Understanding the basics of low light photography

After dark isn’t the only time you can encounter dim lighting; it can also occur inside buildings ““ or even in narrow streets between tall buildings that block the sunlight. Light can have difficulty penetrating thick forests, particularly under cloud-covered skies. And weather conditions like storms and smoke from bushfires can also reduce ambient light levels.

Scenery often appears at its most attractive when the sun is low in the sky, either just before sunrise or after sunset.

However they originate, low light levels give photographers less light to work with. As a result, effects that can be relatively minor in brightly-lit conditions can become either obstacles ““ or opportunities for creativity ““ that can challenge the way you use your equipment and influence the types of images you can produce.

However, you do need to understand your camera’s controls and the limitations of the equipment you are using. So we’ll begin this guide by reviewing the fundamentals of exposure, sensitivity and noise.

Exposure

Since the first photographs were taken, exposure has been made up of three factors:

1. The brightness of the scene.

2. The sensitivity of the recording material.

3. How long the recording material is exposed to the light reflected from the scene.

Interestingly, photographers can control all three factors, although some permit a wider range of more precise controls than others. For example, aside from waiting for a different time of day or different weather conditions, you can only control the brightness of the scene by adding light with a reflector or artificial light source or changing your shooting angle to one where the subject is better lit.

The time of day, ambient light levels and distribution of tones within the scene can all affect exposure parameters.

The sensitivity of the recording medium (in today’s world, the image sensor) is adjustable across a huge range of values, compared with the range available to film photographers. However, there are consequences associated with these adjustments that must be taken into account.

The camera’s aperture and shutter speed settings are used to control how long the sensor is exposed to the light reflected from the scene. Adjusting these settings will affect how much of the resulting image appears sharp and how movement is recorded.

Use of a fast lens with a maximum aperture of f/2.8 (used for this shot) made the subject appear sharp against a blurred background.

Wide aperture settings are used to reduce depth of field and blur away distracting background details. Very fast shutter speeds will ‘freeze’ movement, while slow shutter speeds will blur it.

These two images illustrate the effect shutter speed can have on how motion is recorded. The top image was recorded at 1/40 second, while the lower one was a 1/5 second exposure, which left the sensor exposed for eight times longer.

Regardless of the subject and the lighting conditions, photographers should always strive for a ‘correct’ exposure. This means one in which detail is recorded in both highlights and shadows and where the tones between them are properly separated.

Overexposure will cause the photosites on the sensor to become ‘saturated’, which means they fill up with the light signal and can’t hold any more. This will produce images with blown-out highlights that are totally white with no detail recorded. There is no way to replace detail that was not recorded in the initial exposure.

Underexposure is more complex because even if only a few photons have been recorded, there’s a chance some detail may be recoverable ““ although it’s unlikely. However, if the sensor’s noise levels are higher than the recorded signal, all you’ll get is noise (which is undesirable).

When faced with the potential for underexposure there are three ways to handle it: add more light, open the lens aperture and/or slow the shutter speed. Choosing which combinations of options to use is part of the art of photography.

Sometimes you want to retain the predominance of dark tones in the subject, while at other times it may be better to have the image tones evenly spread across the image’s dynamic range. Your camera’s histogram can help you to decide on the correct exposure level.

When you want to keep the dark tones predominant in the image, the histogram should be biased towards the left hand side of the graph.

A more even tonal distribution is shown in the histogram for the above image.

Sensitivity

Since the days of film, the sensitivity of the recording medium has been defined by the International Standards Organisation (ISO) which has established a standard scale that can be used by photographers to denote the recording medium’s ability to respond to light. Most image sensors have been designed for optimal performance at low sensitivity settings, typically ISO 100 or ISO 200. To achieve higher settings, the digital signal (produced by the exposure) must be boosted electronically.

It’s common for modern cameras to offer ISO settings between ISO 100 or 200 and ISO 6400. Most DSLRs and CSCs (compact system cameras) will go as high as ISO 25600 and some extend this further to ISO 51200. Some also offer ‘low’ ISO settings of 50 or lower.

The standard ISO range adjusts sensitivity in one-stop increments, which double or halve the value, depending on which direction the setting is changed. Moving from ISO 100 to ISO 200 doubles the sensor’s sensitivity, while going from ISO 100 to ISO 1600 increases it by a factor of four.

Each time you double the ISO, the camera needs only half as much light for the same exposure. So, if your metered exposure is 1/25 second at f/5.6 with ISO 100, moving to ISO 200 enables you to shoot at 1/50 second, while going to ISO 1600 will allow you to shoot at 1/400 second, assuming the aperture is unchanged. In very low light levels, you may need to set the ISO even higher.

The sensor itself doesn’t become more sensitive; instead, the camera’s image processor amplifies the digital signal to change the way the sensor reacts and bring out details in the subject that wouldn’t otherwise be recorded. Modern DSLRs and CSCs use sophisticated processing algorithms to tweak this processing in order to minimise the effect of noise, one of the main problems encountered by low light photographers.

Image Noise

Some background noise is always present in any electronic device that transmits or receives a signal. In digital cameras, that signal is the light captured by the sensor’s photodiodes. Image noise can be produced by the randomness of the photons of light (known as ‘shot noise’), disturbances caused when converting the analog data to digital format (‘read noise’) and ‘fixed pattern noise’ that results from the combined effects of heat and amplification.

Shot noise will predominate when insufficient light reaches the sensor due to an exposure that is too brief for the lighting conditions. Read noise is less common in modern cameras and occurs when the exposure is adequate but there are flaws in the analog/digital conversion process. Fixed pattern noise is most common in long exposures and is amplified by environmental heat.

The more light the photosites can collect, the less susceptible a sensor will be to noise because more light is collected relative to the random electronic discharge, which remains virtually constant for a given temperature. High ISO settings amplify both the light signal and the electronic noise, thereby producing higher noise levels than low ISO settings.

A build-up of this background noise during long exposures produces ‘salt-and-pepper’ speckling where some pixels are not recorded, while others are only recorded in one wavelength (colour). The problem is exacerbated by heat, which explains why the best long exposures are usually captured on cold nights.

An example of an image with visible noise resulting from an exposure at ISO 51200.

A crop from the above image, enlarged to 100%, shows the random specking produced by noise resulting mainly from the high sensitivity setting.

Random electronic noise can be mostly eliminated by dark frame subtraction, in which the camera records a second exposure with the shutter closed to capture the noise pattern. This ‘dark frame’ is subtracted from the image by the camera’s image processor, leaving the picture virtually noise-free. The only problem is that it doubles the exposure time.

Noise Reduction

All modern cameras provide some kind of in-camera noise-reduction (NR) processing, which is applied to JPEG files. Most cameras provide separate processing for High ISO and long exposure noise because they require different adjustments.

High ISO NR is handled by the image processor and is based upon calculating whether the actual differences in pixel values constitute noise or genuine photographic detail. Once noise is identified, the processor will average out the former while attempting to preserve the latter.

Since noise levels increase as ISO sensitivity is increased, processing tends to become more aggressive at higher sensitivities. This can produce a noticeable loss of image quality. Details can be smoothed out and colour irregularities may appear in JPEGs recorded with high ISO values.

If you shoot raw files, noise-reduction is included as part of the file conversion process and you have access to a wider range of adjustments. Different raw file processors provide different adjustments and the best ones allow photographers to fine-tune adjustments to suit individual image files.

Samsung’s Adaptive Noise Reduction technology uses the camera’s image processor to analyse the scene before applying noise reduction locally rather than uniformly across the whole image. The advanced algorithm analyses details and colour in the subject as each shot is taken and processing is applied in-camera, resulting in a noticeable increase in clarity in shots taken with high ISO settings.

 

Excerpt from Low Light Photography