Metering mode for portraits

AJSJones wrote in post #18546028
I am not convinced there is a distinction. The camera’s processor “evaluates” the image - I seriously doubt that either carries the database (from which their algorithms were derived) in the camera and that both apply criteria based on their respective databases. In other words, they basically use the same principle: Historical database of wide range of pictures -> evaluations by users for each image -> develop algorithms to match those different (styles/characteristic​s of) images - load algrithms to camera. But, that is all proprietary so we won’t know.

I agree with you, there is no way that either camera contains within it a huge database of different images and user evaluations. At most, they will contain a reduced set of image 'characteristics' that the camera compares and seeks to find a nearest match against before making its metering decision. Effectively, there will be a set of rules derived from prior analysis of the 'database'. Those rules might be different, however, from the set of rules derived from a fundamentally more prescriptive approach.

Me ... I just don't use evaluative metering. I am no more interested in: a) some formulae derived from a database of other people's photographs and other people's opinions of them; than I am in b) another set of formulae derived from some set of expert rules that have, unfortunately, not been disclosed to me. Either way, I am just passing control over exposure to a black box that I really know almost nothing about.

Wilt wrote in post #18546022
We will never know, will we? But one would think that Canon Marketing would have made specific statements along the line of 'Our algorithms are smart ones' so as potential buyers do not assume that Nikon is better than Canon in that regard. The fact that nothing is said is either

• Canon's evaluative is simply not as 'intelligent' so there is no basis for a statement to be made, or
  
• They want us to simply ASSume theirs is just as good as the competition's
  

While it might be confirmed that Nikon has used databases of images to come up with algorithms used in their meters, I doubt that's all they did in the development of their current meters. Companies like Canon and Nikon have a lot of R&D...and for effectively metering a scene with different situations, they're looking at the physics of the scene (both directly and through recorded images). It's also algorithms that are derived from studies that are used with processing. And as earlier linked, Canon has been highly advertising their "intelligent subject analysis system", that was introduced with the 1DX.

I guess it would be interesting to see evaluations of the 5D4's evaluative metering (and compare it to earlier models like the 5D3).

DaviSto wrote in post #18546035
I agree with you, there is no way that either camera contains within it a huge database of different images and user evaluations. At most, they will contain a reduced set of image 'characteristics' that the camera compares and seeks to find a nearest match against before making its metering decision. Effectively, there will be a set of rules derived from prior analysis of the 'database'. Those rules might be different, however, from the set of rules derived from a fundamentally more prescriptive approach.

Me ... I just don't use evaluative metering. I am no more interested in: a) some formulae derived from a database of other people's photographs and other people's opinions of them; than I am in b) another set of formulae derived from some set of expert rules that have, unfortunately, not been disclosed to me. Either way, I am just passing control over exposure to a black box that I really know almost nothing about.

If it was something that had been just introduced and no-one had had any experience with, I’d be skeptical too. I’d wait for “early adopters’” feedback and get my own experience. Oh, wait, that is all a long way on the past. I’ve learnt how to use it like I use autofocus - sometimes I decide to override it, sometimes I use manual/spot exposure. (I don’t need to know the formulae either, or how the AF processor works, just that it does) Learn to use the available tools - heck, I even use auto-ISO sometimes!

AJSJones wrote in post #18546054
If it was something that had been just introduced and no-one had had any experience with, I’d be skeptical too. I’d wait for “early adopters’” feedback and get my own experience. Oh, wait, that is all a long way on the past. I’ve learnt how to use it like I use autofocus - sometimes I decide to override it, sometimes I use manual/spot exposure. (I don’t need to know the formulae either, or how the AF processor works, just that it does) Learn to use the available tools - heck, I even use auto-ISO sometimes!

... and that's fine ... but I have my own ideas (as I am sure you do, too) about how to expose different scenes in different lighting conditions. All I need is a solid basic platform that I properly understand as a starting point before I adjust the camera's settings to achieve what I want (hey ... I'm not that good ... so let's make that "what I think that I want"). Evaluative metering doesn't give me that. An incident light meter reading, partial metering, or spot metering off some appropriate object (by camera or light meter), will all give me that. I prefer to an incident light meter reading to get that starting point, personally ... because it's so simple, quick and straightforward ... but I know some people think that's a little prehistoric.

DaviSto wrote in post #18546060
... and that's fine ... but I have my own ideas (as I am sure you do, too) about how to expose different scenes in different lighting conditions. All I need is a solid basic platform that I properly understand as a starting point before I adjust the camera's settings to achieve what I want (hey ... I'm not that good ... so let's make that "what I think that I want"). Evaluative metering doesn't give me that. An incident light meter reading, partial metering, or spot metering off some appropriate object (by camera or light meter), will all give me that. I prefer to an incident light meter reading to get that starting point, personally ... because it's so simple, quick and straightforward ... but I know some people think that's a little prehistoric.

Agreed. But fundamentally Evaluative was designed for the snapshooter turned SLR owner because he/she could afford more than a point-and-shoot. So any form of 'intelligence' in Evaluative beat the lack of understanding of the typical snapshooter...'put some of the insight of a pro into the mindless shooting of the snapshooter' was the goal!

I tested Evaluative in a purely mindless mode while shooting, just to see how it would do if my brain was on holiday, and I have to admit it did darned well. Maybe not as well as an understanding photographer might do with some other metering mode and time to think about it, but if you suddenly had to point in a different direction and shoot as fast as you could (imagine yourself in NYC on 9/11 as a jet was looming close to the WTC), it was very good!

In any event, even Evaluative cannot intelligently make up for the contextual subjects -- ones considerably different in brightness from 18% gray.

Wilt, how about taking a frame-filling photo of a landscape poster and then inverting the poster?

Alveric wrote in post #18546069
Wilt, how about taking a frame-filling photo of a landscape poster and then inverting the poster?

That sounds like a great test for you to perform and report back here. But rather than a reflective print, I think it needs to be a 24" x 36" Duratrans transparency backlit on a lightbox, so the dynamic range of a real scene is more closely mimic'd by the photo!

Or evaluative is for the professional photographer who doesn't have the time to spot meter or have an incidence meter (like sports photography on the 1DX). When I was using the 5D, I primarily spot metered scenes (I needed to be more critical since I didn't have much exposure latitude). Now with newer DSLRs, I primarily use evaluative: and I feel a bit relieved that I don't have to spend as much time taking readings within my scene. That is not to say I'm not cognizant of how I want to expose....and will EC in certain situations or pick other modes for conditions like back-lighting.

davesrose wrote in post #18546080
Or evaluative is for the professional photographer who doesn't have the time to spot meter or have an incidence meter (like sports photography on the 1DX). When I was using the 5D, I primarily spot metered scenes (I needed to be more critical since I didn't have much exposure latitude). Now with newer DSLRs, I primarily use evaluative: and I feel a bit relieved that I don't have to spend as much time taking readings within my scene. That is not to say I'm not cognizant of how I want to expose....and will EC in certain situations or pick other modes for conditions like back-lighting.

I think everybody should do what works best for them, individually ... so I am definitely not going to argue this.

DaviSto wrote in post #18546084
I think everybody should do what works best for them, individually ... so I am definitely not going to argue this.

Absolutely....just thought I'd bring up examples as to why evaluative is not just for "snapshooter turned SLR owner". I was a film SLR shooter, I exclusively used spot metering when I needed it, and I see that there's a place for evaluative as well.

davesrose wrote in post #18546085
Absolutely....just thought I'd bring up examples as to why evaluative is not just for "snapshooter turned SLR owner". I was a film SLR shooter, I exclusively used spot metering when I needed it, and I see that there's a place for evaluative as well.

Hey, pros can use full auto and even auto ISO! You use the best tool in the kit that suits the situation. Nothing wrong at all with automation, if one understands the limitations inherent to the automation. If I were a pro on the NYTimes staff, I think Evaluative would have been perfect way to capture on 9/11 as a jet was looming close to the WTC while coming up from the undergraound where I just had exposed some shots down in the subway for a human interest story in manual exposure, and no editor wants me to take the time to tweak my RAW files but wants instant JPG sent.

It does not change the fact that in the late 1970's-1980's when the amateur shooter was being lured to SLRs with compact SLRs like the Canon AE-1, that a member of the Nikon compact F-series Nikon FA launched the multi-segmented (matrix/Evaluative) metering in 1983, which ads stated compared the result of metering with over 100k other photos in its database.

Wilt wrote in post #18546099
Hey, pros can use full auto and even auto ISO! You use the best tool in the kit that suits the situation. Nothing wrong at all with automation, if one understands the limitations inherent to the automation.

+1 (or perhaps +12,800 in some cases)

Wilt wrote in post #18546099
It does not change the fact that in the late 1970's-1980's when the amateur shooter was being lured to SLRs with compact SLRs like the Canon AE-1, that a member of the Nikon compact F-series Nikon FA launched the multi-segmented (matrix/Evaluative) metering in 1983, which ads stated compared the result of metering with over 100k other photos in its database.

But those are different periods of history. The Canon AE-1 came out in 1976 and was advertised for being the first SLR with a CPU. The CPU would be considered extremely primitive by today's standards, but it adjusted exposure for shutter priority. The camera itself used center weighted metering. And when Nikon's first matrix metered cameras came out, I'm sure they didn't state the camera's meter compared over 100k other photos in a database: the camera would not be able to have the memory or processing power for such a thing. They may have said they evaluated 100k+ photos in their development of the meter. Put it this way: if Nikon was the only company advertising the sophistication of their new meters, they would be the ones with the most sales with their "consumer" grade SLRs. The fact that new "intelligent" evaluative meters are also being marketed in the 1D series, shows that it's not just aimed at the amateur, nor solely being marketed by Nikon.

Edit: the wikipedia article on the Nikon FA does go into specifics about the matrix meter and processor. The advertising was that the algorithms were developed from comparisons of 100k photos:

Nikon FA

Oh but they did advertise exactly the 100k database point! I did not claim the entire database was somehow within the FA programming database. Nippon Kogaku said that the program was written after the visual assessment of nearly 100,000 photographs. The ad itself:
http://www.mikeeckman.​com …ads/2016/02/Nik​onFAad.jpg

Eh no where do they use the words "database", nor being used in the camera itself. The point I was raising...since post #32, is that they developed algorithms from R&D. Now both the wikipedia article and your link's advertisement says that they developed the algorithms from visually comparing 100,000 photos. Wouldn't you think that both Nikons and Canons R&D over scene metering has advanced since 1983?