Groups | Search | Server Info | Keyboard shortcuts | Login | Register [http] [https] [nntp] [nntps]


Groups > comp.soft-sys.math.mathematica > #14306 > unrolled thread

Re: Speak errors (was Re: audio)

Started bydebguy <johnandsara2@cox.net>
First post2013-04-25 06:52 +0000
Last post2013-04-27 03:08 +0000
Articles 4 — 3 participants

Back to article view | Back to comp.soft-sys.math.mathematica

This discussion starts older than the indexed window; earlier articles aren't shown. The article labeled Started by below is the oldest one visible, not the original post.


Contents

  Re: Speak errors (was Re: audio) debguy <johnandsara2@cox.net> - 2013-04-25 06:52 +0000
    Re: Speak errors (was Re: audio) Richard Fateman <fateman@cs.berkeley.edu> - 2013-04-26 08:24 +0000
    Re: Speak errors (was Re: audio) debguy <johnandsara2@cox.net> - 2013-04-26 08:24 +0000
    Re: Speak errors (was Re: audio) Richard Fateman <fateman@eecs.berkeley.edu> - 2013-04-27 03:08 +0000

#14306 — Re: Speak errors (was Re: audio)

Fromdebguy <johnandsara2@cox.net>
Date2013-04-25 06:52 +0000
SubjectRe: Speak errors (was Re: audio)
Message-ID<klajqo$3as$1@smc.vnet.net>
?? http://msdn.microsoft.com/en-us/library/ms717077%28v=vs.85%29.aspx
??

{f[x*(x + b*(c + x + z))], f[x*(x + b*(c + x) + z)]}

{f[x (x + b (c + x + z))], f[x (x + b (c + x) + z)]}

So I am to understand mm spoke exactly as asked: because they are not
the same.  Is that right?

?? http://msdn.microsoft.com/en-us/library/ms717077%28v=vs.85%29.aspx
??  Microsoft didn't develop speach.  Linux had it before Microsoft
did.  DragonSpeach didn't develop it either.

The U.S. Government developed it (what dragon speech uses) and much
later released it public domain.  I once tried to find it that same
releease: it somehow is very hard to find even though it's public
domain.

Understanding a language is what can be an issue.  bnf parsers can
interpret grammar.  but the correct context of, even people are not
good at.

[toc] | [next] | [standalone]


#14314

FromRichard Fateman <fateman@cs.berkeley.edu>
Date2013-04-26 08:24 +0000
Message-ID<klddk0$d0q$1@smc.vnet.net>
In reply to#14306
On 4/24/2013 11:52 PM, debguy wrote:
> ?? http://msdn.microsoft.com/en-us/library/ms717077%28v=vs.85%29.aspx
> ??
>
> {f[x*(x + b*(c + x + z))], f[x*(x + b*(c + x) + z)]}
>
> {f[x (x + b (c + x + z))], f[x (x + b (c + x) + z)]}
>
> So I am to understand mm spoke exactly as asked: because they are not
> the same.  Is that right?

I don't understand your question.  Mathematica speaks the two different
expressions the same way, although they are not equivalent.  Mathematica
has an option to insert audible parentheses, but that is clumsy, and it
should not be necessary.  Who knows though. Wolfram might think this bug
is a feature.
>
> ?? http://msdn.microsoft.com/en-us/library/ms717077%28v=vs.85%29.aspx
> ??  Microsoft didn't develop speach.
I didn't say that Microsoft developed (text-to-) speech.  I indicated a 
link to a contemporary web page that could be used as a reference.

Linux had it before Microsoft
> did.
Are you sure?
  DragonSpeach didn't develop it either.

The first system I am familiar with to speak mathematics was AsTer,
written by T.V. Raman

http://www.cs.cornell.edu/home/raman/aster/aster-toplevel.html
in 1994.


>
> The U.S. Government developed it (what dragon speech uses) and much
> later released it public domain.

Citation please?  The original Apple Macintosh had MacinTalk, circa 1984.

You are, I think, confusing speech OUTPUT  (relatively easy, at least if 
you are content with a slightly robotic voice.) with
speech INPUT (a much harder problem. e.g. Siri.)

Good public (free, open source) speech recognition is sufficiently hard
to find and set up that companies like Nuance make a good business 
selling it shrink-wrapped.

  Speech-to-text is available "in the cloud" via  (for example) Google 
Voice.

Here's one open project though...

http://cmusphinx.sourceforge.net/sphinx4/

Setting it up is not for the faint of heart.

Programs to convert spoken mathematics to text (or mathml equivalent)
are around, but probably not ready for prime time.


..snip..
RJF

[toc] | [prev] | [next] | [standalone]


#14316

Fromdebguy <johnandsara2@cox.net>
Date2013-04-26 08:24 +0000
Message-ID<klddka$d15$1@smc.vnet.net>
In reply to#14306
I'm sorry if I got off topic Richard Fateman.

Sorry I cannot test the audio situation I have an earlier mm ver.  But
i have an idea.

"I have tried to work with pausing and the problem that I am
encountering with pausing is that what ends up happening is the
displaying of the disks is delayed and while the buttons are delayed
they are not disabled, so it still allows that audio to become out of
sync."

(a race condition - that is why I suggested simple waiting.  i see no
solution posted above so let me try again....)

TRY THIS:

The Front-End has it's own evaluator separate from the kernel.  I'm
thinking maybe that's where your split (and then race condition)
occurs.  Make sure all expressions effected are done by one or the
other but are not split in-between both.  See "FrontEndEvaluate" for
more info.  I'm thinking that because I beleive many new features todo
with dynamic front end content have some level of integration with the
new front-end evaluator.

[toc] | [prev] | [next] | [standalone]


#14326

FromRichard Fateman <fateman@eecs.berkeley.edu>
Date2013-04-27 03:08 +0000
Message-ID<klfff1$jkr$1@smc.vnet.net>
In reply to#14306
On 4/26/2013 9:06 AM, Louis Talman wrote:
> On Fri, 26 Apr 2013 02:25:16 -0600, Richard Fateman 
> <fateman@cs.berkeley.edu> wrote:
>
>> Mathematica
>> has an option to insert audible parentheses, but that is clumsy, and it
>> should not be necessary.
>
> When I was a graduate student many moons ago, I read research journal 
> articles onto tape for a blind mathematician.
>
> I have to disagree that inserting audible parentheses isn't 
> necessary.  It may be clumsy, but most of audible mathematics is 
> clumsy. There are, I think, few alternatives, short of using something 
> akin to (reverse) Polish notation, for some folks.

Sorry I was not clear on this. _Mathematica's option is *clumsy*. _ I 
agree that some grouping markers are necessary.
Example.

Let s = a*(b + c) + d
SpokenString[s]  is

a times the quantity b plus c plus d
   (which Mathematica 9 pronounces  defectively as     ahtimes ....,   
and fails to group the b+c.)

SpokenString[s, IncludeParentheses -> True] is

a times open parenthesis b plus c close parenthesis plus d

which is, in my opinion,  CLUMSY.   Incidentally, the documentation is 
buggy.
It directs you to type  "IncludeParentheses"-> True,  but the quote 
marks are unnecessary.
Perhaps this is all the work of a summer intern.).

How then to group the expression if not to say o-pen pa-ren-thi-sis    
(6 syllables) and close pa-ren-thi-sis  (5 syllables)?

How about
A times open b plus c close plus d  or
A times quantity b plus c endquantity  plus d
A times left b plus c right plus d

or (considerably more subtlely) as done by AsTer,  see
http://www.cs.cornell.edu/home/raman/aster/sec-06.html

by changing the pitch higher within "the quantity"  and dropping it at 
the end.

Or consider what Mathematica does for Sin[x+y], including parentheses..
sine of open parenthesis x plus y close parenthesis
which is unambiguously spoken as sine of quantity x plus y, so clumsy..

There may also be a bug in

SpokenString[HoldForm[Sin[x] + y], IncludeParentheses -> True]

which says sine of x plus y.

Oh, it seems that if you want to pronounce the variable "a",  you can do 
so by substituting "eh".
This may scramble the ordering, so you need to do something like
HoldForm[eh*(b+c)+d].

This works. But especially for Canadians. :)
http://tx.english-ch.com/teacher/dai/others/eh-is-canadian-eh-/

RJF

[toc] | [prev] | [standalone]


Back to top | Article view | comp.soft-sys.math.mathematica


csiph-web