Long live manual mode!

[Abitconfused]

I could only get an acceptable exposure of the dress using manual mode. The background was so bright that even spot metering did not work.



Even then, some work in PCC was required.

[AlwaysOnAuto]

Yeah, manual is going to be the only way to get that metered correctly for the dress.
Which camera?
And do you have the whole car?

[Manfred M]

I could only get an acceptable exposure of the dress using manual mode. The background was so bright that even spot metering did not work.

Did you try using exposure compensation before switching to manual? Both should give you very similar results.

That being said, the background is now very bright and distracting, which is not giving you the greatest image. The ideal approach in this type of situation is to hit the subject with enough light to overpower the ambient light and through that darken the background. I suspect that was likely not possible for this shot.

[Black Pearl]

An exposure comp dial in conjunction with a good evf is my preferred way to get something like this nailed, faster than switching to manual and you get the exact same result.

Personally I'd have held the highlights a little more then brought back the darker areas in PP.

[Abitconfused]

Yes, EC would be a viable alternative. Here I concentrated on minimizing background highlights at the risk of capturing a grayer world than exists. Is there a way to reduce highlights without making them turn gray?

[William W]

I could only get an acceptable exposure of the dress using manual mode.

How exactly did you achieve the “acceptable” exposure?

I ask because Manual Camera Mode does NOT affect how the TTL Light Meter works nor does it interfere with the information that the TTL Light Meter provides to you.

My point being is that the selection of Manual Mode had little to do with getting the “acceptable” exposure that you desired – what got you the “acceptable” exposure was how you used the information from the TTL Light Meter and translated that to your exposure settings and for example, that could have been using the TTL Light Meter as a starting point and making Exposure Brackets.

As mentioned using Exposure Compensation in any of the Semi Automatic Modes would have worked, also.

The background was so bright that even spot metering did not work.

When you write “did not work” it is assumed you mean “did not give the correct exposure for the dress”.

We should NOT expect a Nikon D810 when set to Spot Metering to give the correct exposure for the dress.

I know of no TTL Meter that is indeed a "Spot Meter" (i.e. Reflective meter with 1 degree angle of coverage)

I think that the D810's "Spot Meter Coverage" is about 3~4mm circle (relative to the area of the sensor's reception)
.
There are several reasons why we should not have the expectation that your D810 Spot Meter would give the correct exposure for the dress and the background’s light intensity is only SECONDARY to these reasons:

> the Red Dress is NOT equitable to Photographic Grey

> the TTL “Spot Meter” Area of Coverage in the SCENE is dependent upon the FL of the lens used

> the TTL “Spot Meter” Area of Coverage in the SCENE can be dependent upon the AF point selected

WW

[Abitconfused]

Oh, well, I didn't think about all that. I looked at the photo in the rear display.

[Manfred M]

Ed - you could try to bring down the highlights in PP to see what that does.

I normally would compensate in shooting and with a scene where the lighting is not good, like in this image, I would bracket three shots, probably at least 1 stop (possibly 2 stops) from what the meter is saying and then work with the best one of the lot. With the autobracket function on the D810 and the high speed burst setting, even with hand holding I will get shots that I can stack and combine in PP.

[xpatUSA]

> the Red Dress is NOT equitable to Photographic Grey

That comment caught my eye, Bill.

Being calibrated in lux, does that mean that the in-camera light-meter (if the dress were spot-metered successfully) would recommend a higher exposure than if the dress were green? And much higher if the dress were blue?

In other words, would the background exposure (with no EC) be thereby affected by the color of the dress?

OT, but I'm currently noticing things like that when shooting full-spectrum. The light-meter doesn't see the IR but the sensor sees lots of it, so I have to apply typically anything from -2/3 to -2 EC to avoid blown highlights.

Guessing the ratio of IR to visible in a scene is quite challenging! As Manfred said, bracketing is our friend.

[William W]

. . .

Being calibrated in lux, does that mean that the in-camera light-meter (if the dress were spot-metered successfully) would recommend a higher exposure than if the dress were green? And much higher if the dress were blue?

In other words, would the background exposure (with no EC) be thereby affected by the color of the dress?

Hi Ted,

Sorry for my delay in responding, I didn’t immediately notice your post.

My first answer, to your question as I understand its (deeper) meaning is: “I don’t know”.

My explanation is –

Part 1.

I think that in the brevity of my dot points I might have added a little confusion to my exact meaning.

Perhaps what I should have written and what would have been more clear, was -
“the SHADE of RED in the dress appears to me that it would not be equitable to Photographic Grey and therefore the TTL Light Meter would have provided a skewed indication anyway.”

From my field experience (NOT necessarily “Field Testing” but a bit of that also), I think that the red dress is a tad lighter than “Fire Engine Red” and “Lush Green Grass” and “Rich Blue Sky”: all three of which are “equitable to Photographic Grey” and all of which I will use, with more than reasonable confidence, for a Spot Meter Reading.

Part 2.

I think that you might have thought I was referring to “RED” per se and not “that SHADE of RED” and I further think that your question might have been more about the relative sensitivities to RGB of the typical Digital Sensors.

If my assumptions are correct, then I trust that this has answered your question?

Sorry if I was not clear the first time.

WW

[xpatUSA]

Hi Ted,

Sorry for my delay in responding, I didn’t immediately notice your post.

My first answer, to your question as I understand its (deeper) meaning is: “I don’t know”.

My explanation is –

Part 1.

I think that in the brevity of my dot points I might have added a little confusion to my exact meaning.

Perhaps what I should have written and what would have been more clear, was -
“the SHADE of RED in the dress appears to me that it would not be equitable to Photographic Grey and therefore the TTL Light Meter would have provided a skewed indication anyway.”

From my field experience (NOT necessarily “Field Testing” but a bit of that also), I think that the red dress is a tad lighter than “Fire Engine Red” and “Lush Green Grass” and “Rich Blue Sky”: all three of which are “equitable to Photographic Grey” and all of which I will use, with more than reasonable confidence, for a Spot Meter Reading.

OK, shades of color, I understand those along with tints, Thanks for the clarification, Bill.

https://en.wikipedia.org/wiki/Shades...ire_engine_red

Part 2.

I think that you might have thought I was referring to “RED” per se and not “that SHADE of RED” and I further think that your question might have been more about the relative sensitivities to RGB of the typical Digital Sensors.

If my assumptions are correct, then I trust that this has answered your question?

Sorry if I was not clear the first time.

WW

Thanks; photographic terminology is against us, but that's what was referring to indeed.

You're right, I was thinking more of the wavelengths reflected by 3 dress colors, say 450, 500 and 650nm (realizing that there are no such monochromatic-colored dresses in the Real World).

Ignoring dresses and metamerism, I was speculating that those three wavelengths, incident upon the TTL light-meter at equal radiometric power, i.e. not lux, would each cause a different EV to be recommended by the meter because the meter is calibrated to human photopic vision where green is "brightest" . . .

[William W]

. . . Ignoring dresses and metamerism, I was speculating that those three wavelengths, incident upon the TTL light-meter at equal radiometric power, i.e. not lux, would each cause a different EV to be recommended by the meter because the meter is calibrated to human photopic vision where green is "brightest" . . .

If you are free this weekend then pop down here and I can take you out and show you exactly what my “Fire Engine Red” and “Lush Green Grass” and “Rich Blue Sky” look like in the real world . . . it is not only jargon and terminology that are an hindrance to precise internet discussion.

***

I have been thinking on this and I reckon that you are on to something. I think it would be in the order of tenths of stops: that’s a postulate and merely a rudimentary one.

Perhaps on first glance seemingly obtuse, yet possibly related: I had a long discussion with a bloke around 2004 about the value of CC (Colour Correction) Filters on DSLRs.

The general view is (and was) “CC Filters are not necessary in any way shape or form for digital – just set the Colour Temperature in camera”.

Lester reckoned that CC filters were often effective for attaining better general Image Quality, he had done some rudimentary trials.

I did a few more and what I found when I plonked the appropriate CC filters on to ‘correct’ low level incandescent lighting (about K = 2000~2400°), moving that scene to about K = 3000° I got a better final image: better still if I shifted the scene to K = 5000°.

In all cases for these tests I used Manual Colour Temperature selection on my DSLR and I was using a Canon EOS 20D.

Maybe my memory is phased, probably is somewhat, but I believe the TTL Meter’s indication was more accurate with the CC Filter on the lens. . . I think this point is related to your thoughts.

Of course the practicality of using CC filters in low level incandescent lighting is limited, even if using CC Filters does assist for a better final image and/or a more accurate TTL Metering, because the ‘stops lost’ in the Filter Factor means that the use is limited to shots where a short shutter speed is not mandatory AND/OR (really) high ISO is acceptable.

Anyway, if you play with your RGB idea, I would be interested to know what you find.

WW

PS had to look up the meaning of “metamerism” - cool word and I shall endeavour to use it wisely, thanks.

[xpatUSA]

If you are free this weekend then pop down here and I can take you out and show you exactly what my “Fire Engine Red” and “Lush Green Grass” and “Rich Blue Sky” look like in the real world . . . it is not only jargon and terminology that are an hindrance to precise internet discussion.

Indeed . .

I have been thinking on this and I reckon that you are on to something. I think it would be in the order of tenths of stops: that’s a postulate and merely a rudimentary one.

out of curiosity I looked at the dress in RawTherapee. It's almost pure sRGB red and pretty close to CIELAB mid-gray!!

Anyway, if you play with your RGB idea, I would be interested to know what you find.

I downloaded the CIE luminous efficacy diagram which can be used to relate radiometric and photometric stuff. I added the sRGB primaries and obtained normalized photometric outputs:



Looks like the red puts out about half the lumens per Watt, about -1EV. And the blue about 1/8 the lumens per Watt, about -4EV . .

All in theory of course . . no doubt the Real World will vary . .

[William W]

. . . no doubt the Real World will vary . .

Yep the first thing we gotta do is pop over to Bakersfield and find that red dress!

WW

[Cantab]

Bill and Ted, I've found your discussion very interesting, even though parts of it are at or beyond my limits of full comprehension.

[xpatUSA]

Bill and Ted, I've found your discussion very interesting, even though parts of it are at or beyond my limits of full comprehension.

Thanks Bruce!

[William W]

Bill and Ted, I've found your discussion very interesting, even though parts of it are at or beyond my limits of full comprehension.

Ta.

I am happy to explain and or comment further if you have any questions - probably best if you start another thread, referencing this one.

I'd reckon that Ex-Pat in Texas would be willing to plonk in few comments too.

WW