Lest there be wailing and gnashing of teeth about the value of metering let me add I first use a meter to set the lights relative to each other and for exposure to get them in the ballpark, but use what the camera actually records and my eyes as the final arbiter. By doing testing I can interpret the results of overexposure warning in the camera playback JPG and know how the RAW file will be exposed.
When setting lights I put a target where the face will be in the portrait which has a full range of tone and most importantly a textured white highlight value represented by the white towel. If the texture starts to disappear its a simple visual clue the file is overexposed. The towel will also start to black out in camera playback set to show the clipping warning, making correct exposure of the highlights pretty much a no-brainer.
But its the fill which is actually the foundation the lighting pattern needs to be built on for the simple reason the key light overlaps it and is necessary to reveal shadow detail. Start with just your fill light on and raise it to the point the desired detail is seen in the darkest parts of the foreground subject. Then turn on the key light and raise it until the brightest parts of the towel barely clip, then back off until the clipping stops.
The set-up below used fill and key light in a centered pattern and a hair light. I wanted perceptual separation of the brightest highlights from the background, but without anything clipping. I set the back lit parts of the towel so they were just below the point of clipping, then adjusted the key light to it perceptually slightly darker to maintain tonal separation.
The greatest illusion of 3D is created when there is contrast between rim lit parts of the subject and the background. So I set the background so it matched the front of the target but there was still visual separation and the illusion of shape created with the rim light.
I brought the background lights up to the point the background started to clip in the playback, then backed off until I had the desired visual separation.
The perception of \"white\" is a relative thing and the viewer almost immediately tunes out the background anyway so there\'s no reason to \"nuke\" a white background. On the contrary if the background was taken all the way to clipping the separation of the rim light and the illusion of 3D it creates would disappear, and relative to the nuked background the foreground would seem darker. That\'s the thing about human perception to grasp when setting lights. The same lighting on a face will look different depending on the perceptual influence of the clothing and background. I shot the photo below on the white background (with a single flash on a bracket) but then Photoshopped clothing and background to illustrate this:
It\'s more important to understand how the brain of the viewer will interpret the scene than making the background 255 \"knock-out\" white or setting ratios \"by the numbers\". A 3:1 highlight/shadow ratio which looks average on a dark background perceptually will look heavier and darker on a white one.
Remove target, insert subject, capture well exposed image...
When setting exposure in the camera I try to keep everything except mirror-like specular reflections below the point of clipping. In the digital realm 255.255.255 = pure light. A solid white object reflects about 98% of the light and will measure 250 with the eye dropper tool. Something white with texture like the towel will average around 230-240 depending how the light hits it. 240 is also the value of the red channel in a well exposed skin tone. Red is the brightest channel in skin tones and setting exposure via the towel in the clipping warning also ensures optimally exposed skin tones.
Click the WWW button for my tutorials on exposure, metering, etc.
Lest there be wailing and gnashing of teeth about the value of metering let me add I first use a meter to set the lights relative to each other and for exposure to get them in the ballpark, but use what the camera actually records and my eyes as the final arbiter. By doing testing I can interpret the results of overexposure warning in the camera playback JPG and know how the RAW file will be exposed.
When setting lights I put a target where the face will be in the portrait which has a full range of tone and most importantly a textured white highlight value represented by the white towel. If the texture starts to disappear its a simple visual clue the file is overexposed. The towel will also start to black out in camera playback set to show the clipping warning, making correct exposure of the highlights pretty much a no-brainer.
But its the fill which is actually the foundation the lighting pattern needs to be built on for the simple reason the key light overlaps it and is necessary to reveal shadow detail. Start with just your fill light on and raise it to the point the desired detail is seen in the darkest parts of the foreground subject. Then turn on the key light and raise it until the brightest parts of the towel barely clip, then back off until the clipping stops.
The set-up below used fill and key light in a centered pattern and a hair light. I wanted perceptual separation of the brightest highlights from the background, but without anything clipping. I set the back lit parts of the towel so they were just below the point of clipping, then adjusted the key light to it perceptually slightly darker to maintain tonal separation.
The greatest illusion of 3D is created when there is contrast between rim lit parts of the subject and the background. So I set the background so it matched the front of the target but there was still visual separation and the illusion of shape created with the rim light.
I brought the background lights up to the point the background started to clip in the playback, then backed off until I had the desired visual separation.
The perception of \"white\" is a relative thing and the viewer almost immediately tunes out the background anyway so there\'s no reason to \"nuke\" a white background. On the contrary if the background was taken all the way to clipping the separation of the rim light and the illusion of 3D it creates would disappear, and relative to the nuked background the foreground would seem darker. That\'s the thing about human perception to grasp when setting lights. The same lighting on a face will look different depending on the perceptual influence of the clothing and background. I shot the photo below on the white background (with a single flash on a bracket) but then Photoshopped clothing and background to illustrate this:
It\'s more important to understand how the brain of the viewer will interpret the scene than making the background 255 \"knock-out\" white or setting ratios \"by the numbers\". A 3:1 highlight/shadow ratio which looks average on a dark background perceptually will look heavier and darker on a white one.
Remove target, insert subject, capture well exposed image...
When setting exposure in the camera I try to keep everything except mirror-like specular reflections below the point of clipping. In the digital realm 255.255.255 = pure light. A solid white object reflects about 98% of the light and will measure 250 with the eye dropper tool. Something white with texture like the towel will average around 230-240 depending how the light hits it. 240 is also the value of the red channel in a well exposed skin tone. Red is the brightest channel in skin tones and setting exposure via the towel in the clipping warning also ensures optimally exposed skin tones.
Click the WWW button for my tutorials on exposure, metering, etc.
Lest there be wailing and gnashing of teeth about the value of metering let me add I first use a meter to set the lights relative to each other and for exposure to get them in the ballpark, but use what the camera actually records and my eyes as the final arbiter. By doing testing I can interpret the results of overexposure warning in the camera playback JPG and know how the RAW file will be exposed.
When setting lights I put a target where the face will be in the portrait which has a full range of tone and most importantly a textured white highlight value represented by the white towel. If the texture starts to disappear its a simple visual clue the file is overexposed. The towel will also start to black out in camera playback set to show the clipping warning, making correct exposure of the highlights pretty much a no-brainer.
But its the fill which is actually the foundation the lighting pattern needs to be built on for the simple reason the key light overlaps it and is necessary to reveal shadow detail. Start with just your fill light on and raise it to the point the desired detail is seen in the darkest parts of the foreground subject. Then turn on the key light and raise it until the brightest parts of the towel barely clip, then back off until the clipping stops.
The set-up below used fill and key light in a centered pattern and a hair light. I wanted perceptual separation of the brightest highlights from the background, but without anything clipping. I set the back lit parts of the towel so they were just below the point of clipping, then adjusted the key light to it perceptually slightly darker to maintain tonal separation.
The greatest illusion of 3D is created when there is contrast between rim lit parts of the subject and the background. So I set the background so it matched the front of the target but there was still visual separation and the illusion of shape created with the rim light.
I brought the background lights up to the point the background started to clip in the playback, then backed off until I had the desired visual separation.
The perception of \"white\" is a relative thing and the viewer almost immediately tunes out the background anyway so there\'s no reason to \"nuke\" a white background. On the contrary if the background was taken all the way to clipping the separation of the rim light and the illusion of 3D it creates would disappear, and relative to the nuked background the foreground would seem darker. That\'s the think about human perception to grasp when setting lights. It\'s more important to understand how the brain of the viewer will interpret the scene than making the background 255 \"knock-out\" white.
Remove target, insert subject, capture well exposed image...
When setting exposure in the camera I try to keep everything except mirror-like specular reflections below the point of clipping. In the digital realm 255.255.255 = pure light. A solid white object reflects about 98% of the light and will measure 250 with the eye dropper tool. Something white with texture like the towel will average around 230-240 depending how the light hits it. 240 is also the value of the red channel in a well exposed skin tone. Red is the brightest channel in skin tones and setting exposure via the towel in the clipping warning also ensures optimally exposed skin tones.
Click the WWW button for my tutorials on exposure, metering, etc.
Chuck
Feb 24, 2009 at 10:57 PM
Previous versions of cgardner's message #6763494 « Setting hair light and kicker exposures? »
