[Blindmath] Missing and unwanted pauses in MathPlayer output

[J.Fine]

Hi Stephen

Thank you for sharing your experience, particularly the quality assurance process.  My examples come from a somewhat difference qa process.

Let me clarify my first example.  Please assume or pretend that you are reading this message using a screen reader.  In Sarah's sample there is a multiline MathML equation that when read sounds exactly like this - numberofwaystoarrange 9 books.  What it should sound like is - number of ways to arrange 9 books.

Similarly, provided you are reading this message with a screen reader, the second example sounds like " - 0. 1498  plus 0.0 741 " when it should sound like "-0.1498 plus 0.0741 ".

Let me put it another way.  MathPlayer supplies speech text to JAWS, which turns it into audio.  Rather than attach audio files I am placing the speech text in the email message.  Hence the confusion about my message not containing MathML but being turned into speech.

Best regards



Jonathan



> -----Original Message-----
> From: Stephen L Noble [mailto:steve.noble at louisville.edu]
> Sent: 14 April 2011 12:45
> To: 'blindmath at nfbnet.org'
> Cc: Neil Soiffer
> Subject: Re: [Blindmath] Missing and unwanted pauses in
> MathPlayer output
>
> Hi Jonathan,
>
> MathPlayer doesn't do anything with the text on the page, it
> only pays attention to MathML islands on the page.  In your
> example - "number of ways to arrange 9 books" - that should
> just be plain HTML markup, there would be no reason to mark
> up that text within a MathML island, so I don't understand
> the connection. Not sure about your other example, but I
> assume that is from a simple addition problem where perhaps
> it is indeed marked up as MathML. It is possible that
> different voice families will impact some aspects of the
> reading of a series of numbers, regardless if it is within a
> MathML island or just plain text on a page.
> Since I do a lot of conversion with MathType and routinely
> listen to lots of pages with synthetic speech as part of the
> quality assurance process, I have certainly noticed this.
>
> Neil Soiffer from Design Science usually monitors this list.
> You might want to send him a copy of one of the pages upon
> which you found the problem. His email is neils at dessci.com
>
> Best regards,
>
> --Steve Noble
>
> >>> J.Fine <j.fine at open.ac.uk> 4/14/2011 5:01 AM >>>
> Hi
>
> First, please let me know if this message is not appropriate
> for this list.
>
> I've found that MathPlayer sometimes gets pauses wrong, even
> with MathML generated by MathType.  Here's an example, from a
> sample file Sarah Jevnikar posted to this list on 3 April.
>
> What one should get is - number of ways to arrange 9 books ...
> MathPlayer plus Jaws10 gives - numberofwastoarrange 9 books ...
>
> The last example misses the spaces between the words.  Here's
> another example, taken from a source file for an Open
> University course.  It has extra spaces, this time inside numbers.
>
> What one should get is - 0.1498 plus 0.0741 plus 0.4293 plus
> 0.0280 plus 0.2218 MathPlayer plus Jaws gives -  0. 1498
> plus 0.0 741  plus 0. 4293  plus 0.0 28 0  plus 0. 2218
>
> Sarah's sample file does not have these problems, even though
> also produced using Word plus MathType plus MathPage.
>
> That said, from my tests it seems that MathPlayer can do a
> pretty decent job of generating speech text provided the
> MathML that comes in is just right for that purpose.  (These
> tests used an evaluation copy of MathFlow, which Design
> Science kindly provided.)
>
> In both cases the problems is spaces between characters.  It
> seems to me that errors here are fairly easy to detect and
> correct if you can see the text, but not if you can only listen to it.
>
> The rest of this message is a brief technical explanation.
> In the first example MathPlayer ignores the non-breaking
> space characters between the words.  Hence the funny
> pseudoword wastoarrange, which should be ways to arrange.  In
> the second example MathType has grouped the digits and
> decimal point in a strange way and MathPlayer does not
> correct this grouping.
>
>
> Best regards
>
>
> Jonathan
>
> --
> The Open University is incorporated by Royal Charter (RC
> 000391), an exempt charity in England & Wales and a charity
> registered in Scotland (SC 038302).
>
>
> _______________________________________________
> Blindmath mailing list
> Blindmath at nfbnet.org
> http://www.nfbnet.org/mailman/listinfo/blindmath_nfbnet.org
> To unsubscribe, change your list options or get your account info for
> Blindmath:
> http://www.nfbnet.org/mailman/options/blindmath_nfbnet.org/ste
> ve.noble%40louisville.edu
>
>
>
>
> _______________________________________________
> Blindmath mailing list
> Blindmath at nfbnet.org
> http://www.nfbnet.org/mailman/listinfo/blindmath_nfbnet.org
> To unsubscribe, change your list options or get your account
> info for Blindmath:
> http://www.nfbnet.org/mailman/options/blindmath_nfbnet.org/j.f
> ine%40open.ac.uk
>