Exposing for different skin colours
- Thread starter Jaffster
- Start date
- Messages
- 8,247
- Name
- Terry
- Edit My Images
- Yes
So how should he meter if both bride and groom are black ... what would that do to the dress. and his dark suit.?
Things like this make for a good teaching learning exercise...
- Messages
- 2,803
- Name
- Derek
- Edit My Images
- Yes
Nah, you can't win with the PC mob these days - my wife is 'black' but takes offence at that, as she's actually mid-brown. She wrote something about 'coloured' on a forum once and someone went PC on her calling her a racist pig..... not realising that she was actually non-white... coloured covers everything from white to black she says. 
- Messages
- 23,200
- Name
- Richard
- Edit My Images
- No
True
But are you trying to start another war.
I'm not looking to start a war either. I just want to make a couple of comments
The point you make about how adjusting the ISO on a hand meter being somehow different to adjusting it on a camera. There is absolutely no difference between changing it on a hand meter and adjusting the exposure compensation on the camera.
In both cases you are changing the calibration of the meter, and therefore shifting the 18% grey tone point. How the sensor, or film, then reacts to that change is a separate and independent matter in both cases. And so, I don't see the relevance there.
Secondly, and this is the hub of what I'm trying to say about the basic problem of incident light measurement when applied to digital. While adjusting the ISO to customise an incident reading works up to a point, it is not very acurate because an incident reading, however you custom calibrate it, assumes the important white tone to be a fixed value above 18% grey. The reality is that it changes according to the subject, and to what is decided by the photographer to be an important white tone, and what is not important.
The fundamental difference between digital and slide film is that there is a post processing stage, that is variable and can make big changes to the final output. (you can vary it in-camera too, via the JPEG pre-sets). No such variable exists for slide film.
Shooting digital is similar to shooting negative film in this respect, and you would not use an incident reading to set an optimum exposure level for negative film. Ansel Adams would turn in his grave.
- Messages
- 8,247
- Name
- Terry
- Edit My Images
- Yes
I'm not looking to start a war either. I just want to make a couple of comments
The point you make about how adjusting the ISO on a hand meter being somehow different to adjusting it on a camera. There is absolutely no difference between changing it on a hand meter and adjusting the exposure compensation on the camera.
Adjusting the Iso on a meter simply adjusts the readings.
Adjusting the ISO on a Digital camera adjusts the sensitivity.
However Adjusting the exposure compensation on a digital camera is not the same as changing the ISO. compensation shifts the exposure not the ISO.
The above answer clears that up.
An incident meter does not set a white point at all. it sets an 18% point and thereby all points above and below that ( it is a sliding tone scale) How far it can be slid is what you are calibrating. The ultimate white that can be recorded from a white surface, what ever the illumination, is that obtained from magnesium oxide only spectral reflctions will be brighter. If you set your incident meter to achieve this. highlights will never burn out. and you will have set as far to the right as you can.
It is an extremely accurate system, which is why it is used by the film industry.
[quote)
The fundamental difference between digital and slide film is that there is a post processing stage, that is variable and can make big changes to the final output. (you can vary it in-camera too, via the JPEG pre-sets). No such variable exists for slide film. [/quote]
True
The latitude in medium and fast black and white negative film is so huge
That it matters little how you set your meters or which type you use. in days gone by you tended to expose for the shadows and developed for the highlights. This is still an excellent method with large format cut film. however it is does not work for roll films as not every shot requires the same treatment.
If you are using slow black and white film you latitude is greatly reduced and I have always used an incident meter for this, as it reduces the chance of burning out the highlight detail. By using a beutler type developer you can both increase the shadow detail and increase sharpness.
I have always exposed Negative colour film as if it were transparency for constant results the Latitude on neg colour films falls somewhere between black and white medium speed film and transparency. As loss of detail in highlights is usually the more serious problem the use of an incident meter is again preferable.
Ansel Adams used "the expose for shadows develop for highlights method", modified to place the various tone points where he wanted them.
I do not know what meter he used But Incident metering is often described as providing a means of setting his tone values.
It is a mistake to think that light measurement is any different for digital or analogue. In both cases the measured light is used to set an exposure, that ensure all tones are reproduced on the film or sensor neither blocking shadows nor burning highlights. in Digital and transparency work the greatest problem is in the highlight areas as they are unrecoverable. The latitude in normal black and white film work is so large that the final print could never encompass all the available tone range. this is what we mean by "latitude"
In negative colour work over exposed highlights are also a problem as the options available to a hand printer are rarely available. and it is always easier to open out the shadows.
I tend to think about what I want to achieve, rather than what automatic systems a camera will allow me to achieve.
This is why when taking panoramas it is vital to set everything to manual.
left to its own devices and in what ever mode, a camera will set different exposures and colour setting to each individual shot, in an attempt to standardise its settings to the reflected information.
The fact is, the light has not changed between shot nor has the overall view.
What the cameras reflected reading gives is a patch work.
What is require is a singe fixed exposure to all parts of the scene.
This actually demonstrates quite well the superiority of the incident method , as in all cases a reading for each shot would arrive at the same result.
- Messages
- 18,182
- Name
- Geof
- Edit My Images
- Yes
white caucasian skin tone (palm of hand)in the light which illuminates the group as a datum then +/-..after all digital film is cheap and you get it processed next second..
there has to be a mid way between white and black...so that wipe out and mud dont appear
as you point out though...some latitude in the quantity,direction and quality of the light
white should show white, white caucasian should be ...white caucasian and black as in negro should compliment that tone and not grey it up or blacken it
and any black shadow areas...not too dense...
"to hesitate is to hold back" mao tse tung
- Messages
- 18,182
- Name
- Geof
- Edit My Images
- Yes
white caucasian skin tone (palm of hand)in the light which illuminates the group as a datum then +/-..after all digital film is cheap and you get it processed next second..
there has to be a mid way between white and black...so that wipe out and mud dont appear
as you point out though...some latitude in the quantity,direction and quality of the light
white should show white, white caucasian should be ...white caucasian and black as in negro should compliment that tone and not grey it up or blacken it
and any black shadow areas...not too dense...
"to hesitate is to hold back" mao tse tung
- Messages
- 23,200
- Name
- Richard
- Edit My Images
- No
Thanks for your reply Terry. I appreciate it.
All I can really say is that I disgree with pretty much everthing you've said! And what you say about colour neg film is not true - it has the most exposure latitude of any film, because it has development inhibiter release (DIR) in the colour couplers. That's why it is so successful in fixed exposure disposable cameras.
But I don't want to take the debate way off into what is mostly technical hypothesis, even though you and I would probably enjoy it!
I'll just say this though, which is really the only point I want to make and is not particularly aimed at you. That is, there is a general belief that an incident exposure reading will somehow deliver a perfect digital result, even when applied directly without any modification.
This is not true. It generally results in slight underexposure (depending on the camera) and is rarely anywhere near optimum if you expose to the right, which is one of the gifts of post processing digital.
Thanks again
- Messages
- 8,247
- Name
- Terry
- Edit My Images
- Yes
We clearly disagree, and both have systems that work for us. so best leave it there.
in today's world it is unlikely incident meters will ever be more than a minority interest for stills photographers. except for a few film users.
The cine world will continue to use them as they always have.
I am not sure why we disagree about negative colour as the latitude is well known and published .
what ever the latitude... you get far better prints from a correctly exposed shot than one that takes advantage of the latitude, when using it.
- Messages
- 6,481
- Name
- Tim
- Edit My Images
- No
If you can shoot a man wearing a white shirt and dark suit, and retain detail in both, then surely it is a piece of cake to shoot pale skinned people and dark skinned people in the same frame. Two obvious options....
- Meter and expose to the right for the highlight details - e.g. spot meter off the brightest part of the dress at (close to) +3 for a Canon camera shooting raw;
- Meter for the incident light - I'm happy to spot meter off my own palm at +1.3. That will simulate metering at 0 from a grey card and will place pale skin tones exactly where they belong in the tonal range, also holding highlight detail in pale fabrics in the same light, should there be any bright tones.
Both should get you results you can work with. If you can arrange for low contrast lighting (open sky, overcast, shaded or softened with fill/reflector etc.) then your life will be easier.
Here's an example with mixed skin tones and easily captured in one exposure with no adjustments. Soft, even lighting made my job very easy....
Here's another example showing the white of the dress and the dark of the suit fully captured in one frame with only a little insignificant clipping at the extremes of the dynamic range, both of which could be addressed with a small tweak to levels/curves or highlight recovery. I've enabled clipping alerts so you can see just how little clipping there is. Flesh tones of any shade should fit comfortably within the dynamic range of this scene.
If you shoot JPEG then you might find you want to ease up on the contrast a little. If you shoot raw then you can shoot to capture maximum data (ETTR) and then fine tune the appearance of the scene at your leisure later on.
- Meter and expose to the right for the highlight details - e.g. spot meter off the brightest part of the dress at (close to) +3 for a Canon camera shooting raw;
- Meter for the incident light - I'm happy to spot meter off my own palm at +1.3. That will simulate metering at 0 from a grey card and will place pale skin tones exactly where they belong in the tonal range, also holding highlight detail in pale fabrics in the same light, should there be any bright tones.
Both should get you results you can work with. If you can arrange for low contrast lighting (open sky, overcast, shaded or softened with fill/reflector etc.) then your life will be easier.
Here's an example with mixed skin tones and easily captured in one exposure with no adjustments. Soft, even lighting made my job very easy....
Here's another example showing the white of the dress and the dark of the suit fully captured in one frame with only a little insignificant clipping at the extremes of the dynamic range, both of which could be addressed with a small tweak to levels/curves or highlight recovery. I've enabled clipping alerts so you can see just how little clipping there is. Flesh tones of any shade should fit comfortably within the dynamic range of this scene.
If you shoot JPEG then you might find you want to ease up on the contrast a little. If you shoot raw then you can shoot to capture maximum data (ETTR) and then fine tune the appearance of the scene at your leisure later on.
- Messages
- 8,247
- Name
- Terry
- Edit My Images
- Yes
The scenario of two dark skinned people and dark and light clothing. gives little real choice, and is faced by wedding photographers all the time.
Burnt out wedding dresses are simply not acceptable so you have to expose so as not to clip the highlights. On a dull day the result will be fine with an exposure based on your own hand. Details in the dark faces and suit will also be apparent.
However in bright sun the best option is usually to put them with their backs to the sun and use fill flash. It works every time, the rim light will burn out bits of the dress but that always looks good.
The flash is best mounted on an old fashioned flash bracket a little way off to the side. It gives better modelling on the dress and faces.
I was first shown this by an south African wedding photographer who subbed for me in the 70's
It actually gives fine results what ever the skin colour.
It also means the pesky guests trying to get in on the act get terrible results.
- Messages
- 8,247
- Name
- Terry
- Edit My Images
- Yes
Your post demonstrates what I was saying very well... Nice shots.
However back lit shots really jump out using fill flash.
Like all photograph it is all about management of light, exposure then becomes the easy part.
The real difficulty comes when you have the very Black Black skinned types, they really want to see more detail in their faces than you can easily retain by exposure alone.
- Messages
- 23,200
- Name
- Richard
- Edit My Images
- No
LOL! On this one, perhaps we don't disagree after all! Just shows how semantics can sometimes confuse things. Exposure latitude, dynamic range, exposure tolerance etc, can mean the same thing, or quite different things.
Yes, you will get a much better result if you expose colour neg film correctly, even if you can abuse it, blast it to bits and still get away with it. Not the same thing
- Messages
- 18,182
- Name
- Geof
- Edit My Images
- Yes
hey!
how do our eyes cope
we look and we see
the camera cant...its just one exposure setting per frame...shot
so you have to bracket or move up or down from the datum you choose..
unfortunately i havent had experience with weddings...nor black people..my son in law is sort of fresh prince of bel air colouring
and recent shots came out ok not withstanding any other colours or tones
what did i want...i wanted his skin to come out ok...everything else was not important
setting fences to jump sometimes means you set them too high to prove a point
how do our eyes cope
we look and we see
the camera cant...its just one exposure setting per frame...shot
so you have to bracket or move up or down from the datum you choose..
unfortunately i havent had experience with weddings...nor black people..my son in law is sort of fresh prince of bel air colouring
and recent shots came out ok not withstanding any other colours or tones
what did i want...i wanted his skin to come out ok...everything else was not important
setting fences to jump sometimes means you set them too high to prove a point
- Messages
- 6,481
- Name
- Tim
- Edit My Images
- No
Well, I would usually shoot for the highlights and process for the shadows, and that ought to allow me to pull up shadow detail a little if I need to. So long as my exposure is basically sound, and my ISO not too high, pulling up shadow detail should be no problem.
Here's an example where I did screw up the exposure and I've had to boost the exposure in Lightroom by 1.3 stops and lower the black level from 5 to 1. The flesh tones have come up well and I can't see any concerns with noise in this underexposed 200 ISO shot, even when shown at 100%.
Of course, I don't advocate underexposure, but this example illustrates easy recovery of what was recorded (by accident) as very dark skin indeed.
Here's the resulting image after adjustment....
I know it's a crap shot - flash shadow, poor composition, terrible background, underexposed, so don't rip it to pieces for that. It's just an example of dealing with very dark skin.
- Messages
- 8,247
- Name
- Terry
- Edit My Images
- Yes
Our eyes can't do it in one either.
They are more like a cine camera taking one picture after another and adjusting the aperture constantly.
They even switch to black and white to increase sensitivity, when there is not enough light .
They are also subject to faults...
focal length
focus adjustment.
lack of transparency
Colour fidelity
astigmatism... to name a few.
as for the rest ..
people Pictures usually come out "about OK" for our own needs.
However if someone is being asked to pay for them , "OK" is not good enough. Then you get hit with what someone else thinks is important.
- Messages
- 18,182
- Name
- Geof
- Edit My Images
- Yes
its always in the eye of the beholder....the final eyeballer
need to be a mind reader ...to some extent...
not everyone likes what they see even if its accurate...hence the spot and soft filter for the modelling scene
go for it
- Messages
- 18,182
- Name
- Geof
- Edit My Images
- Yes
Well, I would usually shoot for the highlights and process for the shadows, and that ought to allow me to pull up shadow detail a little if I need to. So long as my exposure is basically sound, and my ISO not too high, pulling up shadow detail should be no problem.
Here's an example where I did screw up the exposure and I've had to boost the exposure in Lightroom by 1.3 stops and lower the black level from 5 to 1. The flesh tones have come up well and I can't see any concerns with noise in this underexposed 200 ISO shot, even when shown at 100%.
Of course, I don't advocate underexposure, but this example illustrates easy recovery of what was recorded (by accident) as very dark skin indeed.
Here's the resulting image after adjustment....
long live intelligently used equipment and post processing..
- Messages
- 8,247
- Name
- Terry
- Edit My Images
- Yes
The problem with that example, is that the contrast is too high in the corrected version, and even his collar is burnt out, as would be a white dress.
But it does demonstrate that the tonal detail is still in the dark skin.
- Messages
- 18,182
- Name
- Geof
- Edit My Images
- Yes
Uploaded with ImageShack.us
"the masai girl"
to get skin tone nearly right...to my eyes...the highlights got burnt out...
- Messages
- 6,481
- Name
- Tim
- Edit My Images
- No
Well, technically my adjustment is not throwing any pixels into clipping, with nothing on the collar brighter than 98% when sampled in Lightroom, following my exposure boost, but I did note the apparent problem and have since added some highlight recovery to pull down the top end of the highlights a little. I think the issue is that the angle of illumination, which is clearly from my flash, is hitting that part of the collar as flat light and the texture has disappeared entirely, causing the appearance of blown highlights. The collar will take 1.5 stops of exposure increase without clipping, so 1.3 stops should not be a problem. It just looks that way.
Anyway, my post was more about recovery of the dark tones, since the highlight data is all intact and can be positioned wherever I like.
p.s. Is your monitor calibrated accurately? I'm sure it is
)) but a poorly calibrated monitor can mash the 250-255 range of values together and make a pig's ear of the highlights. I've got one laptop that really screws the highlights, even though the histogram indicates no problem.p.p.s. I just downloaded the JPEG file from my post and there are about 10 clipped pixels in the collar, so even in the conversion to JPEG it has not gone tits up.
- Messages
- 8,247
- Name
- Terry
- Edit My Images
- Yes
Uploaded with ImageShack.us
"the masai girl"
to get skin tone nearly right...to my eyes...the highlights got burnt out...
Looks great to me the spectral highlights on the jewellery would burn out anyway. a bit over sharpened though... Picky!!!
- Messages
- 6,481
- Name
- Tim
- Edit My Images
- No
By the look of things, if you had shot that raw, I reckon there's a good chance you might have recovered the sky. Skin tones look good though.
However, the sharpening looks a bit enthusiastic. You've got some wicked sharpening halos going on there.
- Messages
- 18,182
- Name
- Geof
- Edit My Images
- Yes
By the look of things, if you had shot that raw, I reckon there's a good chance you might have recovered the sky. Skin tones look good though.
However, the sharpening looks a bit enthusiastic. You've got some wicked sharpening halos going on there.
:bang:
i know...i have the original on my master hard drive which i was too lazy to plug in and took this off a forum where i put the shot originally...
i must agree when i posted it i thought...what hairy shot
its not raw but jpeg and from a 4 Mp camera..i even ran noiseware over it it was so chunky
cheers
geof