Re: Media queries for stereoscopic 3D displays?

On 09/03/2012 06:44 AM, Florian Rivoal wrote:
> 
> On Sat, 01 Sep 2012 11:10:14 +0200, Douglas Crosher
> <dtc-style@scieneer.com> wrote:
> 
>> Has there been any consideration of using media queries to detect
>> stereoscopic 3D displays?
>>
>> For example: anaglyph variants, interlaced, checker-board, left or right
>> image of a parallel or cross eye stereoscopic view.
>>
>> Perhaps it would be best to just target a left and right view and leave
>> it to the display software to convert this to a suitable format, but
>> knowing a display needed anaglyph, interlaced, or checker-board formats
>> would make it possible to serve up suitable images.
> 
> I don't recall such a discussion taking place in this group. Fundamentally,
> I think being able to use media queries to provide specialized styling
> to 3D displays is worth pursuing, but I am not sure the timing is good.
> 
> As far as I can tell, 3D screens are still relatively rare, and there
> is no generally agreed way for browsers to deal with them. If there were,
> it would definitely be useful to provide a way for authors to detect
> if their page is being rendered in such an environment and to act on
> it. But until it is well established what an author could do to take
> advantage of being in such an environment, I don't think introducing
> such a media query will help, or that it can be done well.

3D displays have become quite common. It has been reported that 3D TV
sales are around 30% and with smart TV sales are expected to be around
60% so this is a large market without support.  3D monitors are now also
a commodity item: there are 3D interlaced monitors that use only light
weight polarised filter glasses, lenticular and parallax displays that
do not need glasses, there are monitors that work with shutter glasses,
and head mounted displays.  There are 3D laptops, and 3D mobile phones
have been very successful in some markets.

There does appear to be a need to consider this market.  For example,
using CSS to style the UIs of these devices, or to write Web Apps that
take advantage of 3D displays, or to customize websites to visitors with
these devices.

A large body of current CSS styling uses focus, photometric stereo (drop
shadows etc), highlight cues, focus cues, transparency clues, among
other effects, to create a feeling of depth.  These would appear to be
useful and/or desirable attributes of a UI.

Keep in mind that consideration of stereoscopic 3D displays could have a
large impact on existing and developing proposals, and could help guild
developments.  The current body of 'flat' 3D effects may age quickly if
they can not be adapted to stereoscopic devices.   Even on 2D devices it
would be handy to be able to have a consistent photometric effect under
the control of the user, and to add wiggle transition effects that
reveal depth.

Simply adding some media queries could help specific devices by allowing
them to present different images based on the device mode - switching
between 2D mode and 3D mode.  For example the media queries could be
used to present the left or right image of icons when rending, and to
adjust positioning of elements for the left and right view to create depth.

Even the CSS 3D transform could be adjusted using media queries for the
left and right view to create stereoscopic depth - although for
consistency of the perspective with the stereoscopic disparity a more
specific design would help.

There may be potential for extensions to work with WebGL too, so that
WebGL 3D content could be presented in stereoscopic 3D within a page.

Youtube now supports a range of stereoscopic 3D modes and extensions
that communicate the 3D display technology could help improve the
usability of such websites.  A lot of smart 3D TVs will likely support
youtube, and this is an example of a real need.

It is desirable for the stereoscopic disparity to be user adjustable to
suit the display size and viewing distance so it may be better to
support a flexible approach rather than just a left/right flag.  Perhaps
an extension to calc() would be appropriate too.

If there is no interest on this mailing list then some help pointing me
to the process of creating a working group to examine ideas and develop
proposals would be appreciated?

Regards
Douglas

Received on Monday, 3 September 2012 00:17:56 UTC