[Dtb-talk] Best tool for generating speech on a Mac?

[Greg Kearney]

Here are some assorted tips that I have developed over the lats few  
years.

In producing books I use mostly Alex due to its quality and the fact  
that it does not require licensing to use in digital talking books.  
That said it can produce some unexpected results, as can all TTS  
system. I would urge you and others to report these at once to :

accessibility at apple.com

One thing that could be done is for Apple to support SSML in the say  
command and in say for AppleScript. This would help a great deal  
indeed. You might also mention that to the address above.

It is possible to control the TTS system to a very fine degree with  
the embedded speech commands. more information can be found here:

http://developer.apple.com/documentation/UserExperience/Conceptual/SpeechSynthesisProgrammingGuide/Introduction/chapter_1_section_1.html

Short of this there are things that I do in producing general reading  
(non-technical) books that can be a real help.

1. I change all words that appear in all upper case letters to mixed  
case letters. I have an OpenOffice script that does this.

2. I change all case of roman numerals to arabic numerals.

3. I change all case of em and en  and double hyphens to a comma and a  
space.

4. Make sure that all paragraphs have any new lines removed.

Here is my much simplified, process for producing DAISY digital  
talking books. This assumes you have the latest version of DAISY  
Pipeline which is accessible to screen readers including VoiceOver.  
You will also need the odt2dtbook OpenOffice plugin. and OpenOffice 3.  
There is  similar DAISY export tool for MS Word under Windows.

1. Bring the text into OpenOffice 3, which is VoiceOver compatible.

2. Mark the major and minor heading with the Heading1 to Heading6 as  
needed. Remember they must be in number order Heading1 can be followed  
by another Heading1 or a Heading 2 but not a Heading3

3. I also do thing like footnotes and such but that will depend on  
what your working on.

4. Export the file to DAISY.

At this point you could simply bring the resulting dtbook XML file  
into your stream and you would have all the navigation in place but  
let's move on and make a full text full audio book with DAISY  
Pipeline...

5. Open Daisy Pipeline and create a new Narrator project. Narrator is  
going to use the system default voice from the Mac.

6. Choose the dtbook XML file you created in step 4. Set the location  
for the book directories and other options.

7. Run the job. Depending on the size of the dtbook XML file this can  
take anywhere from a few minutes to many hours. I have a quad  
processor MacPro and the whole Bible take 10 hours to generate.

What you will get is two directories one named z3986 which is a DAISY/ 
NISO 2005 book and one called daisy202 which is the older DAISY 2.02  
format. Either one can be played in the stream and most other modern  
DAISY player.

Pipeline by default always keeps the MP3 files in playable order so if  
you took all the DAISY files and put them on a CD you could play it in  
a MP# enable CD player. Pipeline can also generate iPod style playlist  
so the books could be played as such in MP3 players and iPods.  
Pipeline can do tones of other things as well like make files for  
producing large print and braille books and so on.

Hope this is a helps.

Greg Kearney
535 S. Jackson St.
Casper, Wyoming 82601
307-224-4022
gkearney at gmail.com



On Feb 17, 2009, at 10:31 PM, T. Joseph Carter wrote:

> This might really be a question just for Greg, but what tool or  
> tools do people find effective for getting text into a form suited  
> to desktop and portable spoken form using a Mac?
>
> I generally work with two forms of text--plain text with little to  
> no formatting, which is fairly easy to work with, and PDF files,  
> which aren't so easy.
>
> In the case of the text files, I find usually the best solution is  
> to put them on the Stream and just use that.  I prefer Alex for  
> speech quality, but Alex is a total idiot who thinks he is clever  
> when it comes to handling punctuation.  Often he totally ignores  
> punctuation that is context-vital but misreads things he thinks he  
> understands how to "say correctly".  I can usually fix it--it's  
> plain text after all--but it's usually faster to put it on the  
> Stream since the Stream is a little more intelligent.
>
> In the case of PDF files, columns and hyphenation are the bane of my  
> existence.  I can sort of deal with the columnation, but the  
> hyphenation is just really annoying.
>
> It seems like a tool designed to get text into spoken form on the  
> Mac might be able to solve both problems, but I don't know of  
> anything.
>
> Any thoughts?
>
> Joseph
>
>
> _______________________________________________
> Dtb-talk mailing list
> Dtb-talk at nfbnet.org
> http://www.nfbnet.org/mailman/listinfo/dtb-talk_nfbnet.org
> To unsubscribe, change your list options or get your account info  
> for Dtb-talk:
> http://www.nfbnet.org/mailman/options/dtb-talk_nfbnet.org/gkearney%40gmail.com