Not all Matrix_meters are the same...

Not all Matrix_meters are the same ! I took all 3 pictures with about a 50mm lens, iso=400 and 3 different cameras all set to Matrix metering. The Canon 6D, the Nikon D7000 and the Pentax k5 IIs. The first picture is the Canon, the second the Nikon and the third is the Pentax. Too me it looks like the Canon meter was fooled less by the dark foreground ? I'm not really sure if this is a fair test ??

Not a fair test. The first picture has significantly more dark areas than the others. And the cameras are being asked to make an impossible decision about whether to expose for the very dark or very light areas.

 

Even if they were all exactly the same framing, there would still likely be different exposures, With matrix metering, you are depending totally on the camera makers method of calculating the exposure from the various sections of the matrix. And some of these methods include prioritising the area which the AF has focused on, assuming that this is the most important feature of the scene.

 

Your initial statement says it all. I often take pictures of cameras to sell - they too are difficult subjects for the metering, consisting as they do of mostly black and white or silver with few mid tones. With a Pentax DSLR I get good exposures with +1 compensation. With other makes I don't dial in compensation, although I nearly always have to adjust levels slightly.

Edited May 6, 2020 by John Seaman

Not a fair test. The first picture has significantly more dark areas than the others. And the cameras are being asked to make an impossible decision about whether to expose for the very dark or very light areas.

 

Even if they were all exactly the same framing, there would still likely be different exposures, With matrix metering, you are depending totally on the camera makers method of calculating the exposure from the various sections of the matrix. And some of these methods include prioritising the area which the AF has focused on, assuming that this is the most important feature of the scene.

 

Your initial statement says it all. I often take pictures of cameras to sell - they too are difficult subjects for the metering, consisting as they do of mostly black and white or silver with few mid tones. With a Pentax DSLR I get good exposures with +1 compensation. With other makes I don't dial in compensation, although I nearly always have to adjust levels slightly.

I went back and made sure that the distance, focal length, foreground, ISO etc. were the same(I just did not download those pictures) Again I got exactly, or pretty much exactly the same results. The Canon meter seems to prioritize the area around the focusing point . The other two cameras, even though I set the Pentax to link the Exposure to Focusing Point seem to take the entire scene into consideration. Try it out yourself(if you have all 3 brands) and let me know what you discover...

Edited May 6, 2020 by hjoseph7

You can often find how manufacturers devised the matrix algorithm in technical articles a few months after introduction. Over the years, after experiencing a variety of what I would consider exposure miscalculations from what matrix metering proposes, switched in most cases to a different weighting scheme. For 95% of my work, when I'm using reflective metering I find central weighted metering more to my taste. But given my choice, I most often use separate incident metering...it just doesn't ever fail me.

As a starting point I check exposure metering by shooting a plain grey surface and seeing if the spike of the histogram is right in the centre. It's worth mentioning too that lenses can also introduce exposure differences. I has a Canon 50mm which consistently overexposed by around 1 spot, even at full aperture, for no apparent reason.

My bottom line is that I would never use matrix metering for something like this. This is a situation that requires control by the photographer, not an algorithm.

 

It's not surprising that different companies' matrix metering programs would provide different results. The engineers have to make decisions about what weights to give to different areas, and to do that, they need to make assumptions about what is being photographed and how the photographer is acting.

Given this the Canon and the Nikon have done a job that works pretty well. Nikon is easily salvageable. Not sure about the Pentax though.

My bottom line is that I would never use matrix metering for something like this. This is a situation that requires control by the photographer, not an algorithm.

 

It's not surprising that different companies' matrix metering programs would provide different results. The engineers have to make decisions about what weights to give to different areas, and to do that, they need to make assumptions about what is being photographed and how the photographer is acting.

 

Setting your camera to Average Meter mode, or even Spot Meter mode would probably solve that problem...

Setting your camera to Average Meter mode, or even Spot Meter mode would probably solve that problem...

 

In the image in question, it looks like the range from light to dark exceeds the dynamic range of the camera. In that case, averaging isn't a good option, as it is likely to clip both ends of the histogram. If I were taking that shot and not bracketing exposures, I would choose an area that I know how I want to expose and spot meter off that, adjusting up or down as required.

 

In other words, I would leave no decision to the camera. I would make the decisions myself.

Why are you using 400ISO in bright sun conditions?

Any meter, be it CWA, MWA, Matrix, multipattern or reticulated, will have difficulty balancing highlight and shadows when using a high speed film in conditions it cannot record well.

 

A handheld multispot/duplex meter (not a camera's own meter!) can work wonders in this situation (one of very high differences of contrast) by hedging highlights down and balancing shadows through duplex multispotting and careful mean averaging. This takes skill and experience through observation and applied technique.

Why are you using 400ISO in bright sun conditions?

Any meter, be it CWA, MWA, Matrix, multipattern or reticulated, will have difficulty balancing highlight and shadows when using a high speed film in conditions it cannot record well.

 

This comment seems a bit off the mark, on two counts.

 

A TTL Camera's Light Meter doesn't concern itself one iota with what ISO is set: the setting of the ISO is totally irrelevant.

 

Secondly, balancing highlights and shadows for FILM, doesn't factor into the conversation at all - all three cameras are digital.

 

WW

I chimp at the display and adjust to look right. Then snap. I usually use center weighted. Then I can moe the camera until the exposure is right and lock exposure and reframe. Similar to when you use center focus. Lately I've been using my M4/3 as a meter for my 4x5 film shooting.

My Nikon is usually pretty good but will occasionally get smallish bright areas a bit over exposed. For those scenes I often knock the compensation back a stop or just go by the histogram. The EVF also lets you know quite quickly if things aren't quite right.

I agree it’s not a fair test. You are asking the camera to work out what it is you want to record accurately, but making it very difficult for it to determine that. Each manufacturer, as pointed out, has different ways of going about this, so it’s not surprising you get three different results. In this situation, don’t rely on matrix meter. Spot meter from an average tone, or what I’d do would be to walk into the scene, exclude the window frame, meter from the garden in matrix, hit the exposure lock, retract to original scene location and expose.

My bottom line is that I would never use matrix metering for something like this. This is a situation that requires control by the photographer, not an algorithm.

 

It's not surprising that different companies' matrix metering programs would provide different results. The engineers have to make decisions about what weights to give to different areas, and to do that, they need to make assumptions about what is being photographed and how the photographer is acting.

 

I agree with this - proper exposure is a decision for the photographer, not the camera, because the camera can't know your intent.

I agree with this - proper exposure is a decision for the photographer, not the camera, because the camera can't know your intent

 

Agreed. In addition, the camera can't know what the scene looks like. The metering systems are engineered with the expectation of certain characteristics that your scene may not share.