same word, diff. pronunciation

[john.goley]

Jim,

How do you get a voice to read words spelled the same, and get them correct within the grammtic sturcture?

Such as:
I would like to "read" that book.
I have already "read" that book.

Or

The "wind" is blowing through the trees.
I am going to "wind" up a ball of string.

Can this be done? Or is this wishful thinking?

[PHenry1026]

Greetings John,

The homographs (read and wind) and the examples you provide can easily be made to pronounce to correctly using regular expression (or simpler substitutions) in the Pronunciation Dictionary Maintenance. In your first read example, you would respell read as reed. In your second read example, you would respell read as red.

Homographs such as wind require phoneme usage to pronounce the word properly. Only AT&T Natural Voice allow you to use phonemes from its dictionaries on the respell line. So if you are using AT&T NV, the wind examples you provided are as easily corrected as the read examples.

However, if you are using non-AT&T voices, you have to approximate the phoneme sound for wind and use this approximation on the respell line in the Pronunciation Dictionary Maintenance.

So the answer to your question is that most experience TA users have many dictionary entries that handle the examples you provided.

[john.goley]

Percy

I have been using the Pronunciation Dictionary Maintenance for years but have never gotten it to take the same word, but with different pronunciation within the grammatical structure of a sentence, when the same word is within a single passage, paragraph or book, and get it to say it correctly.

I tried again what you suggested with "read". I created two respells for read, one as "reed" and the other as "red".

When the "reed" respell was first in the dictionary order I got the following:

I would like to read ("reed") that book.
I have already read ("reed") that book.

When the "red" respell was first in the dictionary order I got the following:

I would like to read ("red") that book.
I have already read ("red") that book.

I created a second dictionary and put the respell "reed" in it, and the first dictionary had the "red" respell. If the "reed" dictionary was in the first order I got:

I would like to read ("reed") that book.
I have already read ("reed") that book.

If the "red" dictionary was in the first order I got:

I would like to read ("red") that book.
I have already read ("red") that book.

In other words I am still having the same problem that I have always had. To take the same word (read), but with different grammatical structure within different sentences, when that same word (read) is within a single passage, I cannot get TA3 or the voices I have (NeoSpeech Paul or Bridget or for that matter even Microsoft Sam) to pronounce them as the grammar requires.

Your explanation sounds like it should be easy to get what I want, but I cannot make it work. What am I missing?

John

[PHenry1026]

Greetings John,

Lets start fresh; the first thing you need to do is delete the entries you made for reed and red noted in your comments above.

There are two ways you can get the correct pronunciation you want. The first uses regular expression and is the best solution.

Method 1

Enter the Pronunciation Dictionary Maintenance and enter the following new dictionary entry at the top of your dictionary using Text Matching: Regular Expression



(?#Reed)(?m)(?:^|\s|['"‘“(]|\p{Pi}|\p{Ps}|\p{Pd})I would like to +\Kread(?= +that book[’”\p{Po}\p{Pe}\p{Pf}]{0,2}(?:\s|$))


On the Respell line enter

reed


Next click OK to save the new entry


Using the example above enter next Regex/Respell combination:



(?#Red)?m)(?:^|\s|['"‘“(]|\p{Pi}|\p{Ps}|\p{Pd})I have already +\Kread(?= +that book[’”\p{Po}\p{Pe}\p{Pf}]{0,2}(?:\s|$))


red



Exit the Pronunciation Dictionary Maintenance Dictionary and test:


I would like to read that book.

I have already read that book.




Method 2 (Not Recommended)


Enter the Pronunciation Dictionary Maintenance and enter the following new dictionary entry at the top of your dictionary using Text Matching: Simple Text



I would like to read that book

On the Respell line enter

I would like to reed that book



For the second entry again using Text Matching: Simple Text


I have already read that book.

On the Respell line enter

I have already red that book.





P.S. Since whenever "to" precedes "read", read is always pronounced as reed, the first regex above can be generalized as follows:


(?#Reed)(?m)(?:^|\s|['"‘“(]|\p{Pi}|\p{Ps}|\p{Pd})to +\Kread(?=[’”\p{Po}\p{Pe}\p{Pf}]{0,2}(?:\s|$))


On the Respell line enter

reed

or using Simple Text entry as follows:

to read

On the Respell line enter

to reed

[john.goley]

Thank you Percy

I got it to work with read, wind, and lives! But wow, this means you must stop and create this regex expression every time you hear a mispronounced word. Your P.S. at the end really simplifies it but is this the only way?

I have used TA since version 2.1 and tried to get words to pronounce correctly using only simple text. I finally just gave up and let my brain correct the pronunciation as I heard it. But now I am trying to create mp3 files for many books for future use and to share with friends. I never investigated regular expressions before but now it seems that I need to get into Regex.

Did you use RegexBuddy to write these expressions? Or do you know Regex so well that it’s all in your head?

I am disabled with a huge chronic pain syndrome such that I am in bed 50-70% of my waking hours. It is hard to hold a book, sit up, or do much of anything; so much of my time is spent with TA3 listening to news, stories, books, etc. Anything that I can get in a supported text format has kept me sane.

Thank you so much for TextAloud and this help with Regex

John