Chriscrawfordphoto
Real Men Shoot Film.
Scenes that are backlit, or strongly side-lit, are very difficult to expose correctly. The high contrast light means that the brightest and darkest parts of the scene might fall outside the dynamic range of your digital camera.
A Reflected Light Meter, like that built in to your camera, will underexpose the scene. The bright light from behind the subject will fool the meter, causing the backlit subject to render very dark, possibly so dark that little or no detail is visible.
An Incident Light Meter is usually the best type of meter for digital work because it ignores the subject and just measures the light that hits the subject. Normally, an incident meter cannot be fooled by very bright or very dark subject matter, as a reflected light meter can.
Backlit and strongly sidelit scenes are an exception. The incident meter is used by standing at the subject position and pointing the meter's white dome back toward the camera lens. The hemispherical dome simulates the three-dimensionality of an average subject; and it can take into account light from straight on, as well as side-lighting, if the lighting angle is not too severe. In a backlit scene, the whole dome is in shade and does not see the light coming from behind. The result is an overexposed photograph.
Solutions
One solution is to use a handheld spotmeter and carefully measure the brightness difference between the darkest and lightest parts of the scene that you want to retain detail in, using those readings to calculate exposure settings that will leave all or most of the scene's brightness range within the dynamic range of your camera's sensor.
This works wonderfully, but is cumbersome and is difficult for beginners as it requires some testing to determine your camera's dynamic range, and experience in knowing what areas to measure in the scene. If you're interested in learning to use a spotmeter, see my Spotmetering Tutorial.
An Easier Way: Incident-Reflected Average
A simpler solution is to take two readings with a handheld meter, a reflected light reading and an incident light reading. Then, you average them!
This technique is fast and easy to learn, and works well for most backlit and sidelit scenes. If the brightness range in the scene is very wide, you may have to use the shadow and highlight recovery sliders in your RAW processing software (like Lightroom or Capture One). This technique works best if you shoot RAW, rather than JPEGs.
This technique was invented in the 1960s by Jack Dunn, a British engineer who did pioneering research into exposure metering in the mid 20th Century. Dunn wrote, with George Wakefield, a book called "Exposure Manual." If you're as big of a meter geek as I am, you want to get hold of this book.
I also have a tutorial video on YouTube showing how I edited a photograph that I made using this exposure technique.
Example #1
Exposure determined with a reflected light meter.
Exposure determined with an incident light meter.
Exposure determined by averaging an incident light reading and a reflected light reading.
The final image after adjusting the shadow and highlight recovery sliders in Lightroom.
As you can see in the finished image above, this technique produced a beautiful rendering that has good detail in the highlights and shadows, despite the very high brightness range in the scene.
It was very simple to do in the field, taking about 30 seconds to make the two readings and average them.
Modern cameras have "Evaluative" metering systems (also known as Matrix Metering) that analyze the scene to compensate automatically for such difficult lighting situations. In my experience, they sometimes work well and often do not.
The incident-reflected average technique has been far more accurate for me.
A Reflected Light Meter, like that built in to your camera, will underexpose the scene. The bright light from behind the subject will fool the meter, causing the backlit subject to render very dark, possibly so dark that little or no detail is visible.
An Incident Light Meter is usually the best type of meter for digital work because it ignores the subject and just measures the light that hits the subject. Normally, an incident meter cannot be fooled by very bright or very dark subject matter, as a reflected light meter can.
Backlit and strongly sidelit scenes are an exception. The incident meter is used by standing at the subject position and pointing the meter's white dome back toward the camera lens. The hemispherical dome simulates the three-dimensionality of an average subject; and it can take into account light from straight on, as well as side-lighting, if the lighting angle is not too severe. In a backlit scene, the whole dome is in shade and does not see the light coming from behind. The result is an overexposed photograph.
Solutions
One solution is to use a handheld spotmeter and carefully measure the brightness difference between the darkest and lightest parts of the scene that you want to retain detail in, using those readings to calculate exposure settings that will leave all or most of the scene's brightness range within the dynamic range of your camera's sensor.
This works wonderfully, but is cumbersome and is difficult for beginners as it requires some testing to determine your camera's dynamic range, and experience in knowing what areas to measure in the scene. If you're interested in learning to use a spotmeter, see my Spotmetering Tutorial.
An Easier Way: Incident-Reflected Average
A simpler solution is to take two readings with a handheld meter, a reflected light reading and an incident light reading. Then, you average them!
This technique is fast and easy to learn, and works well for most backlit and sidelit scenes. If the brightness range in the scene is very wide, you may have to use the shadow and highlight recovery sliders in your RAW processing software (like Lightroom or Capture One). This technique works best if you shoot RAW, rather than JPEGs.
This technique was invented in the 1960s by Jack Dunn, a British engineer who did pioneering research into exposure metering in the mid 20th Century. Dunn wrote, with George Wakefield, a book called "Exposure Manual." If you're as big of a meter geek as I am, you want to get hold of this book.
I also have a tutorial video on YouTube showing how I edited a photograph that I made using this exposure technique.
Example #1
Exposure determined with a reflected light meter.
Exposure determined with an incident light meter.
Exposure determined by averaging an incident light reading and a reflected light reading.
The final image after adjusting the shadow and highlight recovery sliders in Lightroom.
As you can see in the finished image above, this technique produced a beautiful rendering that has good detail in the highlights and shadows, despite the very high brightness range in the scene.
It was very simple to do in the field, taking about 30 seconds to make the two readings and average them.
Modern cameras have "Evaluative" metering systems (also known as Matrix Metering) that analyze the scene to compensate automatically for such difficult lighting situations. In my experience, they sometimes work well and often do not.
The incident-reflected average technique has been far more accurate for me.
Chriscrawfordphoto
Real Men Shoot Film.
Example #2
Metered with a reflected light meter.
The very bright backlit sky fooled the meter into underexposing the photo.
Metered with an incident light meter.
The sky is blown out, and even the sign is overexposed, since the shadowed side of the sign facing the camera has been lightened too much.
For this exposure, I took an incident light reading standing in front of the sign, then walked back to the camera and did a reflected light reading. I averaged the two readings, and this was the result.
Because of the very wide brightness difference between the foreground and the very bright sky, the average exposure gives dark tones that are slightly too dark and lights that are still a little too light.
The same exposure, but this time I used the Highlight and Shadow Recovery sliders in Adobe Lightroom to open up the foreground shadows and bring down the bright clouds.
Midtones seem a little too dark, though.
The final image. I took the version with the adjusted highlight recovery and shadow recovery sliders and applied a curves adjustment to the image to lighten the midtones.
Now it looks great! The clouds are bright but still have full detail, while the sign and foreground areas look like they should.
Metered with a reflected light meter.
The very bright backlit sky fooled the meter into underexposing the photo.
Metered with an incident light meter.
The sky is blown out, and even the sign is overexposed, since the shadowed side of the sign facing the camera has been lightened too much.
For this exposure, I took an incident light reading standing in front of the sign, then walked back to the camera and did a reflected light reading. I averaged the two readings, and this was the result.
Because of the very wide brightness difference between the foreground and the very bright sky, the average exposure gives dark tones that are slightly too dark and lights that are still a little too light.
The same exposure, but this time I used the Highlight and Shadow Recovery sliders in Adobe Lightroom to open up the foreground shadows and bring down the bright clouds.
Midtones seem a little too dark, though.
The final image. I took the version with the adjusted highlight recovery and shadow recovery sliders and applied a curves adjustment to the image to lighten the midtones.
Now it looks great! The clouds are bright but still have full detail, while the sign and foreground areas look like they should.
Chriscrawfordphoto
Real Men Shoot Film.
Example #3
Exposure determined with a reflected light meter.
Exposure determined with an incident light meter.
Exposure determined by averaging the reflected light and incident light meter readings.
This is a straight, unmanipulated rendering. The only adjustments done to the RAW file in Lightroom were sharpening and white balance.
The final image after adjusting the shadow recovery slider in Lightroom. I might actually prefer the one with the darker shadows!
This was a very difficult scene because of the bright snow, and the fact that the sun was directly in front of me. The large amount of lens flare visible in the scene is an unavoidable consequence of shooting directly into the bright sun.
The challenge was to maintain detail in the snow, while still keeping it at a value that looks white. Too little exposure, and the snow will go gray, and too much will cause loss of texture and detail. At the same time, we want to have detail visible in the shadows.
The Averaged exposure actually works very well, without the need to do any further adjustments. If you want more shadow detail, pulling up the Shadow Recovery slider in lightroom produced the final result shown above.
Exposure determined with a reflected light meter.
Exposure determined with an incident light meter.
Exposure determined by averaging the reflected light and incident light meter readings.
This is a straight, unmanipulated rendering. The only adjustments done to the RAW file in Lightroom were sharpening and white balance.
The final image after adjusting the shadow recovery slider in Lightroom. I might actually prefer the one with the darker shadows!
This was a very difficult scene because of the bright snow, and the fact that the sun was directly in front of me. The large amount of lens flare visible in the scene is an unavoidable consequence of shooting directly into the bright sun.
The challenge was to maintain detail in the snow, while still keeping it at a value that looks white. Too little exposure, and the snow will go gray, and too much will cause loss of texture and detail. At the same time, we want to have detail visible in the shadows.
The Averaged exposure actually works very well, without the need to do any further adjustments. If you want more shadow detail, pulling up the Shadow Recovery slider in lightroom produced the final result shown above.
Chriscrawfordphoto
Real Men Shoot Film.
Equipment Requirements
Ideally, for this technique, you'll need a handheld exposure meter capable of both incident light metering and reflected light metering. The reflected light metering should not be a spotmeter, it needs to be a wide-area reflected light meter (30 to 40 degree measuring angle).
This rules out many of the recent Sekonic meters, like the L-508, L-608, L-588, L-758, and L-858. The aforementioned meters have a one degree spotmeter as the only choice for reflected light metering.
Minolta meters, like the Flash Meter IV, Flash Meter V, Autometer III, Autometer IVF, and Autometer VF can be used if you have the Reflected Light Attachment (the flat disk with a hole in it, not the spot viewfinders). That said, they're rather cumbersome to use in the field because switching from incident to reflected metering requires you to remove the white incident metering dome and replace it with the reflected light attachment. This is a bothersome two-handed job and if you're not careful you will eventually drop or lose either the dome or the reflected attachment, and they are expensive to replace.
My favorite meter for this technique is the long-discontinued Gossen Ultra Pro (Mastersix outside the USA). The incident dome is made to simply slide out of the way to do reflected readings. You can do it one-handed, the dome cannot fall off, and there is no need for a separate attachment for doing the reflected light reading.
Gossen has made many meters utilizing this simple method of switching from incident to reflected, including the old Luna Pro, and the later Luna Pro SBC, Luna Pro F, and Multi-Pro. I like the Ultra-Pro because it is a digital meter with a built-in averaging function that saves me from having to do math. It is relatively uncommon meter because back in the 80s and 90s when it was made, it was Gossen's top of the line meter, and it cost nearly $1000 when it was finally discontinued. I do see them on Ebay for around $100 fairly often.
If you have a good incident meter that doesn't do wide-area reflected readings, and you don't want to buy one, you can use your camera's built in meter for the reflected reading. You'll have to do the averaging math in your head. Make sure you do not have the camera set to use Matrix or Evaluative metering; the meter should be set to do centerweighted metering. That will give results similar to a handheld reflected light meter.
More Examples
Here are some more more photographs that I made using the Incident-Reflected Average metering technique.
Ideally, for this technique, you'll need a handheld exposure meter capable of both incident light metering and reflected light metering. The reflected light metering should not be a spotmeter, it needs to be a wide-area reflected light meter (30 to 40 degree measuring angle).
This rules out many of the recent Sekonic meters, like the L-508, L-608, L-588, L-758, and L-858. The aforementioned meters have a one degree spotmeter as the only choice for reflected light metering.
Minolta meters, like the Flash Meter IV, Flash Meter V, Autometer III, Autometer IVF, and Autometer VF can be used if you have the Reflected Light Attachment (the flat disk with a hole in it, not the spot viewfinders). That said, they're rather cumbersome to use in the field because switching from incident to reflected metering requires you to remove the white incident metering dome and replace it with the reflected light attachment. This is a bothersome two-handed job and if you're not careful you will eventually drop or lose either the dome or the reflected attachment, and they are expensive to replace.
My favorite meter for this technique is the long-discontinued Gossen Ultra Pro (Mastersix outside the USA). The incident dome is made to simply slide out of the way to do reflected readings. You can do it one-handed, the dome cannot fall off, and there is no need for a separate attachment for doing the reflected light reading.
Gossen has made many meters utilizing this simple method of switching from incident to reflected, including the old Luna Pro, and the later Luna Pro SBC, Luna Pro F, and Multi-Pro. I like the Ultra-Pro because it is a digital meter with a built-in averaging function that saves me from having to do math. It is relatively uncommon meter because back in the 80s and 90s when it was made, it was Gossen's top of the line meter, and it cost nearly $1000 when it was finally discontinued. I do see them on Ebay for around $100 fairly often.
If you have a good incident meter that doesn't do wide-area reflected readings, and you don't want to buy one, you can use your camera's built in meter for the reflected reading. You'll have to do the averaging math in your head. Make sure you do not have the camera set to use Matrix or Evaluative metering; the meter should be set to do centerweighted metering. That will give results similar to a handheld reflected light meter.
More Examples
Here are some more more photographs that I made using the Incident-Reflected Average metering technique.
seany65
Well-known
Thanks for this Chris. I'd been wondering about this recently myself. I'd already used incident and reflected readings separately on various cameras and not got good results as the film couldn't cope with the contrast, and with it being winter and the sun low and usually on the other side of my 'test subjects' (I'm testing all the cameras which I've bought over the year) I've had to wait until the sky was bright but overcast to limit the contrast but get hand-holdable speeds.
However, I was wondering what could be done if someone doesn't have access to film or print scanners to 'even out the contrast'/recover the shadows?
For instance, The averaged 'sandpoint' sign photo: Would it be an idea to do the averaging and then add 1/2 a stop, or would that make little difference?
However, I was wondering what could be done if someone doesn't have access to film or print scanners to 'even out the contrast'/recover the shadows?
For instance, The averaged 'sandpoint' sign photo: Would it be an idea to do the averaging and then add 1/2 a stop, or would that make little difference?
Richard G
Veteran
This is great. I had such a scene yesterday morning and I use a Gossen DigiPro F. The incident dome needs to be screwed off for the reflected reading. It works and is seldom required.
MCTuomey
Veteran
Much more precise than my "go-to" open up 2 stop compensation method for backlit subjects, thanks Chris!
Chriscrawfordphoto
Real Men Shoot Film.
With digital, you wouldn't want to add any exposure because you would be at risk of losing highlight detail.
Same with slide film, though in that case you'd have to deal with the problem of too much contrast causing blocked up shadows in some photos. That's always been something slide shooters have had to deal with when shooting in any kind of contrasty light, not just backlit scenes.
I haven't tried this with black and white or color negative films. The tutorial was written primarily for digital, because the technique gives a highlight-biased exposure as required for digital. Because negative films require a shadow-biased exposure (meaning you need to give enough exposure to maintain shadow detail, even if it overexposes the highlights), then it might be a good idea to give more exposure. How much would depend on the amount of contrast in the scene.
I think for negative films, you're better off using a spotmeter and setting the shadow tones where they need to be.
Spanik
Well-known
When you say "average" do you just "sum and divide by two" or is this more elaborate? Could you give some examples with numbers?
Very interesting series! Thanks.
Very interesting series! Thanks.
Chriscrawfordphoto
Real Men Shoot Film.
If one reading says 1/125 at f16 and the other says 1/125 at f8, then you shoot at 1/125 at f11.
If one reading is 1/125 at f5.6 and the other is 1/125 at f16, then you shoot at 1/125 at halfway between f8 and f11.
Most modern meters with digital readouts can memorize and average readings for you so you do not have to do the math. Basically you want the midpoint between the two readings.
creenus
Established
Thanks for doing this, Chris. Extremely useful stuff. This will help lots of folks, including a reminder for me. =)
I shoot Ilford HP5 with an M2 and Gossen DigiSix, and average the incident/reflected reading quite often. Depends on the subject and its light value/tone, of course.
I shoot Ilford HP5 with an M2 and Gossen DigiSix, and average the incident/reflected reading quite often. Depends on the subject and its light value/tone, of course.
Chriscrawfordphoto
Real Men Shoot Film.
Steve M.
Veteran
This is why I love my Nikon n8008s. It has a spot meter with AE lock. I simply lock exposure on what part of a scene I want properly exposed. It's impossible to properly expose all parts of a shot if you have too much spread in light values, so as long as my subject is exposed properly I let the rest get blown out or muddied.
Steve M.
Veteran
Duplicate post.
Chriscrawfordphoto
Real Men Shoot Film.
You should shoot in manual mode with your Nikon n8008s if you use the spotmeter.
In manual, you'll see a graph in the viewfinder and on the top deck that shows where on the exposure scale the thing you have the spotmeter pointed at will fall at the aperture and shutter speed you have set.
It is very easy to meter something and peg it to the correct place on the exposure scale. Photographing a middle gray subject? Adjust your settings to put the exposure readout on the Zero at the center of the graph. White subject? Adjust your settings so that the exposure readout says +2 stops.
This kind of readout was pioneered by the Olympus OM-4 if I remember correctly, and was later found on just about every AF 35mm SLR and now its found in digital SLRs too.
leica M2 fan
Veteran
Well done Chris. Excellent and valuable information here to help
everyone no matter how advanced anyone is, they can still pick up
something. I've enjoyed it immensely. Thanks.
everyone no matter how advanced anyone is, they can still pick up
something. I've enjoyed it immensely. Thanks.
ChrisPlatt
Thread Killer
Interesting method and no doubt accurate, but anything involving incident metering is just too
time-consuming for fast shooting, which is the main reason I choose to use miniature cameras.
IMO experienced users can get similar results consistently simply using their cameras built-in
reflected light meter, intelligently used, without walking back and forth to set up the shot.
Chris
time-consuming for fast shooting, which is the main reason I choose to use miniature cameras.
IMO experienced users can get similar results consistently simply using their cameras built-in
reflected light meter, intelligently used, without walking back and forth to set up the shot.
Chris