Posts tagged ISO
Calculating the dynamic range of your digital camera.

Why do this test? Years will go by where you guess what your camera's exposure latitude may be, you may wonder why in some shots the skin tone has gone grey. You may want to know if you can shoot that scene at midday, or what those night shots will come out like and what different ISOs do to the detail and colour of a scene.

In doing this test you will quickly learn the quirks of your camera, you will gain a better undstanding of exposure and light. Eventually these numbers will become a part of how you work with your camera, you will be able to engineer shots that are much richer in colour and detail. You will know what you and your camera gear are capable of.

It will take about 40 mins the first time, if you choose to run the test at various ISOs, it will take about 20 minutes each test.

Why can't you trust the histogram on your camera? It is based on your in-camera processing, you know you can pull more out in your raw converter. The camera's histogram is only a guide, as is the highlight warning tool. Those flashing highlights are only an indication of what you have captured. Careful metering based on an understanding of what your camera is capable of is the only way to get perfect exposure for any scene.

What you need.

  • Camera with manual function, this test is for digital cameras, but same principals can be used for film.
  • Set the camera's metering to spot rather than center weighted or matrix as you will need to meter off a specific part of the scene.
  • Tripod or other method to lock the camera position.
  • Consistent light source.
  • Pure white object that is not too reflective, textured white fabric is good.
  • Something to cast a deep shadow.
  • Colorful object that has both pastels and strong colour, include a skin-like tone and other pastels as well as some saturated colour.
  • Grey card for metering and white balancing.
  • Good quality calibrated monitor to evaluate the results.
  • Photoshop or Lightroom or any program that can display digital images accurately.

This test is partly subjective. You need to decide what you would accept as a white with detail and how much shadow noise is acceptable.

Ideally the test camera will be capturing raw files, while the test can be performed with jpegs, you will be sacrificing dynamic range shooting jpegs. You may also want to consider ensuring your camera is set to the maximum bit depth, It is a little known fact that the Nikon D700 ships in 12bit mode when it is capable of 14bit mode! That is a strange decision by Nikon as those extra bits contain useful data.

How a light meter works. It is important you understand how a light meter works. They are simple tools, all they do is calculate the exposure required to make what they are pointed at a mid-toned density (12% grey to be precise). Therefore, metering off pure white will make that white expose as a mid grey. Metering off a black will make that black a mid-grey. So in proper use, you only want to meter off a mid tone to get correct exposure. A white in the scene needs to be over exposed, and a black, under exposed.

The setup. Set your light close to your subject so you get a fairly dramatic subject with shadows and highlights. You want enough shadow and highlight to meter off accurately. Ensure the setup fills the viewfinder. Ensure the camera is focussed and it will pay to leave the focus in manual mode so it doesn't change.

White balance the camera to your lighting (do not choose auto white balance). Set the camera on manual. The camera can be moved for the metering otherwise, everything must be locked down. To make the test easier to evaluate, you can vary the framing for each of the three tests slightly, so you can tell one from the other easily.

Set the camera's ISO to a typical setting. This test should be repeated at both high and low ISO to see how ISO affects dynamic range with your camera. You may also want to test with different lenses and filters if you have the time.

If you want to compare several cameras in a truly scientific manner, the test must be performed identically each time with each camera, so be precise where you meter off and do not change the set up between tests.

Highlight test. Point the camera in spot metering mode at the whitest part of the subject. If the white material chosen is shiny (highly reflective), it may pay you to avoid the specular highlight as this should be blown our as it is a direct reflection of the light source. You are tying to meter off the area where you would want white with detail.

The camera will be in manual mode, choose a high-ish f-stop and set the shutter speed so the exposure meter indicates correct exposure. Then adjust the fstop to the lower numbers (not adjusting shutter speed) as you gradually over-expose each frame, letting in more light. Ideally you want to open up at least 5stops through the test. Most dslrs fstops' move in 1/3 increments, in this case you will make 15 exposures. You will note on the back of your camera the image getting brighter and eventually blowing out the whites.

Shadow test. Point the camera in spot metering mode at the darkest part of the subject, the deep shadow you arranged. You are tying to meter off the area where you would want minimal detail.

The camera will be in manual mode, so set the shutter speed so you get correct exposure at a low-ish fstop (perhaps f4). Then, adjust the fstop to higher numbers so you are letting less light in, gradually under exposing. Ideally you want to close down at least 5stops. As in the highlight test, expose at each 1/3 stop closing the lens up to minimum aperture (f22 or whatever your lens is capable of). You will note on the back of your camera the image starting with out blown out whites, then getting darker and eventually loosing a lot of detail as you go through the steps.

Midtone Colour test. Use the spot metering mode to meter a mid-grey that is in the scene. Ideally a grey card should be used. The grey card should be facing the camera on the same plane as the scene (the closer to the light the brighter it gets, so meter as precisely at the subject as you can) and with grey card pointed directly at the camera.

Next, in manual exposure mode shoot three stops up and three down from the ideal metering. Use the smallest increments your camera offers. Your starting fstop needs to be midway on the scale (f8) to allow for the 3 up and 3 down. If your camera shoots in 1/3 stop increments, you will shoot 9 frames up and 9 down from that mid point.


Evaluation of shadow and highlights. This is where you need your opinions. Open up all of the files in your editing software. I use adobe Lightroom as it allows for quick evaluation of the raw files. Make sure all presets or effects are off. A good quality, properly calibrated monitor will ensure you are seeing the detail you should see. Many monitors struggle at the extremes and we are testing extremes.

Start at the first exposure for white, which will look ok, but dark. Move through the successively lighter exposures until you find one that you can't recover the highlight detail satisfactorily. Use both exposure adjustment and highlight recovery tools. The image prior to this is your is your 'white with detail' point. Count back to the start exposure. Count the full stops. This is your highlight latitude.

Next look at the exposure for black where the shadow you metered off is grey and the image will look very light. Again move through the successive exposures, getting darker, looking for the image where the shadows are strong but can be lightened so detail appears, but noise is acceptable to you. Use exposure and shadow lightening methods, you are testing all the tools you have to make those extremes work for you. Again count the full fstops to the first exposure for the shadows. This is your shadow latitude.

Evaluation of midtones. Start with your ideal exposure and work both up and down the apertures noting at what point the pastel, or midtone colours begin to change. You may see a desaturation or loss of colour in some of these softer colours. As they drift away from ideal, note both the under and over exposure point at which this happens. This will give you your midtone latitude.

In conclusion and in practise. If midtone colour and highlight and shadow detail are critical to you then you need to know your camera and how to use it's light meter.

If you know white with detail is 3 stops over exposed, then, when shooting a scene, meter the white you want detail in. If it is 3 stops or less over the midtone exposure, then all is great!

If your black with detail is 4 stops under exposed, and your scene metering matches this, then also great!

If the scene falls outside of these boundaries...

  • You have to accept a less than ideal result. Which you will understand what it will look like as you will have performed this test.
  • You can change the lighting by using repositioning, reflectors, or flash to manipulate the dynamic range of the scene.
  • Or use HDR techniques to take multiple frames varying the exposure for the problem areas, and combine the frames in post production.

We are bound by the laws of physics. In knowing your boundaries, you can play at your fullest potential.

Paul Atkins.