Le 28/04/2014 14:24,
pierre.meindre(-at-)free.fr
a écrit :
Here is what I get:
On a 55" TV, 55 inches is the diagonal of the screen (about 140
cm) so if the screen is a 16:9 (and if I haven't made a mistake)
its width is 122 cm.
On a HDTV horizontal resolution is 1920 pixels so each pixel is
0.63 mm wide.
So 69 pixels are 44 mm on the screen not the 55 mm you are talking
about.
------------------------
Hi Pierre,
First a question : is it possible with SPM to change the
parameter of automatic mount (like Optimized anaglyph permit) to
obtain the right positive parallax pixels of space depending of
the monitor, TV, screen projection size ?
69 pixels is for a 55" 3D-TV (LG and PANA tested), in fact 79
pixels, but like TV "eat" 10 pixels it is reduce to 69 pixels.
On a 42" 3D-TV this become 93 pixels (in fact 103 - 10 pixels
"eat" by TV). But if 69 pixels fit also on a 42"(with 3D reduce a
bit), the reverse not. For a 65" TV 69 pixels will not fit.
The space of 55 mm is for projection screen, not for standard
3D-TV and I have to make a correcting on my picture bellow to
avoid any confusion :( I add in red "only"
It is wrong to think if the positive parallax fit for a monitor it
will fit too on any 3D-TV and projection screen.
3D movie in cinema will need a special encoding to permit to the
projector to adjust automatically the parallax positive (first for
small children eyes space, especially if it is a 3D movie for
children).
In add of that many 3D cinema and theme parks have too big screen
for the first spectator distance.
An example is the last demand I got from a screen installation
wanted a size of 11.26 x 5.20 meter for the FUTUROSCOPE, a 4m for
the Confluence museum in Lyon city (France) and a 4m for the
museum City of Sciences in Paris. All with too big screen : (
Now regarding your second question :
> on a 3D-TV of 55" verify your distance of "positive parallax"
between two homologous dots to infinity, don't must be over 69
pixels.
Can you explain us how you get this number of 69 pixels ???
It is not of my own but from a partner calculation making progress
since many years. He is exactly in the same way as a Russian
stereographer team editing critic of 3D movies (link edit by
Olivier in the SCF monthly revue under the French title "Le relief
ça fait mal aux yeux... On commence à savoir pourquoi" / "3D hurt
eyes... we starting to know why"). First with others partners
making 3D movies, Ophtalmic test, 3D screen mount in cinema, 3D
viewers in museum. We notice some children have trouble with some
of our pictures (still and motion). We notice also to reduce depth
and make more pop-up avoid children trouble, so the trouble was in
the positive parallax we an over depth limit giving similar effect
than diplopia (see the mixing I permit to do bellow for fun, sorry
Duke and Shahrokh... you can't see in the same time the eagle eye
and the background bridge... (or reversal) imagine now this
happen much more often to children than to adult because the
parallax mount is actually done for adult standard eye space of
65mm, not 55mm for children comfort.
But we had not a real knowledge of stereoscopic geometry. Than
this man call me by phone and explained to me his work came first
from the book "3D aerial photography" of Louis Phllippe CLERC
(1920) and from the book of a old director of the French "National
Geographic Institute "IGN" (using also 3D aerial photography to
built maps), his name was the General Hurault (also in the past
president of the French stereo club... if I'm right). The trouble
is all this work fall down with his dead and French stereo club
successors forgot it.
With the help of this new partner since two years, we discover
the stereoscopic geometry with precision and his Euclid equations
result and draws was similar to our practice result also similar
to the General Hurault and Clerc and the Russian team. When
several people (theory and practice) obtain very narrow the same
result that mind for my partners and me... could be the true and
we have to make attention to study this new partner work. But like
hard to understand (I'm not a mathematician) I propose him to make
my own draws, he made correcting. Now I offer free the result for
stereoscopic
up-grade and will be please everybody verify too. This man want to
stay discrete for moment but he is ready to answer with my
intermediary to any question from any 3D list.
Bellow is his answer in French to your question. I prefer let you
to,make the translation if somebody have interest :
----------------------------------------------------------------------------------------------------------------------
Le calcul du nombre de pixels pour obtenir 50mm entre
infinis
sur une TV de 55"
(0,871×55"x25,4mm/1" = 1216mm de largeur horizontale)
(pour une autre taille, seul le 55" est à remplacer par
l'autre taille dans le calcul ci-dessus !)
vient du rapport 1920pixels / 1216mm × 50mm = 79pixels,
Le calcul
pour obtenir "55mm" entre infinis homologues
donnerait bien 1920 / 1216mm × 55mm =
87 pixels, mais
hélas, sur l'écran LG de 55" qui a été utilisé, il a été constaté
que la réalité mesurée dépassait déjà 65 mm !
Comme si en réalité l'écran montrait moins de 1920 pixels de
large ! ... donc la TV mangeait des marges, donc le taux
théorique de 0.63mm/pixels évoqué par Monsieur Pierre Meindre ne
correspond donc pas à la réalité de la TV 55" utilisée.
Il a fallu ensuite faire des tests en diminuant progressivement
le nombre de pixels entre infinis homologues pour aboutir à
obtenir réellement 55mm sur l'écran.
Maintenant a-ton eu raison de faire la supposition que tous les
écrans de 55" réagissaient pareils ???
les clubs ou les professionnels ont-il fait une enquête ???
... c'est le moment que chacun indique :
- son
nombre de pixels habituels entre infinis homologues
lors
de la création de son couple
- fasse la
mesure réelle en mm
sur sa TV, mesure
entre ses infinis homologues
- et le fasse savoir
Va-t-on constater que l'on obtient une constante réelle en
Pixels/mm attachée à un format de TV ? si c'est le cas, on a eu
raison ; si ce n'est pas le cas, ce serait dommage ... car cela
voudrait dire que les fabricants d'écran de TV3D ne sont pas
conscients que cela est important ! mais les stéréoscopistes
devront trouver hélas un moyen de régler cela par un biais
quelconque en attendant ... c'est ce qui a été fait sur la TV
utilisée, .
------------------------------------------------------
Don't change the picture size...if you change and take some
distance fom monitor, will permit to fuse front and background...
you see the relation between screen size and first spectators
distance !
Transparency permit to change more quickly eyes adjustment.
Try on this one... same trouble !
Of course Duke original picture without front subject do not
disturb adults on monitor.
Best regards,
Jm
Le 28/04/2014 14:24,
pierre.meindre(-at-)free.fr
a écrit :