Re: [pygame] Sound to String

["FT" <chester_lab@xxxxxxxx>]

    I was given NVDA weeks ago and you can download skeleton drivers but
like most places they are very short on documentation. Below, at the bottom
of this message is the response I got last night. I have not looked at it
yet, but I am new at this and documentation is a must in order to see what
they have done. So the links are below at the bottom and anyone who looks at
the sources it would be a great help to develope such an engine. I
understand that Linix and EMax Speak has gotten much better but I still use
windows and would like to  keep it that way for now.

    The NVDA project does have a speech/screen reader for windows, but it is
only object oriented and does not use the screen, just the objects. So
anything spoken with color only is lost. Which prevents you from knowing
such things as Read, Attachment, UnRead when looking at your Outlook Express
mail.

    If anyone can look at the drivers and such with this project and give me
some info on it would be a great help.

    When talking with the project person he told me he had to learn it on
his own, so he was not going to give me any help. In other words, do it
myself, but "Open Source?"

        Bruce


----- Original Message -----
From: "Laura Creighton" <lac@xxxxxxxxxx>
Sent: Thursday, February 28, 2008 5:38 AM
Subject: Re: [pygame] Sound to String


In a message of Thu, 28 Feb 2008 04:25:53 CST, Luke Paireepinart writes:
>Ian Mallett wrote:
>> You could always record the sound you want to play, but actually
>> rendering sound from text might be difficult.  If anyone knows, I'd
>> like to know too.
>Rendering sound is a pretty common operation called Text-to-Speech.  I'm
>sure you could find an open-source C library for this and wrap it with
>ctypes.  If there isn't one, someone needs to start one immediately.
>If anyone takes up wrapping this, let me know, cause I'm interested
>too.  Also, if you don't know how to wrap it but you find a suitable
>library let me know as well.
>Also it's quite possible there's already a Python library for TTS.
>I haven't done any research into TTS in about 5 years.

The University of North Carolina wrote one.
http://www.cs.unc.edu/Research/assist/doc/pytts/

Laura



On 28 Feb, FT <chester_lab@xxxxxxxx> wrote:

> I am a new programmer and using python to learn how to make my own
> games and such with a built in text to speech.

> I read that you wrote it in C and python is a higher level of C and
> easier to use then C. I hope you can help or give me the tools or
> where to get the tools/drivers to connect the speech to the sound
> card.

Unfortunately I have not used Python and I don't know anything about it.

I know that the NVDA screen-reader project:
http://www.nvda-project.org
uses Python to use eSpeak, so you can probably find some information
and examples there.

There are two ways to use eSpeak on Windows.  NVDA uses both of these,
as alternatives.

1.  The eSpeak API.  This uses eSpeak as a DLL. The eSpeak API is
defined in file: speak_lib.h.  A project file to make the eSpeak DLL is
included in the directory platforms/windows/windows_dll in the eSpeak
download.

or
2. The Microsoft SAPI5 interface.  This provides a common interface to
different text-to-speech engines.  Details are available in the
Microsoft Speech Development Kit ("SDK") which you can download from
the Microsoft website,
http://www.microsoft.com/downloads/details.aspx?FamilyID=5e86ec97-40a7-453f-
b0ee-6583171b4530&displaylang=en