Ensure that descriptions are accurate and unfiltered


Audio description aims to replace important visual content with equivalent audible content. Equivalence requires accuracy, objectivity and completeness. A description that is inaccurate or subjective cannot possibly be equivalent. Given the time constraints and the amount of visual information in a scene, absolute completeness is seldom possible. But descriptors should strive to be impartial in what is included and what is left out. If not, they are interpreting or censoring the content, which is discriminatory by nature.

Directions and techniques

Describe what is seen

The describer’s intention should always be to state what they see, not their interpretation of what they see. This is often described as the ‘first golden rule’ of audio description. For example, motivations and intentions should not be described because they are not directly seen. However, visuals that are important for revealing motivations and intentions should be described.

On the surface this ‘rule’ seems quite simple but it is sometimes difficult to achieve. For example, suppose a character in a drama uses body language to demonstrate a particular emotion that is very important to the plot. The ‘golden rule’ would say not to name the emotion by saying the character is “angry” or “suspicious”, but to describe the body language that gives the emotion away, such as “clenching her fists” or “frowning”. The problem is that body language can be subtle and multifaceted. It may be difficult to describe in a way that gives the same information as can be got from seeing it. A person clenching her fists may be either angry or scared. A person frowning may be either suspicious or confused. It may be clear by seeing the frown that it is a ‘puzzled expression’ or a ‘suspicious look’, but saying either of these could be seen as interpretation. However, describing the action as “frowning” may risk provoking the wrong interpretation. Particularly in cases where the correct interpretation relies on combining a number of contextual clues including a collection of expressions or gestures, and describing them all is not possible. In some cases, the clues may be intentionally imprecise so that different viewers will come up with their own interpretations. These subtleties can make this guideline difficult to know how to follow but this should not lead to it being ignored. It still stands that viewers do not want the describer’s interpretation, they want to make their own interpretations and this is the crux of the matter. The guideline is pointing at the general idea that it is not up to the describer to tell the viewer what to think. Dealing with challenges like this is what makes audio description a creative art that can be helped by guidelines and honed by experience and audience feedback, rather than a science that can be precisely described in a set of exact rules.

Do not censor

Do not leave out objects or actions just because they may be offensive. This is unnecessary and condescending. If the content is deemed suitable for its intended audience, it is also suitable for a vision impaired member of that audience. The describer’s comfort level when describing content such as sexual acts or violence should not be a factor.

Be precise and consistent

Always use the unique names of people and objects rather than pronouns like “he” if there could be any confusion about which is being referred to. If there is only one in the scene or it is obvious for some other reason, pronouns can be used.

When referring to a character, place, object or other item by name, use the same name throughout the programme.

Use precise, vivid words. People frequently “walk” but they also “amble”, “stagger”, “shuffle”, “saunter” or “stroll”. These words may more accurately describe the scene and may enable the viewer to create a richer mental image.

Take account of what viewers already know

If an item has already been mentioned, e.g. “a vase”, refer to it as “the vase” as long as it is clear which vase is being referred to.