Open Sapi - RAW PCM

[Thomas Lloyd]

Hi All,


Right I have been a busy bee and now have some questions about Raw Audio
Streams.

At the moment I am directly reading the RAW PCM data from SAPI. I can do
what I want with it now. I have a number of questions though. 

Do I need to be sending speechd a wav header so it can know what it is
dealing with or can I just send it the raw data. I think I will need to
send the wav header but a more in-depth explanation might help having to
recode later on. 

The second question is around how to implement the transfer of the RAW
data. I also have the ability to send stream rendering tracking
information that speech dispatcher likes. I was thinking that I could
interweave audio data and the tracking feedback into the same stream to
help with sync issues. Can people see a problem with that? I am planning
to have two communication channels between client and server one for
normal messages and the other for audio/tracking data.

The third question is about audio data transfers. I would like to get
the size of the audio data to be as small as possible so when this gets
running across the net it will function using less bandwidth. Can I
supply speech dispatcher with say ogg formatted audio data or Mp3?

Sorry I am asking so many questions but I have not got the inside
knowledge of speech dispatcher yet. 

Lastly I asked about getting the open-sapi configuration file for the
module included in the speech dispatcher GIT repository. Can I ask how
to go about that again, also Luke is there a possibility that it will
make it into Lucid if I ask really nicely?

Tom