Nikon Light Metering

This page was deemed necessary as there seems to be recurring questions on the 'net about how the Nikon ambient light and flash TTL metering is performed. The information in the Nikon brochures gives no exact details but enough to allow you to understand what the system can do. This is general information applying to F70/90/100/5 and their Nxx USA namesakes.

Firstly, there seems to be the die-hard brigade that thinks that they can do better with their hand held meters, maybe so. But of course you have to add many years of experience to the use of the hand held ambient or reflected light meter to get the best results.

The SLR camera metering is always going to be more accurate for reflected light because it is measuring the scene that will be the picture, whereas the hand held meter is a bit vague about what it is seeing. The advantage of the camera is that it can be switched (the upper end Nikons) between matrix mode, centre-weighted mode and spot mode. All you have to do is apply a bit of common-sense to the situation you are seeing and decide what mode to use and how to use it.

Now, why you may choose to switch between metering modes.....

If you are presented with an everyday scene then the matrix mode is fine because it works out the best result based on contrast between the segments of the meter and by guessing where the subject is (that assumes use of the D feature lenses). The Nikon system bases its exposure upon the result of test pictures taken under all sorts of conditions. How exactly it does this is not explained clearly anywhere but rest assured there's a very busy computer in there. The F100 brochure indicates that 30,000 photos were taken to make the database the computer uses to choose the best exposure.

If the scene is of a sunset and you wish the bright setting sun and clouds to be exposed correctly and have not much detail or interest in the foreground and edge detail, then switching to centre-weighted makes more sense. This mode is also useful for situations where you may be shooting through a doorway and wish the scene to be perfectly exposed and the edge of the doorway not to be.

If the scene is a bit trickier than that, like having someone's face surrounded by very dark or very bright background then a switch to spot mode is called for to meter exactly and only on the face. In recomposing the scene the reading may change so the use of the exposure lock key is called for, or even going to metered manual mode and setting the controls to suit the spot metering on the face and then no matter where you point the camera the exposure setting is preserved.

Once in an effort to compare what the various modes did with a difficult scene, I took shots of boats lying on dark shiny wet mud looking directly into the setting sun using all three metering modes on my N8008s (F801s). All prints and negatives ended up looking the same. Hmmm. Maybe the situation wasn't difficult enough. I had the sun off centre and used the spot meter on the mud and the other two as it was framed.

Later.... As I have read lots of Nikon brochures and camera manuals I have gradually come to the realisation that the matrix mode is not the magic metering mode that can be used for everything. Information states that if using neutral density filters, or grad filters, or strong coloured filters for B&W - do not use matrix mode, use centre-weighted or spot instead. It appears that matrix mode is very clever and  is designed to "know" the lens maximum aperture, the scene brightness, the contrast between segments etc etc and then make an exposure decision. If you alter the scene brightness or contrast with filters, then the matrix logic may not make the best exposure decision. I guess that also holds for polarising filters. I suppose the main reason would be that the matrix metering decisions are based on a set of normal photographs, so if you do abnormal things with filters, then you are working outside it's range of knowledge of the prevailing conditions. Anyway, with matters like this you try a few comparison shots between matrix and centre-weighted or spot when you think you may have set up an abnormal condition and then compare the results. The block below was written earlier and still applies but the thinking may be modified a little by this paragraph.

Some criticisms have been raised on the odd exposures the matrix mode sometimes yields but the camera is not magic, there are some conditions it can't handle. That is mainly due to the fact the exposure latitude of the film is limited and the camera can only make a best guess in difficult conditions and that may not suit the film you are using. It also tries to guess where the subject is in the frame and that logic may be fooled at times as you may have other ideas. Slide film has the lesser latitude so may show the most "failures". You have to be able to recognise when the system may fail and make appropriate adjustments.

The F70 camera handbook says in the section on exposure compensation...

Nikon's 3D Matrix Metering employs methods of exposure calculation that automatically apply exposure compensation, depending upon scene brightness and contrast and distance information. As a result, your subject, whether it is centred in the viewfinder or not, is given corrected exposure in most lighting situations. We do not recommend using any manually or automatically applied exposure compensation when using Matrix Metering. If you identify an extreme condition under which Matrix may have some difficulty, such as severely backlit or one with extremes of contrast, we recommend using your camera's other built-in meters. Centre-Weighted or Spot. Ultimately, only you know what the subject or part of it requires in terms of exposure measurement. That's why the F70 camera incorporates three meters plus a variety of exposure compensation systems. The photographer's creativity is always the deciding and controlling factor.

I guess that's Nikon-speak for saying "if it goes wrong, it's your fault"!

Confusion exists also about the two major metering systems in the body, there's the ambient light eight segment meter (five in older models and more complex in F5) that lives on top of the prism near the viewfinder. It measures the scene the lens sees when the mirror is down and can be switched between matrix, centre-weighted and spot modes.
Be very aware that light coming back in through the eyepiece also gets to this meter and can cause underexposure. This usually happens when you are taking tripod shots and your eye is not blocking the light from getting in. It can be avoided by blocking the eyepiece with the little plastic clip provided or with the eyepiece shutter, it depends on the model. The effect can also be avoided by using metered manual mode, set the aperture/shutter while the eye is at the eyepiece, then when the extra light gets in when you move your head away it doesn't matter as the settings are already "frozen".
The other meter is the five segment meter (single segment in older Nikons) that measures the light from the flash. This meter is situated in the base of the mirror box and is only illuminated when the mirror is up. It either measures the flash light reflected back off the film surface or measures the Monitor Pre-flashes reflected back off the front side of the shutter before it moves.
The older single segment TTL flash control gave a sort of bottom centre-weighted measurement, but the newer five segment meter gives a much better evaluation of the flash light for the whole frame. Plus when the clever logic is added if using the newer flashes and D type lenses some very clever things happen and allow you to take those impossible shots easily.

The main pain with the early days manual flashes was that careful calculations had to be made based on distance, flash strength and film speed to arrive at an aperture to use. Easy to make mistakes.

Self-auto flashes removed the calculations but they only received the bounced light back off the scene in front of them and didn't "see" the same scene as the camera lens, they shut off the flash output when the sensor on the front of the flash had seen enough light for the aperture you had chosen on the camera and on the flash unit. Getting better, reduced the mistakes.

Next TTL (Through The Lens) flash metering measured the flash light bounced off the film surface and shut down the flash when enough was seen. This had the advantage that the aperture settings didn't matter as the meter saw the light as it was received at the film surface via the closed down aperture, no matter what setting it was. Plus, of course, the flash light from the actual scene the lens was seeing was being measured. Much more accurate but still prone to mistakes because the meter doesn't know what's important in the scene.

The Nikon got a bit smarter and brought in a five segment TTL flash meter that when used in conjunction with the clever flashes and D type lenses has a good guess at what's important in the scene and only measures the flash light from the significant segment(s). This ensures an off-centre subject is exposed properly, or improves the situation if the subject is in front of something very dark or very reflective. Much better chance of getting a good flash shot every time.

Probably here I should just quote the Nikon N90s brochure of 1995 as their description is as much information as you'll learn about the way the system works. First the ordinary light metering and then the flash metering.
The following is copyright property of Nikon 1995 from brochure (9503/B) 8CE41701. I reproduce this without permission but in the hope it provides more understanding of how the metering works. This is the best description I've found amongst all the Nikon brochures.
 

Starting on page 28. Quoted exactly as it appears.... Guy 3D Matrix Metering With 3D matrix metering, the camera's computer analyses the following data to determine exposure: #1 - Brightness data from each of eight segments (configured in various combinations) #2 - Contrast between segments #3 - Distance Information from the D-type Nikkor lens in use #4 - Focus status data from the camera's AF system (during composition and after recomposing with AF-lock) On the basis of these four types of data, the N90s's computer selects the most suitable algorithm to use for the exposure calculations from many stored in the camera's memory. (Note: an algorithm is a set of procedures for solving a mathematical problem.) First, the computer calculates the exposure value for the scene in two ways: as an eight-segment pattern and as a five-segment pattern (center segments are regarded as one segment). Then both exposure values are combined to offer the best possible exposure according to  data #1, #2, #3 and #4. Distance information (#3) is used to balance the readings, for example, if the subject is close to the camera, the computer biases its calculations in favor of the five-segment reading: because when shooting a subject nearby, especially in extreme close-up photography, too much segmentation may mislead the meter because of the high magnification ratio. 8-segment Matrix Metering also works with a non-D-type AF or AI-P Nikkor lens, but without using Data #3. Finally, focus data (#4) determines if the subject is in the center of the frame or off-center. For example, if the amount of defocus detected is not great, the computer judges that the main subject is in the center and the information from the center-segments is appropriately weighted. If the defocus amount is large (out of focus), the computer judges that the main subject is not in the center (as when focus lock is used for recomposing off-center) and the information from the peripheral segments is weighted (based on their relative brightness and contrast). When Focus Lock is used, if the meter judges the scene to be strongly backlit or the main subject to be strongly illuminated against a dark background, data detected prior to recomposition is also considered for final evaluation. The foregoing also applies to manual focussing with Electronic Rangefinder. In addition, fuzzy-logic algorithms are employed for smooth data processing and to prevent sudden changes of exposure in continuous shooting. (for the next section, read "SB-26 and SB-28" for where it says "SB-26", also the F70/N70 and its pop-up flash use the same logic........Guy) 3D Multi-Sensor Balanced Fill-Flash With the SB-26 AF Speedlight and D-type AF Nikkor lenses, the N90s uses the five-segment TTL Multi Sensor to its maximum potential - the ultimate in balanced fill-flash control. This is how it works: 1) The D-type Nikkor lens sends subject-to-camera distance information to the N90s. 2) The SB-26 fires a series of weak flashes, just after the mirror goes up but before the shutter moves, as a Monitor Pre-flash for the TTL Multi Sensor. 3) The TTL Multi Sensor meters the light reflected from the gray shutter curtain to each of the five segments. 4) The camera's computer compares the amount of light actually metered by each segment of TTL Multi Sensor with the theoretical light amount calculated from the Distance Information from the lens, the guide number of the Monitor Pre-flash and the lens aperture in use. 5) The camera's computer analyses and decides: i) which segment of the TTL Multi Sensor to use for TTL flash control according to a relative reading of the reflected light amount, and ii) what amount of flash is necessary to balance with ambient light according to the exposure meter used, including 3D Matrix, Centre-Weighted and Spot. 6) The shutter opens, the main flash fires and the camera's computer controls flash output based on data from 5). This function ensures correct exposure even in difficult situations which ordinary methods cannot accommodate, that is, scenes that include a very reflective object such as a mirror, white wall, or something very close to the camera other than the main subject. It also accounts for the sun being part of the scene, the background being very distant and other similar conditions. Multi-Sensor Balanced Fill-Flash Type 1 With the SB-26 and a no-D-type AF or an AI-P Nikkor lens: Without distance information, computation is less refined than when using a D-type lens. However, results are much superior to single TTL flash sensor systems. Type 2 Without Monitor Pre-flash: The N90s's Matrix sensor provides the information on too-bright or too-dark areas included in the scene; the TTL Multi Sensor then controls the main flash output. This method is superior to single TTL sensor systems, but not as effective as when the Monitor Pre-flash is used.

Well, there you have it. The description above also applies to the F70 and the F100. The F100 is a little different, details below. The F5 has cleverer ambient light metering introducing colour sensing so the story may differ there a bit.

The central small circle is the spot metering element, the bits that make the small oval shape round the centre spot cover the AF area. The larger circle with the chopped top has to be the centre weighted area, and the four outer areas make up the rest. When the meter goes to five segment mode it must be combining everything in the centre weighted central area.  This meter lives above the viewfinder prism and is only in action when the mirror is down.

Again the same basic arrangement but now the five squares are placed in the pattern to line up with the five AF spots, so spot metering will actually happen on the spot that is focussed on. This meter lives above the viewfinder prism and is only in action when the mirror is down.

Sorry, but I know little about the F60/N60, here's the meter layout from the brochure. It's called a six segment meter and they only provide full matrix or centre weighted. The spot in the middle does not mean there is a spot mode, maybe that's reserved for a future model. There is no mention at all of the method of flash TTL metering so I have to assume that it's the single sensor type as per my N8008s/F801s (which works very well, anyway).

The components 3 & 4 live in the bottom of the mirror box and are only in action when the mirror is up. If the camera is using the monitor pre-flashes then the item 1 is really the shutter surface before it opens, then after the shutter opens the metering is taken off the film surface.

Remember that the metering systems will do the best they can even if you have an old Nikkor or a "foreign" lens on the camera or the lens isolated by something like a bellows unit. When the camera can't get information from the lens about its maximum aperture then the metering system won't provide Matrix Metering, only Centre-Weighted or Spot. The flash TTL five segment meter still works and does the best it can with measuring and comparing the flash light reflected on to the five segment flash TTL meter.

Even though the flash system is very clever people still complain about washed out subjects in their prints. This is 99% sure to be a photo lab printing problem where the printing machine tries to average out the scene so, particularly when there are large areas of black night sky behind, the subject ends up overexposed in the printing process and the sky comes out dark grey instead of black. Examine the negatives with a magnifier and determine if the subject is really exposed correctly, then request the lab to re-print with more density.

The Nikon flash metering system is reputed to be the smartest on the market and makes less mistakes than other brands may. Next project will be to dig up similar information on Canon, Minolta and Pentax to see how they do things.