Re: The API hooks related to hooking up with the capture system.

A bit better, yet short description [1]:


interface MediaStreamRecognizer
{
   attribute MediaStream input;
};

SpeechInputRequest implements MediaStreamRecognizer;


If the user agent supports a MediaStream API, it may expose input
attribute in SpeechInputRequest objects. Initial value for the attribute
is null.

When input attribute is null, SpeechInputRequest object should use the
data from the default microphone as its input. If input attribute is
non-null, SpeechInputRequest object should read the data from the stream.





-Olli


[1] Is it not yet clear which MediaStream API to use.
It can be for example something WebRTC WG defines
or
http://www.whatwg.org/specs/web-apps/current-work/multipage/video-conferencing-and-peer-to-peer-communication.html#mediastream
or
http://hg.mozilla.org/users/rocallahan_mozilla.com/specs/raw-file/tip/StreamProcessing/StreamProcessing.html
or probably a merge of all those three.




On 07/20/2011 08:49 PM, Olli Pettay wrote:
> Hi all
>
> to hook up the generic capturing API to speech API we should
> reuse Stream API. So the addition to the Speech API would be simple:
>
> interface SpeechInputRequest
> {
> ...
> attribute MediaStream input;
> };
>
> Endpointing should, IMO, be handle by the Speech API, so that we can
> use the generic media streams.
> If input is not set, Speech API should use whatever capturing mechanism
> the UA provides and .input attribute would be null in that case.
>
> Scripts can get the Stream using for example the API proposed in
> http://www.whatwg.org/specs/web-apps/current-work/#video-conferencing-and-peer-to-peer-communication
>
>
> .input would work also well with
> http://hg.mozilla.org/users/rocallahan_mozilla.com/specs/raw-file/tip/StreamProcessing/StreamProcessing.html
>
>
>
> If we want to support <reco>, the non-default input stream should
> be set using scripts.
>
>
> -Olli
>
>

Received on Friday, 29 July 2011 15:34:01 UTC