W3C home > Mailing lists > Public > public-media-capture@w3.org > April 2013

Re: Overconstrained

From: Randell Jesup <randell-ietf@jesup.org>
Date: Wed, 03 Apr 2013 10:37:59 -0400
Message-ID: <515C3EC7.9020001@jesup.org>
To: public-media-capture@w3.org
On 3/27/2013 1:20 PM, Martin Thomson wrote:
> In a large proportion of the remainder, the user will be the one
> deciding which input device to use.  Most likely, they will have done
> so in the OS or chrome beforehand.
> For users, the single most important factor in determining which
> camera to use will be where it is pointed.  Applications too.  In both
> cases, this will probably be to the exclusion of all other factors.
> In the vanishingly small subset of cases where the above is not
> sufficient, it is not a problem for an application to ask for another
> camera.  I'm willing to bet that those cases are going to be
> specialized applications with better-informed users that are
> sufficiently motivated and informed to comply.
> Facing therefore becomes the most important constraint.  The sourceId
> constraint allows for stability of selection over time, especially for
> those cases where a default doesn't work.

I'm 100% in agreement on this.  As I said at the Interim (and others did 
too, including Martin), the most (and virtually only) interesting 
constraint for a user is "what is the camera pointed at".  Now for 
microphones, the question gets more interesting (attached mic on camera 
vs secondary mic), but once again it's a question of "where is the mic" 
not "what is the mic".

(Browser) UI's where the user selects cameras with images of what they 
can see have an advantage here.  For microphones, there are similar 
ideas possible.

Setting camera "constraints" (really settings) after acquisition is a 
separate question, but on getUserMedia() I'm wholly on-board with 
eliminating constraints (outside of SourceID perhaps initial *settings* 
(like preferred resolution -- not selection criteria)), or at most using 
them to order choices/provide feedback to the user (and I don't think 
there's enough benefit to the user to do that).

Randell Jesup
Received on Wednesday, 3 April 2013 14:40:11 UTC

This archive was generated by hypermail 2.4.0 : Friday, 17 January 2020 16:26:16 UTC