[opendtv] Re: LCD unfit for 3D viewing ?
- From: Jeroen Stessen <jeroen.stessen@xxxxxxxxxxx>
- To: opendtv@xxxxxxxxxxxxx
- Date: Wed, 31 Aug 2005 13:24:48 +0200
Hi,
olho_avatar_i@xxxxxxxxxx (who is that ? Olivier ?!) wrote:
> We have in the family this heap of stereoscopic photographs on glass
> plates from my great grandfather, and i was thinking of scanning them
> for digital preservation, and considering the possibility of viewing
> them in 3 D with sequentiel display of left and right picture and
> synchronized glasses.
> Using a CRT, it is possible to have a high enough refresh rate for each
> eye to see a flicker-free picture (as only half the refresh rate will be
> available for each eye).
Right. If you set the TV to 100-120 Hz then you'll have 50-60 Hz to
each eye, and it will be enjoyable. The decay time of CRT phosphor
is short enough to keep the two images separated.
> But refresh rates are lower for LCDs.
Most LCDs don't allow addressing faster than 75 Hz, though there have
been some prototypes at 100-120 Hz. The faster speed is suggested for
improving motion portrayal, not for field-sequential stereo images.
> Now those are supposed to be flicker-free,
Yes, the typical LCD is entirely flicker-free because the light valve
is open for the entire frame period. The image is not interrupted by
black, so there won't be any flicker. But this is bad for the motion
portrayal, as it leads to motion smear on the retina of the tracking
eye. Proposals to cure that (scanning or blinking backlight, black
field insertion) will lead to flicker again, and then you need 60-75
Hz refresh rate to suppress that. The only remaining alternative is
120 Hz refresh rate and motion-compensated up-conversion on the video,
and then you get twice as good motion portrayal without any flicker.
> but as they also use some kind of internal scanning for refresh,
Not really. Each LCD cell acts as a sample-and-hold capacitor, like
a Dynamic RAM, and keeps the transmission factor constant between
refreshes. This is entirely passive.
> i was wondering if there would not be perceptible problems at
> 30 Hz, for example?
In the first place, if you would use optical shutter glasses to
allocate 16 ms to one eye and then 16 ms to the other eye, both eyes
would perceive horrible 30 Hz flicker. This is not because the
display itself flickers (though the naked eye would see the edges of
objects flicker), it is because of the glasses !
In the second place, the LCD depends on two successive images being
identical, because it uses polarity inversion of the display drive
for avoiding DC voltage. DC voltage can cause image retention, and
inversion of the DC to 30 Hz AC prevents that. If you want to show
two different images then the two voltages do not add up to zero DC
anymore. Even line flicker from interlacing can cause burn-in.
So you would still have to run the LCD at 120 Hz, and apply some
other tricks, to get it as good as a CRT again. Not nice.
> Also, the display must me fast enough for two successive pictures
> to be cleanly separated.
Which it is not. Depending on the design, the LC molecules take
between 2 and 50 ms to rotate to a new state. This is typically too
slow for keeping two images separated. Also the TFT transistors and
line electrodes have too large an impedance to drive the cells to
the required voltage in a very short time. It won't work (yet).
> Any opinion ? Anyone tried something like that ?
There do exist stereo LCDs, but they do not apply temporal (frame
sequential) multiplexing. Instead they make use of the high resolution
that comes with LCD technology, allocating the odd columns to one view
(eye) and the even columns to the other view (eye). A parallax barrier
is placed in front of, or behind, the LCD, and this keeps the 2 views
separated. Sharp sells such a display, where the parallax barrier can
also be switched off to get a high-resolution 2D display back again.
You need to put and keep your two eyes in a narrow sweet spot in order
to have separated images, and this can be quite uncomfortable !
This principle can be extended to multiple views (e.g. 9), but then
the parallax barrier should be replaced with a lenticular lens array
to keep the efficiency up. Or a reflective parallax barrier is used
(behind the LCD). Multiple views give two advantages: the head position
becomes less critical, and you get a "holographic" effect, you can move
your head to look around objects. The biggest disadvantage is that a
lot of spatial resolution is sacrificed (e.g. a factor of 9).
Philips has given demonstrations on the CES and SID exhibitions.
Obviously, mono and stereo source material must be converted to
multiple views. That is an interesting challenge (for us).
Best regards,
-- Jeroen
+-------------------------------+------------------------------------------+
| From: Jeroen H. Stessen | E-mail: Jeroen.Stessen@xxxxxxxxxxx |
| Building: SFJ-5.22 Eindhoven | Deptmt.: Philips Applied Technologies |
| Phone: ++31.40.2732739 | Visiting & mail address: Glaslaan 2 |
| Mobile: ++31.6.44680021 | NL 5616 LW Eindhoven, the Netherlands |
| Pager: ++31.6.65133818 | Website: http://www.apptech.philips.com/
|
+-------------------------------+------------------------------------------+
----------------------------------------------------------------------
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:
