see also:

• exposure
• Zone System for reflective light metering
• spot meter modes of Olympus cameras

• most digital cameras display RGB histograms after the photo has been taken - that is a graphical display of how much of the image is recorded at each brightness level for each of the 3 digital color channels Red, Green & Blue.

• the main exposure problem in digital cameras is the loss of detail in over-exposed regions which cannot be retrieved by later processing (although if you shoot in RAW mode, some of it can be retrieved).

• thus, most digital photographers prefer to check the histograms to ensure none of the main image peak goes past the right side of the histogram in any of the 3 channels - although, inevitably with high contrast situations where there are 2 peaks, the right peak for very bright sky, reflections or the light source itself, will usually be to the right of the limit of the histogram and is thus ignored, while the right edge of the left peak representing your subject does not reach the right limit.

• some cameras allow a histogram in Live View - see live histogram in Olympus cameras

• light values is a way of estimating or determining the amount of light ambient on a scene which then enables one to predict what combination of aperture, shutter speed and ISO one will need for a given scene by using the S+L=A+T exposure system.

• why bother?

• let's say you want to take photos of paintings in the Louvre and you have only the one day in your life to do it.

• given that flash photography will be practically useless, and tripods not feasible, one will generally need to rely on hand-held photography, so will your camera/lens combination cut it in taking sharp photos without camera shake?

• you could just buy the most expensive camera and hope it will do, but the alternative is if you know the light value that you are likely to be working with (which I give below), then you can determine your requirements more accurately.

• given my calculations for paintings in the Louvre of -1 to +2 light value, this will mean that you can use ISO 200, f/2.8 and shutter speed will need to be 1/4 to 1/15th second which is too long hand held unless you have an image stabiliser or you use a higher ISO or wider aperture. Thus most consumer cameras would not be adequate to get good results with low noise.

• this same principle can be applied to a multitude of situations such as a sports event or model photography using ambient light from the sky.

• a professional photographer using a Canon 1D Mark III with a Canon EF 70-200mm f/2.8 L IS lens at 200mm:
  • can usually get away with using f/2.8 at 1/50th sec handheld with IS on (ie. A+T = 3+6 = 9) and thus the light levels he can shoot hand held will be determined by the ISO he selects:
    • ISO 100 ⇒ LV = 4 eg. sports ground with night lighting
    • ISO 400 ⇒ LV = 2
    • ISO 800 ⇒ LV = 1 eg. average living room lighting; heavy cloud,
    • ISO 1600 ⇒ LV = zero
    • ISO 3200 ⇒ LV = minus 1 eg. twilight
    • ISO 6400 ⇒ LV = minus 2
  • conversely, let's say he is shooting a sports ground under night lights with a LV of 4 and shooting at f/2.8, what is his fastest shutter speed to stop the action:
    • ISO 800 ⇒ EV = 8+4 = 12 ⇒ T = 12-3 = 9 which equates to 1/500th sec
    • ISO 1600 ⇒ EV = 9+4 = 13 ⇒ T = 13-3 = 10 which equates to 1/1000th sec
    • ISO 3200 ⇒ EV = 10+4 = 14 ⇒ T = 14-3 = 11 which equates to 1/2000th sec
    • ISO 6400 ⇒ EV = 11+4 = 15 ⇒ T = 15-3 = 12 which equates to 1/4000th sec
  • a portrait photographer using a Canon 1D Mark III with a Canon EF 85mm f/1.2 at f/1.2 for its DOF:
    • full sun at ISO 100: LV = 10, S = 5, A = 0.5 ⇒ T = 10+5-0.5 = 14.5 which equates to 1/20000th sec but as min. shutter is 1/8000th, this will be 1.5 stops over-exposed, thus need a ND or polariser filter but will not be able to use fill-in flash at that shutter speed.

• the difference of 1 for a light value equates to 1 f-stop difference.
• if you double the distance of a light source, this will result in the light reducing to 1/4 which equates to 2 f-stops less or 2LV less.
• camera manufacturers often give exposure range for their light metering, auto focus capability, etc in EV ranges, to determine exposure value (EV) from LV, just add 5 to LV assuming you are using ISO value of 100 for determining EV.

• An EV (exposure value) is equivalent to a combination of f/ratio, shutter speed at a given ISO value (usually 100ISO), with a value of 0 being 1 second at f/1.0., thus EV = S+L = A+T.
• The S+L=A+T exposure system uses this concept, but expands on it to create a value for each of the 4 variables of Speed (film speed), Light intensity, Aperture & Time (shutter speed) such that an integer difference in any of the values equates to 1 f stop difference, and the sum of S+L equals the sum of A+T which equals the EV at ISO 100:
  • Speed: ISO film rating of 100 has a value of 5, each time you double the ISO value, you increase this value by 1, conversely, halving it drops the value by 1, thus ISO 1600 = 9. (S = log2 (0.3 x ISO))
    • ISO 1600 = 9
    • ISO 800 = 8
    • ISO 400 = 7
    • ISO 200 = 6
    • ISO 100 = 5
  • Light: bright sunlight = 10, for more see below (I use abbreviation LV for this) also LV = log2 (brightness in foot-Lamberts)
    • see examples below
  • Aperture: f/4 = 4, each f-stop change gives an integer change in the value, thus f/8 = 6 as it is 2 stops different to f/4. A = log2 (f-stop^2)
    • f/16 = 8
    • f/11 = 7
    • f/8 = 6
    • f/5.6 = 5
    • f/4 = 4
    • f/2.8 = 3
    • f/2.0 = 2
    • f/1.4 = 1
  • Time: 1sec = 0, each time you halve the shutter speed, you decrease exposure by 1 f-stop, and you add an integer to the T value, so that 1/125th = 7 and 1/1000th = 10. T = log2 (1/exposure time in seconds).
    • 1/1000th = 10
    • 1/500th = 9
    • 1/250th = 8
    • 1/125th = 7
    • 1/60th = 6
    • 1/30th = 5
    • 1/15th = 4
    • 1/8th = 3
    • 1/4 = 2
    • 1/2 = 1
    • 1sec = 0
    • 2 secs = -1

• THUS, bright sunlight at ISO 100 = 5 + 10 = 15EV, and thus any combination of A & T that sum to 15 can be used such as f/5.6 and 1/1000th sec.

• of course you will need to adjust settings to take into account loss of light from any photographic filters used, and if using photographic film in a digital age with long exposures, adjust for reciprocity failure (see below).

• add 5 to the light value for EV at ISO 100

• see Zone System for reflective light metering

• the above S+L=A+T exposure system only applies to light sources that have a duration longer than the shutter speeds used.
• electronic flash units have a duration of 1/300th sec to 1/20,000th sec depending on the unit and output selected, thus the camera shutter speed selected only determines ambient lighting effect and does not affect the flash lighting which is dependent on aperture and ISO, hence the use of the guide number (GN) system for determining flash output.
• lightning also is very brief and its brightness is dependent on many factors including its intensity, distance and the amount of cloud or rain between the lightning and your camera. For lightning a few kilometres away, consider using f stop of f/5.6 to f/8 at ISO 50 and then set shutter speed according to desired ambient lighting effects needed (eg. for night streetscapes try 8-30 secs). A long shutter speed is usually needed to maximise the chance a lightning strike will occur during the exposure, unless you have a light-sensitive shutter trigger that takes the exposure immediately lightning occurs.

• transparency film photographers often set the transparent film ISO to be 1/3rd of a stop lower than stated to improve colour saturation and prevent highlights being blown out by its narrow dynamic range.
• B&W film photographers often shoot at a different ISO (eg. when they expose for the shadows) and use developing to modify the contrast range.
• digital cameras:
  • most digital cameras ISO values are around 0.5EV lower than the stated value, although Olympus tends to be around 1EV lower than stated - this can be an issue if using external light meters.
  • in general, it is best to exposure so that the histogram just reaches the far right and so highlights are not blown out.
  • you have much more latitude in lowering the exposure value in RAW development than you have in increasing the exposure which tends to add noise to the shadows very rapidly BUT as long as you haven't blown the highlights as these areas will always remain white without any detail.
  • if you capture using jpeg mode, you should aim for the exposure that looks best without any adjustment needed.
  • if you capture using RAW mode you should aim for the exposure which will capture the greatest amount of useful capture information as you can always use contrast control to adjust the mid-tones later.
  • shadow detail and highlight detail are extremely important in fine full-toned prints, to have black shadows with no details or blown out highlights makes the image look amateurish.

• illuminance onto a surface (E) = luminous flux / surface area, (unit is lumen per square meter = meter-candle) for a point source:
  • E = (I/s^2)cos a
  • where I = luminous intensity, s = distance from source, a = angle of source from the perpendicular of the surface
  • hence, the intensity of sunlight is less in winter than summer as the angle from the zenith is greater, compounded by some atmospheric extinction as there is more atmosphere through which it must pass.

• 1/10th sec ⇒ + 0.5 stop
• 1 sec ⇒ + 1 stop
• 10 sec ⇒ + 1.5-2 stops
• 120 sec ⇒ + 2.5-3 stops

• digital sensors, like transparency film can only deal with 5-6 f-stops of dynamic range, thus for high contrast scenes such as landscapes when one wishes to show detail in the clouds/sky as well as in the scene shadows then one must resort to other means of obtaining it such as:
  • use a higher dynamic range film such as B&W (expose for the shadows, adjust development for the highlight range needed) or color negative film
    • see http://www.erik-krause.de/index.htm?./cn_hdr/index.htm - expose for the shadows then adjust the scans in PS to correct the highlights
  • use a higher dynamic range sensor such as medium format sensors (but these are very expensive) and save as HDR file.
  • use a gradient filter on the lens - but this adds to lens flare and may not match the geometry of the highlight region
    • see http://www.singh-ray.com/neutral-density-filters/
    • see Gary Maico's pics - US seascapes using Singh-Ray variND filter (~$US399), ND grad filter, Clor Grad filter, Blue & Gold Polariser.
  • use a polarising filter to darken the sky - but only if sky is approx. 90deg to the sun
    • see http://www.vinberg.nu/pages/polarizers.htm
    • see Singh-Ray filters
  • use infra-red filters to darken the sky - but this then creates an infra-red effect
  • contrast mask blend a single image in Photoshop:
    • see http://www.luminous-landscape.com/tutorials/contrast_masking.shtml
    • see http://www.erik-krause.de/index.htm?./contrast/index.htm
    • see http://www.vinberg.nu/pages/contrast_masking.htm - effects of altering Gaussian blur
  • use HDR techniques by taking 3-5 bracketed exposures 2-3 f-stops apart and blend images in Photoshop:
    • may not be suitable if subject moves eg. human content
    • see http://www.luminous-landscape.com/tutorials/digital-blending.shtml shows 3 options:
      • painted mask technique - need to mask area by hand
      • Gaussian blur layer mask technique - here is a snippet:
        • Starting with having done the pasting of the dark image on the light one, add a Layer Mask. This is done by clicking on the second icon on the lower left of the Layers palette. You will now see a white rectangle next to the image on the Layer 1 layer.
        • Click on the background layer (the lighter image) on the palette and the press CTRL-A to select the whole image. Press CTRL-C, copying it to the clipboard. Now hold down the ALT key (Option on the Mac) and click on the white mask rectangle on the Layer 1 palette.
        • The whole image will now turn white. Next, press CTRL-V to paste the contents of the clipboard onto the white mask. You will now see a B&W mask image. With the B&W mask displayed go to Filter / Blur / Gaussian Blur and set the Radius to about 40 pixels. Click on the Background Layer and then flatten the layers.
      • Fred Miranda's Dynamic Range PS action
        • see http://www.erik-krause.de/index.htm?./blending/
      • also Photoshop CS2 has a new dedicated HDR method