skip to primary navigation skip to content

Clari-Fi

Clari-Fi uses blurring to simulate the challenges of viewing small details on packs and display screens

 

Relating the Clari-Fi levels to population data

In order to understand how the Clari-Fi Check Levels are related to population data, it is necessary to understand the survey data that is used in the calculations. This data is used to predict the percentage of people who cannot see particular details, based on the level of blurring at which those details remain visible.

On this page:

Which survey data is used?

The population percentages used in the Introduction to assessing apps and websites were derived from the Better Design Survey (2010). This was a postcode sampled survey of 362 participants, which was conducted inside people's own homes, and tested people’s near vision.

The test intended to measure people’s ‘real-world’ near vision ability (i.e. the level of near vision ability that they have for the majority of the day). So, if participants were wearing glasses or contact lenses before the test, then they kept them on. If they weren’t wearing any glasses or contact lenses before the test, then the test was conducted without glasses or contact lenses.

This survey is particularly relevant for evaluating apps and websites on mobile devices and small portable laptops, because it included a vision test using a handheld test chart, which participants could hold at whatever distance worked best for them. This reflects the way in which people use mobile devices and small portable laptops. The smallest four rows of the test chart are shown below at an enlarged size below.

Participants were given the handheld test card and asked to read out rows of test-chart letters that got smaller and smaller, to determine the smallest row that they could read successfully (i.e. 7 or more letters correct out of 8).

What were the survey results?

The bubble plot below relates the survey participants’ ages to the height of the smallest row that they could read successfully (i.e. 7 or more letters correct out of 8).

The row of bubbles along the bottom indicate participants with excellent near vision, who can perceive very small test chart letters. The scattered bubbles towards the top indicate participants with very poor near vision, who require the test chart letters to be much bigger so that they can read them. This mostly occurs in the older age ranges.

The graph shows that on average, the older half of the sample have much worse near vision than the younger half, and the spread of ability evident within the older half is much greater. Primarily, this is because from about age 45 onwards, age-related long sightedness typically makes people’s near vision ability much worse.

The next graph shows the same data in a slightly different way. For each row of letters on the test chart, the percentage of participants that cannot read that row was calculated and then the graph below was created by drawing a smooth curve through these data points.

Note that in the above graph, the height of test chart letters has intentionally been put on the y-axis, so that it directly matches the previous bubble plot.

In order to relate this curve to the degree of blurring that is specified in the Clari-Fi Check Levels, it is first necessary to understand how Gaussian blurring affects the readability of test chart letters. This is described in the next panel.

How does Gaussian blurring affect the readability of test chart letters?

The effect of Gaussian blurring on the readability of test chart letters was measured experimentally with 47 participants (who didn't have any specific vision issues1). The results from this experiment are summarised in the picture below.

The smallest line of blurred letters that can be read doesn’t usually depend on the vision ability of the person reading the letters, as long as the person has reasonably good vision (i.e generally good enough to drive, and no specific issues with near vision).

Try reading the letters in the chart above yourself. You will probably find that you:

  • Can read the middle row (7 or more letters correct out of Z H R F E K D N);
  • Cannot read the bottom row (2 or more letters incorrect out of F N K E D H R Z).

Try reading the chart again at a different viewing distance, or at different levels of browser zoom, or on different monitors. Within reasonable variations of these parameters, you should find that it makes no difference to the smallest line of letters that you can successfully read, because this is being limited by the degree of blurring that has been applied to the chart.

This effect is particularly convenient because it means that Clari-Fi can be used to evaluate apps and websites on a large screen monitor, and it doesn’t usually matter exactly how large the monitor is, how far away the assessor is from the monitor, or what level of zoom is used for viewing the blurred version of the image.

Footnotes

  1. 'No specific vision issues' means near vision ability equivalent of 20 / 40 (or better), measured at 40cm.
  2. 'Could read this row' means 7 or more letters correct out of 8. 'Could not read this row' means 2 or more letters incorrect.

Applying the survey data to the Check Levels

The Better Design Survey (2010) had a robust sampling strategy, so its results give a good indication of the spread of vision ability in the population. We would have preferred to use a data source with more than 362 participants, but we are not aware of any other household-based, postcode sampled survey, which used a near vision test chart, which participants could hold at any distance they wanted.

If you would like to run a survey with your customers to calibrate the Clari-Fi Levels directly to your customers, please contact edc-toolkit@eng.cam.ac.uk, and we would be delighted to discuss this as part of our consultancy services.

To best of our knowledge, the second graph in the survey results panel gives the best available estimates for the percentage of people in the general population who cannot read test chart letters of different sizes (at a handheld distance).

Furthermore, the graphs in the survey results panel can be used together with the experiment on Gaussian blurring to identify the appropriate degree of blurring that’s required for the 1%, 4%, 20% and 50% Clari-Fi Levels for near vision.

The graph below shows these blurring levels for near vision overlaid onto the corresponding survey data.

The following statements relate the outcome of a Clari-Fi assessment to the percentage of people that cannot perceive the details that were assessed:

  • If the visual details are bigger than the bare minimum needed to pass the 1% Level
    then less than 1% of people cannot perceive these details1.
  • If the visual details are bigger than the bare minimum needed to pass the 4% Level,
    but not big enough to pass the 1% Level,
    then between 1 and 4% of people cannot perceive these details1.
  • If the visual details are bigger than the bare minimum needed to pass the 20% Level,
    but not big enough to pass the 4% Level,
    then between 4 and 20% of people cannot perceive these details1.
  • If the visual details are bigger than the bare minimum needed to pass the 50% Level,
    but not big enough to pass the 20% Level,
    then between 20 and 50% of people cannot perceive these details1.
  • If the visual details are not big enough to pass the 50% Level,
    then more than 50% of people cannot perceive these details1.

Important note: all of these population percentages refer to people who cannot perceive particular visual details, because they are too small. There are many other reasons that might cause people to have difficulty perceiving these details. Clari-Fi assessments should be accompanied by checking contrast, simulating colour vision deficiency, and testing with users, as described in the Important notes section (within the Introduction to assessing apps and websites).

Footnotes

  1. This refers to the percentage of people who cannot perceive the visual details, while wearing any glasses or contact lenses that they usually wear for the majority of the day. For more detail, see the the Calculation assumptions section (within the Introduction to assessing apps and websites).

Feedback

We would welcome your feedback on this page:

Your name:


Your email:


I would like to receive the Inclusive Design Toolkit news bulletin. This bulletin showcases people’s achievements with inclusive design tools, and highlights new tools. It comes out once or twice a year.

I would be interested in contributing to the Inclusive Design Toolkit news bulletin.

I would like a FREE activation code for the Inclusive Design Toolkit's impairment simulator software.

Your comments:


Please leave this field blank (it's a spam trap):

Privacy policy. If your feedback comments warrant follow-up communication, we will send you an email using the details you have provided. Feedback comments are anonymized and then stored on our file server. If you select the option to receive or contribute to the news bulletin, we will store your name and email address on our file server for the purposes of managing your subscription. You can unsubscribe and have your details deleted at any time, by using our Unsubscribe form. If you select the option to receive an activation code, we will store your name and email address on our fileserver indefinitely. This information will only be used to contact you for the specific purpose that you have indicated; it will not be shared. We use this personal information with your consent, which you can withdraw at any time.

Read more about how we use your personal data. Any e-mails that are sent or received are stored on our mail server for up to 24 months.