[BruceBanner]

I understand what the metering system does, but I always just stay in Matrix and deliberately under or over expose using the EV compensation (in Av mode that is). It occurred to me the other day that even when you have some dynamic lighting in your frame and say you want to protect that bright part from having it's highlights blown out (and are content with the rest being plunged into dark shadows), you could set to spot and take the shot, or just stay in matrix and keep underexposing till the bright part is protected. I tried doing this and the results looked the same, no difference with that particular example. In fact... spot was worse because you need to have that bright part of the image central in the frame, for those instances where the bright part in a dynamic scene is off centre then spot won't help you, matrix with ev comp will.

So I've always just stayed in Matrix and adjusted my exposure via EV compensation. My question is, are both routes pretty much the same end result, or is there something quite critically different I'm missing?

• With metering choices, the exposure is kinda more 'automatic', however you do have a little more work to do in terms of changing the metering method, framing the shot (relevant to the method, such as centre for spot), possibly locking exposure and then recomposing for the shot.
• EV Compensation requires no reframing or AE-L per se, but you might have to chimp more till you get the right exposure for the part of the image you want to protect (conversely happy to ruin/let go of the other dynamic range etc).

I know, sounds silly to compare the two, but really.. I hardly ever toggle out of matrix and just wondered if there was something else missing than I can't get from my constant EV Compensation use. I mean I do occasionally toggle to spot for those kinda events that it's really important (public speaking, concerts) where I just want the singers/speakers head exposed well and to hell with the surroundings etc, but yeh...


TIA!

Bruce

[beachgardener]

Centre weighted for me, just like it better for quick non thinking Av shots.

However, for shots that matter, and potential difficult lighting I like to meter areas of the scene manually (usually incident readings) and choose an exposure that matches close to what I am wanting as a result in an image.

[pschlute]

or just stay in matrix and keep underexposing till the bright part is protected

This is a sensible option, reviewing your exposure (using histogram), and adjusting your exposure parameters. This method works with any metering mode, and importantly no metering mode.

In fact... spot was worse because you need to have that bright part of the image central in the frame, for those instances where the bright part in a dynamic scene is off centre then spot won't help you, matrix with ev comp will.

Spot metering can help you by using AE-l or manual exposure mode

just wondered if there was something else missing than I can't get from my constant EV Compensation use

Spot metering is invaluable when you want to know the different ev values in different parts of your scene. If using ND graduated filters for example, the chimping method would work but would be very cumbersome to use. The spot meter allows you to choose the strength of the filter at the outset.

Spot metering was invented in the film era when instant review was not possible. It helped photographers get the exposure right for a difficult scene. maybe I am old school but I prefer to continue this practice.

[Dartmoor Dave]

Spot metering assumes that the area you are pointing it at has a tonal value similar to an 18% grey card and gives you the correct exposure for a grey card. If you are actually pointing at something lighter or darker than 18% grey, you'll need to adjust accordingly. The most sophisticated use of spot metering is the zone system, but that's too in depth to go into in detail here.

Centre weighted metering assumes that the tonal values around the centre and bottom of the frame will tend to average out to 18% grey, so it's only accurate if that's the case. Once again, if the actual scene averages out to lighter or darker than 18% grey you need to compensate -- snow being the classic case.

Matrix metering divides the scene up into zones, meters each zone separately, compares the results against a database of different patterns characteristic of different types of scenes, and tries to work out an exposure from that.

You've mentioned the specific case of metering for brightly lit parts of the frame, so one handy method is to do some tests to work out how many stops of overexposure compared to the meter reading your sensor can handle while still retaining detail. Brightly lit parts of clouds are a good test subject, so find a nice cloud and choose the brightest part of it that you still want to capture some detail in. Now spot meter that part of the cloud, and the spot meter will underexpose it by probably two or three stops because it assumes it's seeing 18% grey. Now increase the exposure by half stops or third stops, whichever you use, until you get to the last point where there is still detail in the cloud.

Let's say it turns out that your sensor can handle overexposing by three stops above the spot meter reading while still retaining detail -- now all you need to do is choose the brightest part of any scene that you want to keep detail in, spot meter that part of the scene (which will actually give you an underexposure reading because it's assuming 18% grey), then overexpose that spot meter reading by whatever amount your previous testing with the clouds decided on. You'll have detail in the brightest part of the scene that you want detail in, and all the other tonal values will be left to fall into place around that.

---

Last edited by Dartmoor Dave; 04-03-2019 at 02:50 AM. Reason: Clarification

[photolady95]

More about technique than General so moved to correct forum, Bruce.

[TonyW]

....

I know, sounds silly to compare the two, but really.. I hardly ever toggle out of matrix and just wondered if there was something else missing than I can't get from my constant EV Compensation use. I mean I do occasionally toggle to spot for those kinda events that it's really important (public speaking, concerts) where I just want the singers/speakers head exposed well and to hell with the surroundings etc, but yeh...


TIA!

Bruce

My experience of matrix both Pentax and Nikon is that generally you will get a perfectly acceptable result under a fairly wide range of conditions. Sometimes we just have to accept that the exposure may be somewhat less than optimal, but with modern sensors the issues of underexposure and noise is nowhere near as bad as it was with some older cameras. Clipping of IMPORTANT highlights of course is another matter

Applying EV compensation to matrix on the fly is a good way to go, as is chimping (although you will be likely underexposing if you are trying to judge by histogram or look on camera LCD) and reducing exposure until highlights protected. The problem may be that doing this you miss the 'decisive moment' as it occurs between shots.

Metering a specific ROI and applying simple compensation can often get us closer to the optimal exposure. One method is to spot meter the ROI, lets say it happens to be the brightest highlight that we wish to retain detail, then with knowledge of our metering system we can apply the required exposure compensation (either pre set EV shift or quick mental calc.).

The first problem I suspect is that we need to make sure that the ROI fills completely the spot meter area in camera. Your camera may not actually display an equivalent of the spot area and you may just have a vague notion of the area occupying a 2-5% ? circle based on your current focus point. The second issue that needs to be dealt with is the amount of compensation required for a given metering of a ROI. This will depend on how the cameras metering system has been calibrated by the manufacturer. I suspect that it is likely that Pentax follows the current ISO standards for exposure meter calibration and that this will mean that the meter is calibrated to about 12.5%.

IF 12.5% is the camera calibration point then by aiming our spot meter at an important highlight (one we require to hold detail - not specular!) and increasing exposure by +3EV we should have held detail in that particular highlight. Should the metering system calibration point be different then the exposure compensation must also change. For instance (and I think it unlikely) if the manufacturers calibration point is 18% then +3EV will clip highlights as the correct adjustment would need to be +2.47 EV (so practically +2.33 EV).

So as already mentioned to make the most of metering (spot particularly) then some practical testing of what your meters calibration point is will reap rewards.

A quick and dirty method applied by some/many 'old school' photographers was to meter a known value and increase the indicated exposure by the required amount to reach optimal exposure/sensor saturation. Grey card is obvious, but we do not always have one to hand or the time or inclination to use. But we do carry a known value with us at all times, namely our hand. By metering our palm (only) in the same orientation and light as our subject matter and increasing exposure by 1 stop will get us close to optimal exposure - check your own palm values first though.

As long as you do your research/testing on your camera metering system first and assuming you have the time and experience to implement your findings you may find that you have a liking for the precise nature of spot reading over pre calculated values and assumptions of matrix and centre weighted sytems. YMMV of course and as long as it catches what you require then one method over another may not be considered worhtwhile

[CarlJF]

I know, sounds silly to compare the two, but really.. I hardly ever toggle out of matrix and just wondered if there was something else missing than I can't get from my constant EV Compensation use. I mean I do occasionally toggle to spot for those kinda events that it's really important (public speaking, concerts) where I just want the singers/speakers head exposed well and to hell with the surroundings etc, but yeh...

Both have their uses, but yes, for most situations matrix and EV compensation will be the most practical and sensible things to do. As you mentioned, spot metering can be useful in some unusual lighting scenes. Backlighting where you don't care if the background is overexposed but you want the subject correctly exposed SOOC would be another example. Obviously, spot metering is also the way to go if you're using a grey card.

Spot metering was much more useful in the old film days, particularly with slides which didn't have much if any leeway for bad exposure. And the relatively rudimentary matrix metering used at the time was far to be as good as it is today.

[stevebrot]

Spot, Weighted and Matrix metering, is it the same as EV Compensation?

No...EV compensation is applied to the set exposure and is not related to metering. Proof? It works with arbitrary values set manually.

In regards to the metering itself, the difference between the three types is as follows:

• Spot only evaluates the area within a circle of limited circumference. Usually, this means the center of the frame unless set somewhere else.
• Center-weighted averaging evaluates the full frame, but weighted to the center, often with reduced sensitivity in the upper portion of the frame (landscape orientation). This type was dominant before the advent of evaluative matrix metering.
• Evaluative matrix takes measurements from multiple areas of the frame, evaluates those values and uses the results to provide a reading to the camera. This form is most prevalent in modern cameras.

All three types are capable of being fooled, with matrix being the most robust when confronted with difficult light. Center-weighted is quite capable, despite being very simple from a technical perspective. Spot metering is a specialized tool and should not be used unless the operator knows exactly what they are wanting to do with the reading (hint: requires either M-mode or EV comp).

All of this having been stated, the statement below remains a strong maxim:

"There is no such thing as an accurate or correct meter reading as it applies to exposure."

Yes, a meter may be properly calibrated and yes, it may be backed by sophisticated algorithms, but for most subjects the meter reading is at best a blunt hint to the photographer/camera as to how much light to admit to the film or sensor. All the meter does is give its best estimate of a middle value that MIGHT provide usable exposure for all elements in the frame.


Steve

[TonyW]

....
All of this having been stated, the statement below remains a strong maxim:

"There is no such thing as an accurate or correct meter reading as it applies to exposure."

Yes, a meter may be properly calibrated and yes, it may be backed by sophisticated algorithms, but for most subjects the meter reading is at best a blunt hint to the photographer/camera as to how much light to admit to the film or sensor. All the meter does is give its best estimate of a middle value that MIGHT provide usable exposure for all elements in the frame.


Steve

The maxim may only hold strong for evaluative metering where values may be 'integrated' to a grey value of around 12% and then some sophisticated algorithm adjust to account for scene brightness.


Spot metering is accurate (or should be if you know the exact area covered and EV range to sensor saturation) and a correct exposure may be applied to maximise SNR and maintain full DR, if that is required.

[Papa_Joe]

I prefer center weighted measure and EV compensation to matrix plus EV. The reason being, that center weighted is quite predictable, while you do not know what the processor thinks about the picture just taken in matrix mode. I can set the EV value in center weighted before taking the shot and will get usable results. In matrix mode I always do one shot to check the outcome and adjust EV afterwards if still necessary.

Else both methods are quite equal in terms of workflow. In critical lighting conditions you always have to check the histogram and set the EV if needed. In practice setting the camera to M mode and using an incident light meter might be sometimes easier.

In the end you have to choose the method you are used to and comfortable with as they will give the best results.
The best workflow is the one that works for you.

[gofour3]

Also note for center weighted metering is that cameras have different meter sensitivity patterns. (At least for my Pentax film cameras)

Example my KX has a different pattern than my K2 and my K1000/KM uses an all together different sensitivity pattern.

You need to know the pattern to make the necessary exposure compensation in some shooting conditions.

phil.

[stevebrot]

Also note for center weighted metering is that cameras have different meter sensitivity patterns. (At least for my Pentax film cameras)

Example my KX has a different pattern than my K2 and my K1000/KM uses an all together different sensitivity pattern.

You need to know the pattern to make the necessary exposure compensation in some shooting conditions.

phil.

Indeed! May I ask where you got meter pattern diagrams for your cameras?

Back in the ancient days, camera reviews would often include a diagram showing the the meter sensitivity pattern so that a buyer might be aware prior to purchase. Usually, this would not be an issue, but on some cameras, the weighting to avoid sky brightness would be abrupt enough to cause problems when doing a vertical composition. My late-1960s vintage Minolta SR-T 101 avoids this with its excellent CLC metering system which evaluates scene contrast as well as brightness using a primitive form of matrix metering (only two cells!). It works amazingly well.


Steve

[dms]

I prefer center weighted measure and EV compensation to matrix plus EV. The reason being, that center weighted is quite predictable, while you do not know what the processor thinks about the picture just taken in matrix mode. I can set the EV value in center weighted before taking the shot and will get usable results.

As said above, matrix metering is making an exposure adjustment by rules that probably give good results for typical scenes, but one does not know what the rules are it uses, and you should be able to meter the scene and then adjust based on what you conclude is needed--e.g. a scene with bright areas due to reflected sunlight of the ocean should be metered/adjusted to ignore this.

Ultimately it is about understanding exposure theory, and studying this will give huge benefits down the road. I really like "Exposure Manual," 3rd ed., 1974, by Dunn and Wakefield. Also Ansel Adams "Camera and Lens" (although we don't expliciely use his zone system, the way of looking at a scene in terms of zones is still valid/useful). They both [Dunn and Adams] discuss exposure meters (incident, spot, averaging) but that is less important than exposure in terms of mid tones and balance of light to dark areas.

These older books tend to also treat (practical) exposure theory, as opposed to (only/largely) saying use matrix here and use spot here, and so on. I have no idea if there are good current books on the subject--the few I have looked at are about mechanics and not practical exposure theory.

For what its worth my usual method for outdoor street/architecture work is as follows.

-- Full sun: sunny 16 rule (adjusted for whatever f-stop I set) in manual mode. This is actually a very tough scene to meter w/ camera but easy w/ sunny 16 rule or incident light meter.

-- Other: usually AV mode, center weighted, and set ev compensation based on my perception of scene proportion of light/dark areas and whether I want to bias exposure for lighter or darker areas.

[gofour3]

Indeed! May I ask where you got meter pattern diagrams for your cameras?

I found them in an old Pentax book that I have:

"How to select & use Pentax SLR Cameras" (Carl Shipman) published in 1977.

It has diagrams for M42, K & M series bodies.

Phil.

[stevebrot]

I found them in an old Pentax book that I have:

"How to select & use Pentax SLR Cameras" (Carl Shipman) published in 1977.

It has diagrams for M42, K & M series bodies.

Phil.

Cool! I found this diagram in a post by PF user @Cuthbert from several years ago.




Steve