon see some kind of pattern emerging. I use Canon, but I know that when I photograph something white or light colored, or what to get 'past' a closer object, I plus compensate my flash--+1 1/3-1 2/3 or 2 normally does it.</p>

<p>You should be plus compensating the ambient exposure too, if using an automated mode.</p>

<p>Now all you Nikon folks--can you clarify and add without getting too technical?</p>

<p>Unfortunately it is not that simple Nadine. The OP need to test his camera and flash in different modes to find out how they work. How's that for a non technical response? :-)</p>

<p>And I'll show a collage of images (jpegs) I shot just now for test purpose having the camera in Program, metering set to Matrix and flash set to TTL-BL. No flash or exposure compensation is changed between any of these images. As you can see they wont average out to middle grey.</p>

<p>More info on TTL / TTL-BL <a href="http://desmonddowns-ttlblflash.blogspot.com/2009/06/nikons-new-ttlbl-flash-and-cls-wireless.html">here</a> .</p><div></div>

<p>PS. First image on the left has a white wall to the left much closer to the camera, that's why it is blown out (and rightly so).</p>

<p>Pete--Yikes. I guess Jeremy also has to find out if the S5 works with the new standards, too.</p>

<p>Hi Nadine<br>

I think you have really answered your own question. Its well explained by you. To add a a rider to "TTL-BL if the flash is for fill only ", yes mainly but I use this setting as often as I can indoors when there is reasonable ambient light. What this setting tries to do is maintain the background ambience and balance the exposure with the subject. Sounds unrealistic but works very well.<br>

I second what <a href="../photodb/user?user_id=1807070">Tom Luongo</a> says about distance info. The flash definitely gives out more power if you misfocus onto a more distant subject.<br>

Digital technology is brilliant and is far away more complex than any film camera ever was. There are so many variables that can be applied to a situation if you wish.( Ala Photoshop )Having said that, if one is prepared to investigate and get it right, the results are nothing short of fantastic.</p><div></div>

<p>Nadine,</p>

<p>Here's my simplified answer (without getting too technical)</p>

 

<ol>

<li>Use iTTL. iTTL is being used when TTL-BL shows on the back of the flash.</li>

<li>Set the camera's exposure mode to matrix. iTTL isn't available otherwise.</li>

<li>Manually set the camera to get an appropriate ambient exposure. That is, so the background is light, but not too light.</li>

<li>Use direct flash.</li>

<li>Use a compatible lens that gives distance information to the flash. </li>

</ol>

<p>#4 above is key. If direct flash is being used and iTTL knows the distance to the subject, then iTTL can calculate what power to output.</p>

<p>Trying to render the scene as an 18% gray isn't really a concern. It's not too different than how a photographer varies the f-stop in a fully manual system. Usually expressed as a chart on the back of the flash, the equation is f-stop = GN / distance.</p>

<p>With iTTL, the flash can vary the guide number (i.e. flash power) based on the selected aperture. GN = f-stop x distance.</p>

<p>I redid my sample exposure to more clearly show this trend. All camera settings are the same from shot to shot. I even put the camera on a support this time. </p>

<div></div>

<p>Upon further reflection, iTTL is doing more than just using the distance information. The exposure difference between 2 ft and 4 ft is about what what I'd expect (1 stop). I don't really know there's such a big jump between 4 ft and 7 ft (about 2 stops) when I would have expected. And though there is a difference between 7 ft and 20 ft, it's less than if distance is the only factor.</p>

<p>I was shooting in a small room. So there's some contribution from flash bouncing off the mostly white walls and ceiling. iTTL can probably factor that in some way. iTTL also does something clever with the 1005 segment (or whatever) meter.</p>

<p>By the way, did the original poster ever say what mode the flash was in? I'd assume the flash was in TTL or iTTL and the camera was set to matrix metering.</p>

 

<p>I don't think that info was given. Still scratching my head, but going to read this again tomorrow.</p>

<p>Hi Tom<br>

Flash was in TTL BL mode with Matrix metering</p>

<p>My head is spinning! :-) Good info. Need more time to digest.</p>

<p>Nadine, not sure if anyone answered this part of your question re Nikon flash -- with Nikon, flash exposure compensation (FEC) and expsoure compensation work together, unlike the Canon system. e.g., you can set FEC on both the flash and the body. (I am assuming a D700, because it has a built-in flash.) So, if you set +1/3 on the flash and +1/3 on the body you get a +2/3 flash expsoure. (I am talking FEC.) Now, if you set exposure compensation to +1, the D700 will adjust the overall exposure +1 while maintaining the same ratio between flash and ambient. Canon doesn't work this way. Put another way, as you adjust exposure compensation up or down, the Nikon system will increase/decrease flash output accordingly to maintain the ratio of flash to ambient.</p>

<p>A neat thing with the Nikon system is that if you have the camera set to manual mode, then the exposure compensation button will affect flash compensation. Confused, yet? :-)</p>

<p>Hope this helps. Oh, I wouldn't use TTL BL indoors, usually.</p>

<p>Whew, lots of pretty complicated answers ... all of which may well be absolutely correct if I could wade through it all without getting a headache ... LOL!</p>

<p>A couple of questions that occurred to me about this issue. My experiences with Nikon flash don't jive with this result. Matrix metering with the flash set to iTTL could have done the job better than this ... not perfect, but better.</p>

<p>The difference is that I never would use 160th of a second in a place this dark when using a 35mm focal length set to f/4.3. I haven't used one single camera yet that wouldn't underexpose in a similar composition situation with those settings unless camera/flash compensations were fiddled with. I'd use Manual on the camera, shutter speed set to 1/50th since there is little movement to contend with, and let the flash do it's job ... with possibly needing a small bit of compensation.</p>

<p>In this case using the OP's settings, if you comp'ed the flash to the plus side enough it'd probably show cast shadows more than now, because of the shutter speed/aperture/ISO combination would force the flash to be the primary source of light even more than it is now. Which is why so many people use "Dragging the Shutter" techniques in situations like this.</p>

<p>Don't sweat subject blur to much. Remember ... the more flash becomes the prime source of light, shutter speed becomes less important as the flash duration is much shorter (quicker) than the camera's shutter. To grasp this concept, think of how they freeze bullets in mid-flight. It sure isn't using any shutter speed known to man ... it is extremely short flash durations.</p>

<p>Another question is how were these processed? If your defaults and profiles aren't set up right for your camera, you can get a ton of underexposed (or sometimes overexposed results) as the post program auto balances each image ... thus adding a layer of unnecessary program effort to make everything middle grey ... exactly the way a meter will. I see this as often as I don't. The wrong ICC profile will also affect the image.</p>

<p>So will poor white balance, which the second example shows. This can throw off the actual exposure, and is very apparent when in a post program where if you white balance a very yellow or pink cast image it'll immediately go darker. Setting a manual WB does wonders for exposures and is usually done once in any given lighting scenario.</p>

<p>Now I have an even bigger headache : -)</p>

<p>-Marc</p>

<p> </p>

<p>Oh Great! Marc if you have a headache, imagine what mine feels like. :-) I gotta get another pot of coffee on and then re-read the thread.</p>

<p>Yeah--my headache started when I began to read all the info about how Nikon flash metering works. Apparently it has changed along the way too. And I agree, all the tech stuff is probably not helping Jeremy figure out what is essentially something every brand of flash metering (and ambient metering) does with light or white subject values (and also dark or black subject values). This is why I wanted to simplify the answer, above, and leave off stuff about dragging the shutter, bouncing, or trying to get the background in. But I'm still interested in how the flash metering actually works, so I'm still going to re-read everything.</p>

<p>Nadine, thanks for trying to steer the valued contributors in the correct direction.<br>

I do want to do some tests myself to try and wrap my head around the myriad of scenarios posted here.<br>

I am getting the majority of what is being said - I obviously need to put a lot of the theory into practise.</p>

<p>Automated exposure modes are always subject to error. White dresses and dark suits perplex even the finest reflective light meters. The OP might have corrected the first photo by boosting exposure compensation (not flash compensation) by 1 to 1.5 stops and reshooting, but if the next photo didn't have a white dress in the middle of it (perhaps a dark tuxedo instead) it would have been grossly exposed at these settings.</p>

<p>The OP could have checked his histogram, adjusted the exposure in one of a number of ways and reshot, but the moment would have been lost.</p>

<p>He could have set a manual exposure based upon a meter reading (incident or gray card) and a manual flash output value, but he would have to adjust the flash power anytime he move closer to or further from the subjects.</p>

<p>Probably the most reliable approach is to start with a base exposure that shows detail in the room (adjust if you want to make it darker or very dark). Then use iTTL flash on top of that exposure, being careful (a) not to blow out the white dress and (b) not to let the dress turn gray. As you move closer to your subjects, turn flash compentation down. If you're farther away from the subjects or if you are bouncing, turn the compensation up as needed.</p>

<p>Don't forget to check the histogram and the blinking highlights displays. Any blinking must be compensated for immediately unless you're willing to lose most or all of the detail in those areas of the photo.</p>

 

<blockquote>

<p>Don't forget to check the histogram and the blinking highlights displays. Any blinking must be compensated for immediately unless you're willing to lose most or all of the detail in those areas of the photo.</p>

</blockquote>

<p>For JPG. Not necessarily for NEF.</p>

<p>Back in the "old days" when I was shooting a wedding every weekend with film cameras (hasselblad) I never used a meter, and I suspect that if I were shooting them today with Digital cameras I wouldn't use one either. Far far too many opportunities for those meters to make a mistake.<br />Anytime I was shooting inside I knew it was time for 400 ISO film and f5.6 at as slow a speed as I could handhold, which was usually 1/8th of a second. Maybe sometimes a 1/4 of a second if it was really dark and I still needed to handhold. If it was really really dark it was time for the tripod and 1/2 second. And then of course use the flash for the subject.<br />That was called "dragging the shutter" back then. I don't know what you would call it today. But it somewhat took care of that inverse square law problem by allowing backgrounds to have detail even though the flash didn't get out there into those backgrounds to lighten them up.<br />Generally speaking, you basically have 5 or 6 exposures during a wedding. (bright sunlight, overcast, shade, indoors with lots of light, indoors with medium light, and indoors with very little light). If you memorize them you can shoot on manual and you won't have to worry about meters.</p>