I am really sorry to learn that some are having bad experience, in using the
Extended Dictionary. And I do apologize for any trouble caused.
For you, Orlando, Well, I thought you and I was sorting things out off-list.
I sat up, and spend three hours in the middle of the night, trying to figure
your problems, off-list, in mails between you and me. I asked for you to
please send me your Dictionaries subfolder, that I maybe could try to figure
why you are having your problems. So far, you did not send me the folder,
and I am unable to reproduce the mispronounciations you told me about. I
really don't want to clutter the list with our offline discussion. If I
could figure why you are having your problems, I would do what I can to
solve them. But then, please keep what is between the two of us, off-list.
In general, I can tell you, that the app does do nothing, except from
replacing what you put in the Original field, with what you put in the
Replacement field. Likely, your entries are somehow badly constructed. You
ask me to point you to where things are handled. Well, they are handled by
your entries. You ask me to let you know what the app does for meddling with
speech. Well, it takes what is to the left of the Equal-sign, look out for
occurrances of it in the string sent to the synthesizer, if found, replaces
it with what you have to the right of the Equal-sign in your entry. That is
the simple nature. It does not do anything but that. But it does so, even
down to detail.
If for instance, you have an entry like:
iv=4
, it will replace the two letters i and v, even if they are part of a word.
That is, even in words like "active", "activate", "massive", "positive" and
"ivona". They all hold the iv combination, hence all are being replaced. If
you want to have the IV only replaced when it is denoting a Roman number 4,
you could either enclose the Original string in spaces, or try changing the
entry to look like this, and note the writing here:
IV=4
.. This will make the replacement only to take effect, whenever an
upper-cased version of the combination occurs, due to the Extended
Dictionary being case-sensitive. The amount of trouble you reported, really
calls for me to have to see your Dictionary files, as I have a feeling
something must have got wrong on your system. But we will have to deal with
that off-list, and rather get back here, when things have been sorted.
For your info, the whole section, holding close to ten chapters, entitled
"Making Your Own Entries", found in the User's Guide of the Extended
Dictionary - it is all dealing with how the entries will affect your speech.
In addition, chapter 11 of the User's Guide, entitled "How Does A
Replacement Actual Take Place", also will be a good thing to read. You ask
me to give details. OK, There you have a long line of details. There's been
spend many hours and days, in writing these chapters. Not really sure, what
more info I possibly could provide. If you have constructive ideas, I am
listening.
Regards,
----- Original Message -----
From: "Orlando Enrique Fiol" <[email protected]>
To: "David" <[email protected]>; "Neville" <[email protected]>; "GW
Micro" <[email protected]>
Sent: Saturday, September 01, 2012 5:07 PM
Subject: Re: new extended dictionary
At 10:28 AM 9/1/2012, David wrote:
>All the date handling ever processed by the extended Dictionary, is based
>on
>entries in the Voice-Specific Dictionary, of any voice. In the English
>Dictionaries, the American Date section, is rather extensive. I won't
>deal
>in details, with the reason for it's being this extensive.
If you want us to use this app, you owe us more information about why
these mispronunciations are happening and how they can be corrected. As I
told you in private email, I have punctuation signs being misspoken or
spoken at a higher pitch, short vowels that are pronounced as though there
were three or four in a row, suffixes like "ING" or "ED" being separated
from words, even final letters being voiced at the ends of words. This has
happened with and without my dictionary entries being added to your basic
dictionary. I combed through the entire shipped and basic Dectalk
dictionaries and found no references to word fragments, vowel sounds or
phonemes. Yet, the synthesizer is obviously receiving input from somewhere
in your app that is changing how it pronounces even the most basic sounds.
You can either deal with this problem constructively or at least give us
the ability to have your app relinquish dictionary processing for any
synthesizer we choose and then kick in when we choose one of its listed
synthesizers. It would also help for your app not to insist on
rebuilding/reloading deleted dictionaries.
>But if you open the Voice-Specific Dictionary of an English voice,
you will find the section
>spanning from line 169, way down to line 1891. It is introduced with the
>Heading "; **** American Dates:", and ends where the next heading is
>reading
>"; **** Other Date Definitions:".
Yes yes, but none of those sections address the mispronunciations I'm
talking about.
>To get rid of the American way of date handling, you will have to remove
>this whole section. That is the chunk of lines, between the two above
>mentioned headings. Once removed, your Dictionary should handle the
>non-American ways of date notations properly. Note, this has to be done
>on
>each Voice-Specific Dictionary, for the voices you want not to handle
>American dates. The benefit would be, you could have one voice handling
>American dates, and another voice which handles dates in the non-American
>way. Switching voices, you then also would be switching the date handling
>automatically.One of the big tayloring possibilities that is part of the
>extended Dictionary project's intensions.
>
If only your assertion were true! No matter what we delete in the basic or
shipped dictionaries, the default information will reappear upon relaunch.
I tried all sorts of combinations: leaving one or both dictionaries blank,
filling them only with my entries and deleting all of yours, etc. All
yielded the same mispronunciations.
Orlando
If you reply to this message it will be delivered to the original sender only.
If your reply would benefit others on the list and your message is related to
GW Micro, then please consider sending your message to [email protected] so
the entire list will receive it.
GW-Info messages are archived at http://www.gwmicro.com/gwinfo. You can manage
your list subscription at http://www.gwmicro.com/listserv.
