SSC logo Scottish Sensory Centre
university of edinburgh

Adapting Video for VI Learners

Design issues and approaches

General principles related to VI needs

Below, we draw together the main hints and tips to keep in mind as you create your new educational video or multimedia pack. Our hints fall into two categories. The first has to to with primary design: things to do or avoid to make a given video clip easy to see. The second category of advice suggests what to do when there is no way that the video can be sufficiently enhanced: when you have to consider supplementary materials.

Primary design:

Educational video will not be novel for VI learners. Like their able counterparts, visually impaired people watch an average of 24 hours of TV per week.

However, the balance of design compromises differs between broadcast programmes and educational materials. Broadcasters use the medium to its contemporary artistic and technical limit to make their work attractive and attention-grabbing. This implies fast action, exotic shot angles, fancy titles and credits, multiple things going on at the same time. Many of these make modern programmes increasingly difficult for VI viewers.

Educational producers must also produce attractive materials, but they can take greater account of perceptual issues raised by mixed audiences. The result will be programmes designed more for understandability, legibility, and effective manipulation at viewing time.

People who are registered as visually impaired or blind have more residual abilities than you might expect: 7 out of 10 can use text if it is clear and large enough. Most visually impaired people lose their sight rather than being born with the condition: around 17 out of 20 have had progressive loss. This in turn means that most will have memories of better vision which will help them interpret audio and text descriptions from real experience. Even for congenitally blind people, visually rooted language including such things as colour descriptions comes to have associative meaning, with words like 'green' used with 'strawberries' being an alternative for 'unripe'.

Here is a quick checklist:

Colour - (see also the References section for links to in-depth coverage)

  • choose distinctive colours, especially for on-topic objects. Try to avoid colours with counter-intuitive meanings (green is a pretty universal 'safety' colour, red is 'hot', 'danger', blue 'cold', and so on).
  • avoid colour combinations which produce on-screen colour artifacts like edge creep
  • check to see if your colour choice works in grayscale.
  • don't rely on colour alone: instead, give multiple cues (such as shape, audio or spoken cues, on-object labels, or fixed object locality on screen).

Example: red / green colour blindness

If your video is on cooking, don't rely on colour alone to help distinguish baking from eating apples...

>Contrast -
  • aim for good, even contrast.
  • avoid glare spots.
  • don't use oversaturated colours.

Example: material with inherently poor contrast

Some material is difficult. In the clips below the intent is to show how cold misty air pushes in under warm clear air: the information is in the fog movement, which is in subtle shades. The second enhanced clip might be offered on a multimedia training pack to compensate.

Enhanced fog(D)

View video without (2.3 M) and with (2.5 M) enhancement>Scene complexity -

  • make background / foreground distinctions clear. Ask yourself if the usual techniques (such as using tight depth of field to defocus background) work for people with poor vision to start with.
  • avoid complex backgrounds.
  • consider carefully before using multiple windows - especially transient ones.
  • avoid rapid changes of scene and unneccessary cross cuts and zooms.
  • avoid overlaying (ghosting) material.
  • Give supplementary cues where possible.

Example: Cueing through dynamic cursors

A simple cursor, pointer, or bounding frame can help the VI viewer locate where critical action is taking place. The video shows such a cursor in operation to highlight a transient event.

>Example: Providing high quality enlargements

Although it is tempting to assume that pictures may be enlarged simply by using low vision aids or by taking advantage of computer operating system level enlargement tools, blowing up images in this fashion can lead to pixellation which renders the image unviewable. It is easy to build in high quality enlargements during development:

Low/Hi quality Enlargements(D)>Movement -

  • avoid multiple action points. A person with restricted fields may miss one part of what is going on
  • help users who need to pause on still frames by avoiding movement blur: use a high shutter speed if possible
  • help VI viewers orient themselves. Provide secondary cues (cursors, audio) where on-screen action is transient or indistinct.

Example: Over-spread action

A clock is timing a physics experiment, with the action (a ball rolling down an inclined plane) being shown in one corner of the screen with the clock in another corner producing problems for those with restricted fields of view. The producer could have tracked the clock with the ball, or overlaid time readings at particular points on the way down the slope, or could simply have called out the times on the audio commentary.

Physics experiment(D)>Example: still frames

Here are two still frames, one from a digital camcorder with a high frames per second capture which makes it easy to see what is going on when the video is paused.

Motion blur

Fast frame rate: no motion blur(D)>Text -

  • use clear fonts at an appropriate size and with good contrast. If you aren't able to offer user resizing and re-colouring, then any choices you make will be a compromise. A starting point for video viewed on a 20 inch screen from a normal living room distances might be text about 1" high with clear space all round and good interline spacing. However, if the text is to be viewed up close (on a monitor, for instance) you can reduce the text size - halving the viewing distance to the screen will allow halving of text height (but remember that the original was a compromise - don't treat such guidelines as sacrosanct).
  • avoid text effects: text which hops about, changes font or size, uses highly stylised fonts, fades, overlays other text, or presented in non-standard reading order.
  • restrict scrolling speeds.
  • if you provide open subtitles, settle on a fixed cue location.
  • use san-serif fonts,and try to use lower case - the word outlines are more distinctive.

Example: Active and stylised text

This mockup shows some of the problems facing VI viewers: for impact, the title designer has used active, ghosted, multifont text with poor contrast and multiple directions of movement.

>Designing supplementary material:

Designing video programmes for mixed audiences can be difficult. If working to the guidelines above becomes too painful, consider designing-in supplementary material. Supplementation built into primary production is both easier and more effective than adding material later.

If your medium is linear (VHS tape, DVD), then all the techniques we have described in the add-on section still apply. You can plan for supplementation, bearing in mind what channels will be available to you on the delivery medium you choose (with more on MPEG-2 systems like DVD, for instance, than on VHS) -

  • Take conventional photographs at the same time as you shoot the video - you can use them in handouts.
  • Keep A4 versions of diagrams for the same reason.
  • You can help audio subtitlers by leaving gaps for their supplementary explanations and descriptions. Don't explain one aspect while showing another: this leaves no room for descriptive manoeuvres.
  • Create script notes as part of the teaching pack, annotated with elapsed times.

However, as the primary designer you can go much further. You can consider providing special versions for VI viewers in ways which add-on modifiers would find difficult. Properly planned for, there is no great incremental cost to producing a second variant of the teaching material with extended audio descriptions, high-quality stills, data bursts and supplementary views of difficult material on multichannel streams

If you are designing non-linear materials then your ability to support learners with special needs is massively extended, and as with the linear materials, need not cost much more to provide. Because you can incorporate extra materials without resorting to tricks such as borrowing audio channels for subtitling, you can develop a single package to serve all users. Planning involves -

  • Collecting the same supplementary materials described above, together with extra video where this helps, and incorporating these assets into the main stream as branching options.
  • Choosing one or more 'hot-spot' selection metaphors and associated cues which match the needs of your target range of viewers. Partially sighted learners will be able to choose from on-screen branch options, provided they are in a fixed location and audio cued. Blind users might be given multiple options through a spoken menu in complex situations: for supplementary audio descriptions a simple sound prompt and mouse button response could be used to pause the main flow and start the supplementary description.
  • Adding text markers to the video, audio and picture assets, and incorporating any notes or scripts as screen viewable text files. Because these text markers and text files are ASCII resources (rather than text incorporated into video or pictures as images), they will enable teachers and students to create printed hard copy of relevant sections and to search both for basic information and supplements.

Example: User defined text sizing

Web Browsers allow the user some control over the choice of font size to be used when displaying text. The two pictures show the same text being displayed in the default font, and then in a font of a more appropriate size. In either case what makes this work is an acceptance on the part of the materials designer that they may not have complete control over text layout. Relinquishing artistic control in favour of user flexibility is hard for many graphical and multi-media designers.

Browser control over text(D)