RE: [rtcweb] Use cases - recording and voicemail

 

> -----Original Message-----
> From: Harald Alvestrand [mailto:harald@alvestrand.no] 
> Sent: 22 August 2011 15:10
> To: Stefan Håkansson LK
> Cc: Ravindran Parthasarathi; Avasarala, Ranjit; Elwell, John; 
> rtcweb@ietf.org; public-webrtc@w3.org
> Subject: Re: [rtcweb] Use cases - recording and voicemail
> 
> On 08/22/11 13:19, Stefan Håkansson LK wrote:
> > On 2011-08-22 10:16, Ravindran Parthasarathi wrote:
> >
> >> ... Currently, let us have common understanding whether recording 
> >> usecase has to be added in RTCWeb or not.
> > Agree. And also _which_ recording use case(s) in that case. I think 
> > this is what John was looking for when starting the thread.
> >
> > My $0.02 says that we need to take some care before adding more and 
> > more usages and reqs - after all the schedules for the WGs 
> are quite 
> > aggressive.
> Indeed. When thinking about the recording use cases, I could come up 
> with a few interpretations:
> 
> - Record one media stream, as it arrives (the "voice mailbox" concept)
[JRE] This was not what I originally proposed, but emerged during discussion. From my point of view a mailbox is more likely to be centralized, not on a user's device, so the web application can divert calls to a mailbox using signalling means, without any impact on the browser. Some people might want to support a local mailbox, in which case browser support would be needed, but that is second priority, from my perspective.

> - Record a conversation: Outgoing and incoming audio and video in an 
> 1-on-1 or 1-on-N context (typically as permanent records of a 
> meeting or 
> conversation)
[JRE] This was what I was originally asking about. Seems to be important to have this capability from day 1, or at least an architecture that can be extended to do this at day 2.


> - Record the application screen as presented to the user, 
> together with 
> the audio tracks as presented to the user's audio device, but don't 
> bother with incoming audio from the user at all (game recording, for 
> instance - the "WoW movie" kind)
[JRE] Not high priority from my perspective.

> 
> In each instance, recording technology might include recording:
> 
>    - As one track, somehow mixed, ready for playback
>    - As multiple tracks recorded separately within a container that 
> requires special playback devices
[JRE] I think the container here is what is known as metadata, from a session recording point of view. That container can be a matter for the web application.

>    - As multiple tracks, recorded separately to separate objects
[JRE] Yes, although the possibility of mixing (e.g., the two directions of an audio stream) would be good to have.

>    - As streams sent to some remote recording entity, that may choose 
> one of the options above
[JRE] Yes, this is my primary concern, coupled with the "record a conversation" above.

> 
> I'd be happier about discussing which ones of these people find it 
> critical to support before we dive into figuring out how to 
> support them.
> 
> Send text!
[JRE] I was just about to - look out for separate postings for local and remote recording (I will leave others to proposed text for mailbox, if they require it).

John


John Elwell
Tel: +44 1908 817801 (office and mobile)
Email: john.elwell@siemens-enterprise.com
http://www.siemens-enterprise.com/uk/

Siemens Enterprise Communications Limited.
Registered office: Brickhill Street, Willen Lake, Milton Keynes, MK15 0DJ.
Registered No: 5903714, England.

Siemens Enterprise Communications Limited is a Trademark Licensee of Siemens AG.


> 
>                        Harald
> 
>

Received on Monday, 22 August 2011 14:41:51 UTC