asterisk users - Aug 2007 - Pre-recorded first and last names audio database

Hi!

My application needs to look up (by spelling) the first and last names of a
person and then insert the corresponding pre-recorded audio file to
personalize the message. E.g. "Hi, John Brown. Your book is due back at the
library." So I need "John" and "Brown" in audio files
along with LOTS of
other names -

Do such databases of sound files already exist or do I have to record my
own? I'm not sure how many first and last names I'd have to record but
it
seems like thousands for both genders first names and then thousands more
for last names to cover a significant proportion of the people in the USA -

Any and all help appreciated!

Thanks,
John

On 8/4/07, John Vogel <johnv at comcast.net>
wrote:
>
> Hi!
>
> My application needs to look up (by spelling) the first and last names of
> a
> person and then insert the corresponding pre-recorded audio file to
> personalize the message. E.g. "Hi, John Brown. Your book is due back
at
> the
> library." So I need "John" and "Brown" in audio
files along with LOTS of
> other names -
>
> Do such databases of sound files already exist or do I have to record my
> own? I'm not sure how many first and last names I'd have to record
but it
> seems like thousands for both genders first names and then thousands more
> for last names to cover a significant proportion of the people in the USA
> -

I haven't seen something like this, but if you figure it out, I'd like
to
know. There's a piece of software called HouseCalls that reminds people of
appointments. The proprietary software prompts the person setting up the
automated reminders to record each name individually. In the beginning, it's
a bear, but over time, it gets better.

I guess something in Asterisk would have to do the same, right? I mean, a
general list of "John Jon Jonh" in some person's voice, and the
rest of the
prompt in another, wouldn't be much better than having Festival say the
name, would it?
-------------- next part --------------
An HTML attachment was scrubbed...
URL:
http://lists.digium.com/pipermail/asterisk-users/attachments/20070804/64ea1ff5/attachment.htm

On 8/4/07, John Vogel <johnv at comcast.net>
wrote:
> My application needs to look up (by spelling) the first and last names of a
> person and then insert the corresponding pre-recorded audio file to
> personalize the message. E.g. "Hi, John Brown. Your book is due back
at the
> library." So I need "John" and "Brown" in audio
files along with LOTS of
> other names -
The only good solution to this problem is getting the "customers" to
record their names when their records are created. This could be done
at the library itself, if that's who is doing this, during a signup
process.

NO TTS will be able to do this. The reason is that names can be
spelled the same and pronounced differently depending on the person.
Even some first names like Stephen have at least two ways of being
pronounced.