[opendtv] 3D at NAB
- From: Mark Schubin <tvmark@xxxxxxxxxxxxx>
- To: opendtv@xxxxxxxxxxxxx
- Date: Mon, 19 Apr 2010 13:58:46 -0400
Thanks for the comments.
There were several no-glasses 3D displays. You might be talking about
the one from Alioscopy. There was also a range at Franhofer: two
8-view, one 9-view, and one 5-view. To me, the 5-view looked best.
In the 3D field, there wasn't 3D at every booth, but there WAS 3D at
every aisle. 3D "products" and "solutions" ranged from dual coax
(Belden 1694D) to Harris claiming to offer "end-to-end 3D" when they
sell nothing on either end (acquisition and presentation).
Two 3D solutions that stood out to me were Pace-Fujinon and Screen
Subtitling, both of them because they recognize the present 3D business
environment: 3D is not ready to be the dominant form of TV, but it is
nevertheless being pursued by a number of programming channels,
including DirecTV, Discovery, ESPN, and Sky. So here's what those two
groups came up with.
In sports, 3D has been shot very differently from 2D. But how long can
content providers afford to do dual production, with dual equipment &
crews? And what happens when the camera position desired for 3D happens
already to be occupied by the camera position for 2D? So, in a tower at
Fujinon's booth, there was the Pace-Fujinon Shadow. It starts out as an
ordinary 2D camera position: long lens, big camera, etc. Then, on top
of the big 2D lens (underneath is an option), Pace rigs two 3D cameras &
lenses. That takes care of the 2D-&-3D-in-one-position problem, but
there's still the operator problem. So the two rigs are tied together
in a range of possible configurations. In one, a zoom of the 2D camera
will be matched by the 3D cameras until they run out of range and the 2D
continues; in another, the zooms are scaled, so the 2D zooms faster than
the 3D.
Is it ideal? Absolutely not. But, for the present dual-shooting
environment, it's at least economically feasible. Kudos to them.
Screen Subtitling addressed the 3D-graphics issue. If a graphic
occludes a part of a 3D picture but is set in the screen plane, then, if
the 3D comes out in front of the screen, the graphic is difficult to
process visually due to conflict between the occlusion cue (most
powerful at all distances) and the stereoscopic depth cues (second most
powerful at short distances). MANY companies at NAB showed systems for
generating 3D graphics that avoid that problem, but, again, that's new,
expensive equipment for what will initially be tiny audiences. So
Screen Subtitling made use of the fact that common distribution of 3D
signals is in a side-by-side format. If the subtitles are added in the
side-by-side domain, then they're simply a long line of repeated text,
with depth controlled by the spacing between the repeats. Existing
equipment works just fine. Kudos to them, too.
There were MANY other terrific 3D innovations at the show and a number
of 3D cautions. In a 3D remote truck, for example, entering the control
room one saw horrendous wrong-eye ghosting even when wearing glasses,
leading the truck crew to instruct visitors to sit down. Some of the 3D
display technologies being exhibited had very restrictive vertical
viewing angles. What works for a director sitting at the front bench
doesn't necessarily work for a producer standing behind the rear bench.
Another caution was nicely addressed in 3ality's dual-camera rigs. If
you shoot 3D for TV, eye divergence might not occur until, say 80 pixels
of positive parallax. But blow the same images up to a theatrical
screen, and 8 pixels might do it, a problem that won't be noticed by
anyone looking at a small monitor. 3ality indicates when there will be
a large-screen divergence problem even when no one can see it.
FYI, for those who couldn't make it to my Digital Cinema Summit 3D
keynote, it has now been posted on my website:
http://schubincafe.com
You may download it directly here:
http://schubincafe.com/files/2010/04/Schubin_2010_DCS_NAB.ppt
One caution: Although the charts in my presentation ARE the ones
presented by Professor Martin Banks at the HPA Tech Retreat (and used
with his permission), I had a discussion with him the night after my
presentation, and he made revised versions. According to the new
charts, vergence depth out to infinity might be comfortable at viewing
distances of about 3.2 meters and beyond. But Professor Banks points
out that the new charts are still based on assumptions, and more
research is needed.
TTFN,
Mark
On 4/19/2010 1:15 PM, dan.grimes@xxxxxxxx wrote:
"I didn't attend this year; but one of my Producer friends did.
Friday, I sat down with him and asked him about all this "3D" buzz.
He looked at me with a puzzled look. He thought the biggest thing was
acquisition using Digital SLR cameras such as the Canon 5D. (An
episode of Fox's "House" was recently shot using one 5D camera.)
"He saw a lot of camera accessory booths prominently featuring the
Canon 5D camera and their accessories.
"It's an interesting observation, as two news photographer friends who
attended mentioned the 3D as well. I guess folks in different fields
took away different perspectives." --Don Moore
Yes, your friend is quite correct that the DSLR was well represented
as a capture device. I didn't pay much attention to it except to say
that I was amazed at the sophistication of the accessories and lenses
that were displayed. I was surprised to see so many DSLRs hung on a
$200,000+ lenses.
But I thought 3D was more prevalent; but that could be because of
sheer number of displays everywhere.
By the way, one company (I wish I wrote down the name) was displaying
3D content on a screen that that was truly 3D without any eyewear.
The effect was fine except there was only depth from the screen
surface to infinity with nothing jumping out to the viewer. I did not
ask about the particulars. The effect remained as one walked to the
sides of the monitor but in stepped increments.
I am always interested in Mark Schubin's view of NAB. He always finds
such great things that are buried amongst the overly obvious.
Dan
----------------------------------------------------------------------
You can UNSUBSCRIBE from the OpenDTV list in two ways:
- Using the UNSUBSCRIBE command in your user configuration settings at FreeLists.org
- By sending a message to: opendtv-request@xxxxxxxxxxxxx with the word
unsubscribe in the subject line.
Other related posts:
- » [opendtv] 3D at NAB - Mark Schubin
