[mymage]

Hi, i want to know your opinion on the possibility to add a metering method in the DSLR camera: using the flexibility of the software update, it will be easy to insert a multi spot mean measuring method accessible via menu. This method works in these way: setting the spot metering method the camera get a reading every time you half shot the shooting button and memorize it. After a series of readings (than can be theoretically much but in practice I think 10 are enough ) the camera calculate the mean based on the setted exposition method (P, Av, Tv, ecc.).
You can say: "why you don't calculate it manually?". Ok, but why don't use the power of your camera?

By

Stefano

[Bobe416]

Not sure I understand the need for such a feature. Can you give an example(s) of when you would use this new metering feature.

[Lowell Goudge]

someone can correct me if I am wrong, but this is the theory behind matrix metering.

The idea is that the multi metering points (11 on all pentax DSLRS except the K7 which is 77 i believe), are used to develop a better model of the scene with highlights and lowlights, to determine the best compromise for exposure.

Center weighted metering is somewhat the same, where it takes a single reading but averaged over the entire frame with some bias. On center weighted metering, it is biased towards exposing the lower half lf the screen more (assumes that the shooter has a bright sky in the background, that can be blown out to some extent) Don't believe it, try shooting the exact same scene best is a landscape with the sky, with the camera right side up and upside down, and look at the difference in metering for hte 2 shots.

[mymage]

Bobe416->
In cases when there are some elements that you want to expose right (or try to best expose) in a situation with "dark and white" areas.
These is only a comfort, you can live and photograph without it.

[mymage]

Lowell Goudge->
Yes, matrix method do it well considering all the image. My suggestion is for focusing on "some" elements of the image.

[Lowell Goudge]

Lowell Goudge->
Yes, matrix method do it well considering all the image. My suggestion is for focusing on "some" elements of the image.

In other words you want to be able to program selectively the matrix metering points, not to have "multiple spot metering"

I think this would be a nightmare, since matrix metering is set up to consider typical shooting modes, and evaluate them based on some rather nebulous concept of scene composition.

I think the best you could hope for is to somehow modify the metering by selecting a "key point" of the scene based upon the AF point used. To me this would be logical because the AF point is normally also where you want exposure to be perfect.

To have a fully programmable matrix metering, I think it would be too difficult and time consuming to set for each different scene, and after all, that is why you are a photographer, it is up to you to check exposure on each scene, if it is that radical, by using spot metering and make the decision yourself.

I don't think the camera has the capacity to understand how you want to shoot every scene. Until someone develops a Mindread.exe function for the camera, I believe this is something best left to the photographer.

[Gimbal]

The thing is that it's probably faster to fire off a test shot, look at the histogram and compensate if needed.

[Lowell Goudge]

The thing is that it's probably faster to fire off a test shot, look at the histogram and compensate if needed.

except the problem here is the histogram is considered for the whole image not for specific zones. Maybe, and this would be a stretch of the imagination, it would be nice to have 9 histograms, one for each zone dividing the scene by 3 in each direction.,

note that this is not impossible in terms of viewing on the K7 for example, with a nice big screen, and also that on even the smaller screen of the k10 there is sufficient space for 4 histograms presnetly, greyscale and each of the 3 colors, so why not overlay 9 histograms on the image directly, greyscale for each zone?

to me that would be much more useful in evaluating the scene than having programmable metering.

[Marc Sabatella]

I think this would be *extremely* unlikely to produce reliable results. Unless you knew *exactly* which spots to pick, and *perfectly* balanced the number of brighter-than-average and darker-than-average spots so that a simple mean was not skewed, chances are you'd just end up shooting yourself in the foot. A far simpler, faster, and much more reliable way to do what you want is to simply use M mode. Hit the Green button to check shutter speed at a number of different points within the scene, then simply spin the dial to choose a speed within the suggested range. That's what I do, and I trust my ability to choose an appropriate shutter speed withn the range far more than I'd trust myself to chosoe exatly the right spots to rig the calculations in the camera to come u with the right answer using your scheme.

[jmbradd]

I see where you are going, sort of like how we have the ability to select an AF points, if you selected a few points in the scene you wanted to meter. It makes sense to me theoretically as there could be some dark spots you accept as dark and do not want to consider them in your metering calculation.

Correct me if i'm wrong but isn't this the old style "zone" metering system from before AE?

Part of me thinks at least one camera company has tried it during development and it failed. The trade-off is paying a bunch of programmers to come up with a mode that few people would likely use. This does tickle my curious side though, if I were a better programmer i'd love to try to program such a thing Guess the next step could be tying it in with HDR mode where you set "priority points" and if those points are not within an acceptable exposure threshold then switch to HDR mode and take a burst of pictures with the appropriate values to get those points you selected and compile the HDR.

Again, it doesn't seem like such a mode is commercially viable but for someone who enjoys a programming challenge, it could be interesting to see what the results are.

[creampuff]

OP, multi spot metering is not new but was pioneered by Olympus in the early 1980s with their OM-3 and OM-4 SLR film cameras.

The cameras allowed the user to take one or several spot meter readings, which were cumulatively used by the camera to calculate the correct exposure. There were also "highlight" and "shadow" buttons which allowed the photographer to identify parts of a scene that should be exposed as white or black respectively. The thing is to be able to select the appropriate area to take a spot reading is always the challenging part for the photographer. It is a skill that the experienced photographer is better able to exploit. However with the immediate feedback offered by digital cameras today (instant playback plus histogram and shadow/highlight clipping warnings), it is really a lot quicker to do a test shot/digital preview and adjust exposure accordingly or compensate by bracketing or using the AE-L button. Makes use of multi-spot metering pretty much redundant.

[MPrince]

You can say: "why you don't calculate it manually?". Ok, but why don't use the power of your camera?

By

Stefano

Because the power of the camera does not come close to matching the power of the human brain?

[Marc Sabatella]

Because the power of the camera does not come close to matching the power of the human brain?

I somehow missed that last line about using the power of the camera; I of course gave an answer that amounted to "why not do it manually". I'd clarify this slightly: the power of the camera is great in some areas, but not in others. If it still relies on the human brain to choose the spots to meter, then it's still limited by the power of the brain in doing so well. That's the hard part, not the figuring out that somewhere between 1/60" and 1/120" lies 1/90".

[Ira]

I'm holding out for this:

A metering system where the sensor identifies the sky and correctly adjusts only THAT portion of the scene, and can also identify the green grass and correctly adjust for THAT portion, and correctly adjust for the skin tone of your subject.

In other words, various exposures in the same frame.

It will definitely happen one day, when we're all dead. But for now, I have enough problems with a fixed ISO and two variables.

[cbaytan]

Aha! I knew this question will be coming one day, as an ex-Canon T-90 user I know the concept very well, instead of typing let me paste wiki's description:

Canon T90 - Wikipedia, the free encyclopedia
"
To these, Canon copied the metering options found on Olympus' OM-4.[5] Multi-spot metering allowed the photographer to average up to eight spot meter readings from different parts of the scene. In another feature borrowed from Olympus, separate Highlight and Shadow spot readings could be taken. These adjusted the camera's metering decisions to ensure extremes of tonal range were not muted and grey in the final exposure.
"

But don't forget one part of the scene can be marked more than once with multi-spot metering, for instance if there is important 3 parts of the photo (say dark part, bright part and a human face), you can mark face 2-3 or 4 times and mark the other areas just once, becasue face is more important to you, then cross your fingers, camera takes all these measurements adds it up and divides to how many times multi-reading button pressed.

I used the method a few times expecting to fail but it never let me down, all exposures were perfect, I think I was being just lucky, because I second the Marc's idea averaging exposure measurements can lead to mistakes.

Is it useful? I doubt it , if the face is important to me, I measure the face and then do the adjustment and check the results on the spot. This is a film era technique you could see the results maybe a week later, I guess thats why photographers preferred these method to ensure their readings are correct probably they were bracketing actual multi-spot exposure later.