W3C home > Mailing lists > Public > w3c-wai-eo@w3.org > January to March 2010

Re: Font recommendation for Accessible Presentations

From: Cliff Tyllick <CTYLLICK@tceq.state.tx.us>
Date: Mon, 29 Mar 2010 11:03:06 -0500
Message-Id: <4BB088E9.41EE.0066.0@tceq.state.tx.us>
To: "eowg" <w3c-wai-eo@w3.org>
Shawn, I think this would work just fine:
>How about adding something like: "Simple fonts with consistent thickness are often easier to read from a distance (as opposed to fonts where parts of the letters are thin, like Times New Roman)." ?
And where you point out:
 
>Many people do not have access to a projector when they are creating their presentations (myself included, btw). (Also, we would need to clearly explain this test, e.g., how do you "take the image slowly out of focus".)
How about:
 
"If you have access to a projector, here is one way to find a suitable font:
1. Prepare a slide with a phrase repeated in each font you are considering.
2. Project that slide on a screen.
3. Slowly adjust the projector to bring the image out of focus.
4. As the image gets farther out of focus, note which font remains legible the longest. This font will generally be the best to use for your slides."
 
I'm not insisting that we add such wording, but it occurs to me that to do so would be much like the approach taken with color contrast in WCAG 2.0. And if we refer to adjusting the projector, then "How do I change the focus?" is best answered by the guide for using that projector, not by us.
 
~Cliff
 
Shawn Henry wrote:
>Thanks for the additional input, Cliff.
 
>> Shawn, I am not opposed to providing research where it exists, but it seems to me that we might not need it in this case.
 
>[Shawn considers and nods head probably in agreement.]
 
>> We also have a fairly simple test everyone can run on any fonts they are thinking of using for their slide presentations: Project an image of the same line of text in each font on a screen, take the image slowly out of focus, and determine which font produces the image that remains intact the longest. If we explain it that way, we might not even need to mention stroke width.
 
>Except that many people do not have access to a projector when they are creating their presentations (myself included, btw). (Also, we would need to clearly explain this test, e.g., how do you "take the image slowly out of focus".)
 
>How about adding something like: "Simple fonts with consistent thickness are often easier to read from a distance (as opposed to fonts where parts of the letters are thin, like Times New Roman)." ?
 
~Shawn


Cliff Tyllick wrote:
> Shawn, I am not opposed to providing research where it exists, but it seems to me that we might not need it in this case.
> 
> In print and on screen, we can't manipulate individual images and see what happens to their clarity. In both cases, the resolution is whatever the medium provides. And there is value in running tests with many subjects to be sure that what seems readable to me is generally considered to be readable by everyone.
> 
> But in a projected image, we can manipulate the image simply by adjusting (or, in this case, maladjusting) the focus. As a result, there is a physically observable event -- specifically, at some point in some fonts the narrower portions of letters disappear. You don't need to sample a population to understand the impact of this event on legibility of the text.
> 
> As a person with worse than 20/400 vision (uncorrected; my eyeglasses do correct it to 20/20), I can vouch for the fact that taking a projected image out of focus gives a close approximation of what I see without my glasses. I suspect there is fairly widespread observation of this phenomenon, so there should be general agreement on this point, too. In other words, I think all can agree, or at least concede, that the manipulation of the projected image does give meaningful information.
> 
> Put those two points together, and we have a physical experiment that can be run by anyone who is interested in the question. In other words, we have crossed into the realm of the hard sciences: "Doubt my conclusion? Run the test yourself, with a sample size of n = 1 observer, and see if you get a different result."
> 
> We also have a fairly simple test everyone can run on any fonts they are thinking of using for their slide presentations: Project an image of the same line of text in each font on a screen, take the image slowly out of focus, and determine which font produces the image that remains intact the longest. If we explain it that way, we might not even need to mention stroke width.
> 
> If there is statistical research out there, I'm all for considering it. I'm just not sure it would matter what that research had to say in this case.
> 
> Cliff
> 
>>>> Shawn Henry <shawn@w3.org> 3/26/2010 4:32 PM >>>
> Thanks for this, Cliff!
> 
> I know there is research on readability of fonts, certainly online and in print - I'm not sure about projected. Before we provide specific guidance, it would be good to have the research to back it up.
> 
> Does someone want to gather research on font face readability and provide a summary?
> 
> Also, I wonder if "stroke width" itself is enough, or if some people would need more explanation of what that means?
> 
> Best,
> ~Shawn
> 
> 
> Cliff Tyllick wrote:
>> I have a thought about this exchange between Shawn and Liam on "Use an 
>> easy to read font face":
>>
>> Liam: Sans is not *always* easier to read than serif - too much of 
>> a generalisation. Suggest mentioning a few commonly available fonts 
>> by name instead.
>> Shawn: um - I wonder if that would be more complicated on our end. Maybe 
>> just saying "Use an easy-to-read font face. Avoid fancy fonts that 
>> are difficult to read." is enough given the scope of this document/page?
>> Liam: Yes, much better.
>> Cliff's thought: I wonder if this be even better: "Use an easy-to-read 
>> font face, especially a font with a constant or nearly constant stroke 
>> width."
>>  
>> Here's my thinking: It seems to me that the main reason serif fonts are 
>> generally less readable than sans serif fonts has to do with stroke 
>> width. Especially among the older serif fonts, the stroke varies quite a 
>> bit. It's rather thin in the horizontal portions of characters and is 
>> much thicker on the vertical portions. By contrast, the older sans serif 
>> fonts have a stroke thickness that is uniform, or nearly so, throughout 
>> each character.
>>  
>> As a result, when the image is out of focus, text can still be readable 
>> in such a sans serif font while portions of the letters in a serif font 
>> have disappeared. Of course, the image is always out of focus for people 
>> with low vision, but we can simulate this experience by slowly taking a 
>> projected image of the same passage in three different fonts -- Times 
>> New Roman, Arial, and Georgia -- out of focus.
>>  
>> Times New Roman and Georgia are both serif fonts; Arial is sans serif.
>>  
>> The stroke width varies quite a bit in Times New Roman varies; the 
>> stroke width is constant or nearly constant in Georgia and Arial.
>>  
>> As you take the image out of focus, the passage in Times New Roman will 
>> become illegible when the passages in Georgia and Arial are both still 
>> readable. The passages in Georgia and Arial eventually become illegible 
>> at about the same time.
>>  
>> So a person at the back of the room could still read slides in Georgia 
>> or Arial when they couldn't read slides in Times New Roman. Either 
>> Georgia or Arial is a good choice; Times New Roman is not a good choice. 
>> By no means would I consider Times New Roman to be either "fancy" or 
>> "difficult to read," but if you were to clue me in to the importance of 
>> the stroke width, perhaps I would look at it more carefully and make a 
>> more careful choice.
>>  
>> Is this analysis on target?
>>  
>> Cliff Tyllick
>> Texas Commission on Environmental Quality
>>  
>> t: +1 (512) 239-4516
>> e: ctyllick@tceq.state.tx..us <mailto:ctyllick@tceq.state.tx.us>
>> w: www.tceq.state.tx.us <http://www.tceq.state.tx.us>
> 
Received on Monday, 29 March 2010 16:04:01 UTC

This archive was generated by hypermail 2.4.0 : Friday, 17 January 2020 20:29:44 UTC