Sweet Spring Smiles

[Mike Buckley]

center weighted average - this is usually considered to be the most consistent form of metering

Considered by whom?

[Marie Hass]

Lols, Mike!

Good question! Several articles that I read and pictures I have taken and seen with the comparisons of the different metering systems suggest this conclusion.

You have other thoughts?

Marie

[Mike Buckley]

You have other thoughts?

Use the metering method that produces the desired results the most reliably for you in a given situation regardless of why that is so. You can attain the same fundamental exposure and histogram regardless of the metering method that you use if you know how to properly use all of them.

So, the key is understanding which methods for which you have (and don't have) a reliable working command. As an example, I have a working command only of center-weighted and matrix metering. Using spot metering requires a particular skill that I don't have.

[William W]

Dear William, Per your request, here is the original file. I have made it 1800 on the long side so you have something to work on. . . .

Thank you.

Firstly, it doesn't appear to be as underexposed for the skin tones as I first thought. I calculate it is about is about ⅔ Stop Underexposed for skin tone. (It’s late here - I have run the file through ⅔ Stop increase exposure and collected the results, but it takes time to collate them into JPEGS to illustrate, so I will do that tomorrow.)

What confused my initial analysis was the amount of universal darkening in Post Production: I thought that the darkening was more isolated to the background and that the hair resembled closer to a correct exposure for it (the hair). In other words I didn’t realize or rather didn’t consider the possibility that the top-light on the hair was so intense.

Knowing better what the Lighting scenario actually was, I’d leave it almost be and have the lighter OoF background. I wouldn’t mind a bit of the hair blowing out – what I probably would have done is moved her out of the harsh direct light on the hair, or shaded it.

I assume that “so you have something to work on” means you’d like me to redo the PP from the original file – and again, I’ll do that tomorrow.

***

Regarding your question about center weighted average - this is usually considered to be the most consistent form of metering, as most photographic subjects are near the center of the frame and rarely fall into the outer four corners.
. . .
Several articles that I read and pictures I have taken and seen with the comparisons of the different metering systems suggest this conclusion.

That’s interesting.

At the risk of sounding adversarial, which I am not, because I am certainly NOT challenging that you have read or have been told that using CWA will be the best Metering Mode for you to use, but:

I would like to read the sources and data for the statement:

“[Canon] center weighted average . . . is usually considered to be the most consistent form of metering usually considered to be the most consistent form of metering, as most photographic subjects are near the center of the frame and rarely fall into the outer four corners.”

***

I would be interested to hear your comments.

As far as my knowledge and research extends:

Canon’s DSLR’s CWA Meter Mode is a modification and adaptation of Canon’s original “Central Emphasis Metering System” which was used in Canon’s SLRs: for example the Canon A-1.

However, after Auto Focus Lenses were invented and subsequently with the advent of Multiple Auto Focus Points on cameras, the development of “Evaluative” Metering Mode progressed to take into account the area of sharp focus / the AF point used, for the TTL Meter’s calculation of the exposure.

In simple terms CWA remains essentially a ‘dumb’ weighted averaging system, but on the other hand Evaluative continuously makes a ‘active’ calculations by including the data from the AF points that are triggered “in focus”.

So yes, on the face of it, one would think that CWA would be the ‘most consistent’, because it is ‘dumb’ (i.e makes EXACTLY the SAME calculation all the time). But it appears to me, based on my own and others' testing (Nadine Ohara ran extensive tests in this area) that CWA isn’t always as ‘dumb’ as the old “Central Emphasis Metering System” that it was predicated upon. That was precisely my point when I wrote:

“I cannot get my head around working out accurately the parameters and functionality of Canon’s CWA Metering – it seems to have a mind of its own and a function of ‘Changing the Rules about how I will average’.”

*

On the latter part of the statement, the rationale for using CWA:
“as most photographic subjects are near the center of the frame and rarely fall into the outer four corners.”

I do not concur with this as an appropriate rationale for using CWA. Three main points:

Firstly, the area of the centre portion of weighting, (before the exposure is averaged), seems to work as a relatively small area of the whole FoV. Because CWA is (supposedly) a ‘dumb’ calculation this weighted area is prone to skewing the result even between shots of the same Subject. For example, if taking two Portrait images of the same person and one shot is a Bust Shot compared to Full Length Shot, but for both shots the Subject is in the centre of the frame.

Secondly, I disagree that centre-centric positioning is the aim of ‘most’ Photographer’s COMPOSITIONS.

Thirdly, if one wants to Lock AE from a particular AREA (not 'Spot'), then Canon Provide "Partial" Metering mode for this function. "Partial" has no weighting and no averaging, so one is assured of consistency. The only variance with "Partial" Metering Mode is the area that is metered changes between camera models, but (unlike the scant data on CWA functionality) Canon provides quite accurate data of the area of the "Partial" Metering for all its DSLR's.

WW

Addendum:

For absolute clarity, Marie, my preceding comments refer to CANON TTL Metering Modes, and NOT Nikon's TTL Metering Modes.

I did make one passing comment regarding the similarity of Canon's "Evaluative" Mode and Nikon's "Matrix" Mode.

I have little data and testing of Nikon's "Center-Weighted" Mode, but I do note that the functionality is DIFFERENT to Canon's "Center Weighted Average" Mode.

[William W]

. . . [Frame 01] doesn't appear to be as underexposed for the skin tones as I first thought. I calculate it is about is about ⅔ Stop Underexposed for skin tone. (It’s late here - I have run the file through ⅔ Stop increase exposure and collected the results, but it takes time to collate them into JPEGS to illustrate, so I will do that tomorrow.)

FYI - From top left to right:
1. Hist. of Original
2. Hist. of whole image with increase overall Exposure, approximating 0.8 Stop
3. Hist. of Skin Tones with increased exposure
4. Hist. of Redo Version II



***

I assume that “so you have something to work on” means you’d like me to redo the PP from the original file – and again, I’ll do that tomorrow.

As mentioned, now knowing better the lighting scenario of the shot, the key elements to this example of Post Production are:
> Crop to remove portion of L. Arm
> Retrieve highlights in Hair
> Correct to expose for Skin Tones
> Remove extraneous Wires and Foliage
> Minor Dodge / Burn and Colour Balance, to taste

The original is on the top, the quick and indicative redo (i.e. only basic functions as listed above), is on the bottom.



Obviously better results would be derived from using the raw file and spending more than just a few minutes on the PP.

Especially, I believe that we could get really good results regarding the highlight retrieval on the hair (It appears there is a lot of data there) and the exposure correction on the face would be enhanced by one of the various specialty Portrait PP tools.

***

SUMMARY:

We’ve covered a lot of ground on Frame 01, so just in summary, the take home points that I want to make, are:

1. Good shot.

2. Good balance, Juxtaposition of positive and negative space and good composition (not mentioned previously – the Subject’s Nose cutting the far Cheek line spoils the Portrait, but only a tad)

3. Nice shutter release timing for the decisive moment

4. Poor choice of Shutter Speed

5. Underexposed for correct rendition of skin tones

6. Possibly hard top direct lighting (on hair) could have caused irretrievable highlights – but my guess is no, it wouldn’t have: I expect that the Dynamic Range of the7D would have accommodated at least another 1Stop of Exposure, and all the highlight could have been recovered quite easily from the raw file

7. Poor choice and/or lack of understanding of Canon’s Metering Modes

*

Apropos General Artistic Critique & Comment:

Frame 01 is the pick of the four. It has better balance and composition than any of the other and is the least contrived. Had the technical aspects been better it would be a “Ripper”. It is still good as is.

Frame 02 is the next best, but the back rim lighting does not work (as is).

Although that image too is underexposed for Skin Tones, so with adequate PP, I think that the highlights of the rim lighting could be maintained and the L. Cheek and L. Eye area easily corrected.

But, in Frame 02, the foreground dominance of the Subject’s L. Hand is a positive entity.

As a positive entity, there is no juxtaposition, but rather, the L. Hand is vying for supremacy against the Subject’s Face and this battle is most disturbing for the Viewer’s Eye.

WW

PS - If you want any of those redo images for your reference please download them in the next few days, because I clear others' work from my holding bay, regularly.

[Marie Hass]

Dear William,

I cannot thank you enough for the more than generous expenditure of your time, detailed explanations and examples. It really helps to solidify what you are showing me and other CiC members.

I have downloaded the contents of this thread. I would like to take more time to examine everything and all of the comparison pictures you have presented.

Marie

[William W]

Good.

You are most welcome.

WW