W3C home > Mailing lists > Public > public-media-capture@w3.org > May 2015

Re: Filtering of enumerateDevices() results

From: Joe Berkovitz <joe@noteflight.com>
Date: Thu, 21 May 2015 08:20:30 -0700
Message-ID: <CA+ojG-ZT2Yac=P0OYu88y2timqDrh-Lbq6-FhLdzLYcsBeu-KQ@mail.gmail.com>
To: Harald Alvestrand <harald@alvestrand.no>
Cc: public-media-capture@w3.org
...connecting the above dialogue to my original question in this thread...

- If the labels can simply reproduce what the OS is willing to tell us
about a device's human readable name, that's good enough for the Audio WG.
We understand there will be limitations.

- If any other device information that is available from enumerateDevices()
is absent because the OS couldn't tell the UA, so be it. Let the OS APIs
improve over time, but let's not cripple the browser API in advance. Web
developers are lazy -- in a good way. I think they are not likely to go to
any effort to scrape manufacturer info from websites, especially for
information like channel counts and sample rates that are not exotic "spec
sheet stuff" but things that WebRTC and Web Audio need to know in order to
function at a basic level.

On Thu, May 21, 2015 at 1:18 AM, Harald Alvestrand <harald@alvestrand.no>
wrote:

> Den 21. mai 2015 04:53, skrev Jan-Ivar Bruaroey:
> > On 5/19/15 10:29 AM, Joe Berkovitz wrote:
> >> 2. How would the WG feel about including more "filtered" information
> >> in the MediaDeviceInfos returned by enumerateDevices(), other than the
> >> device label
> >
> > I'm going to take a stab at guessing that the original intent with
> > 'label' was for web-devs to look up the info they care about themselves.
>
> The original intent with "label" was to enable web-devs to present a
> device picker interface that was no worse than what one could do in
> browser chrome.
>
> The engineers who built this have a great deal of experience with just
> how hard it is to say something sensible about a device - presenting the
> USB device IDs (for things that present themselves as USB devices) is a
> lot more precise than labels, but not very user friendly.
>
> >
> > E.g. label == " Logitech HD Pro Webcam C920" -> look in table of what I
> > care about scraped from
> > http://wiki.oz9aec.net/index.php/Logitech_HD_Pro_Webcam_C920
> >
> > I.e. an end-run around workgroups and browsers having to formalize every
> > spec sheet known to humankind.
> >
> > I should say I see evidence this isn't always working so well, e.g.
> > android phones seem to give nonsense labels like "Camera front", and I
> > see a lot of microphones called "Built-in microphone" or "".
>
> There are a lot of nonsense labels around, yes. This reflects directly
> the state of play in device drivers.
>
> > How about a compromise like 'modelNumber', or maybe come up with better
> > rules for label?
> >
> > Also from a practical standpoint, not every platform is chock full of
> > info here. OSX for example seems surprisingly tight-lipped about its
> > hardware.
>
> Which is the reason for the phenomenon you describe above.
>
>
>
>


-- 
.            .       .    .  . ...Joe

*Joe Berkovitz*
President

*Noteflight LLC*
49R Day Street / Somerville, MA 02144 / USA
phone: +1 978 314 6271
www.noteflight.com
"Your music, everywhere"
Received on Thursday, 21 May 2015 15:20:58 UTC

This archive was generated by hypermail 2.4.0 : Friday, 17 January 2020 16:26:33 UTC