Would like to find a way to interface with the Speech Engine direc

[mthom]

Hi,

I have already in the past developed a dictation/transcription telephony
application on top of Dialogic API and recently on top of HMP (Global Call
library). I would like to support voice recognition of voice commands in my
telephony application. I'm currently evaluating Speech Server 2007 and I
found that it is a great product for general IVR purpose but unfortunately it
seems not to be adequate for dictating and playing long dictation and for
easily giving to the user a full control of the voice cursor example during a
rew, ff, insert ... Because HMP and global call API from Dialogic seems to
give me more control on the audio using SIP and RTP. I would like to see if
it's possible some how to have simultaneously many different user ports that
just communicate with the Speech Server 2007 Engine by feeding it with voice
buffers and getting back text and semantic response based on the Speech
Server grammar builder.

Any suggestions will be appreciate?

Thanks

[Ahmed Stewart [MSFT]]

Unfortunatelly, MSS 2007 does not currently support dictation and there is
currently no way to access the engine directly.

--
Ahmed Stewart
Microsoft Office Communications Server 2007, Speech Server
--
This posting is provided "AS IS" with no warranties, and confers no rights.
Use of included script samples are subject to the terms specified at
http://www.microsoft.com/info/cpyright.htm

"mthom" <mt...@discussions.microsoft.com> wrote in message
news:347DBCD0-026C-471A...@microsoft.com...

[mthom]

Thank you for your prompt response. It will be really great for Speech Server
to provide developer full control for recording and playing back long
dictation. This is for me the most important thing that is missing from
Speech Server in order for me to port my existing telephony dictation
application on it. I hope Speech Server will target dictation market on
future release. We have many Legal and Medical customers around the world who
are using our dictation system.

Martin

[George]

Yes, likewise, I would like to see this feature implemented.

Incidentally, Martin, I am in the medical field, and would love to find out
what speech applications you have done. You can email me at drg...@yahoo.com

George

[Marshall Harrison [OCS]]

While it doesn't do dictation it does have a RecordMessage activity that
allows the user to confirm the recording, start over or playback the
recording.

--
Marshall Harrison
Microsoft MVP Office Communications Server
Gold Systems Inc.
E-mail: mars...@gotspeech.net
Web: http://GotSpeech.Net
My Blog: http://gotspeech.net/blogs/marshallharrison/default.aspx

"mthom" <mt...@discussions.microsoft.com> wrote in message

news:719B24B8-CFE7-4D5D...@microsoft.com...

[mthom]

Yes but unfortunately this activity is a bit primitive for us. I would like
to get better control of the recording so that I can easily at runtime do
basic operations such as rewind, forward, insert new voice buffers or
overwrite voice at specific position in existing dictation voice file,
cancel/delete part of the dictation ... This activity seems ok for recording
short memo, voice mail but for recording long dictation in legal or medical,
dictation could take up to 1 hour and more. Of course during the dictation we
would like to recognize voice command such as: Pause dictation, Rewind
Dictation, Record ... We also do not want when playing back the dictation to
ear back the voice command recorded as part of the dictation.

[Steve Weavers]

A possible alternative would be to use the Speech engine supplied with
Vista. The problem here is that there are no telephony acoustic models
provided. Does anybody know how the acoustic models supplied with Speech
Server 2007 (MS-1033-90-TELE) could be used with the Vista Desktop engine?

"George" <Geo...@discussions.microsoft.com> wrote in message
news:E2FA43E6-62C6-4F08...@microsoft.com...

[Andres Vettori]

Hi, there is a way to record not just the audio for the response but also the
prompt being played to the user?
We need to implement something similar to an "e-sign" mechanism and
therefore we need to be fully auditable by recording not only the answer but
also the question.

Thanks in advance,

Andrés G Vettori
MSCE/MSSD/MCT
http://weblogs.asp.net/andresv