>behaviour is a natural one for the eye Certainly. In fact one cue to depth perception for the brain is the degree the eyes have turned to bear on a point in space.
The old stereo cameras had no such mechanism so far as I know. The two lenses simply pointed straight ahead. It is for this reason that I would not have the cameras bother with turning onto a point independantly ... the feedback which tells the brain anything at all about depth from such an action is both minimal and confusing. It is confusing because it assumes that the eyes of the movie viewer are interested in the single point that the cams are aimed at. This is another reason why I hate DOF in digital rendering - it constrains the viewer unnaturally. The eye after all will constantly move about a scene. N. ________________________________ From: Jean-Sebastien Perron <[email protected]> To: [email protected] Sent: Mon, 8 February, 2010 3:14:30 PM Subject: Re: stereoscopic imagery If I remember on my website the stereoscopic tutorial use center of object as aimpoint instead infinity... This behaviour is a natural one for the eye The future is either connected to the brain or holographic voxels. Jean-Sebastien Perron www.NeuroWorld.ws Neil Cooke wrote: > >Saw an interesting one, an animated gif on loop showing old >stereo prints (for those old hand held stereoscopes) one after the >other with a fairly decent delay. The eye was totally tricked and saw a >3D view. Even the extremes of shift were just about there. My guess is >that these extremes happened due to the stereography having the centre >subject as the aim point rather than centre infinity. Google something >like "animated gif stereo prints old japan" ...... or similar. > > >N. > > > ________________________________ From: "[email protected]" <[email protected]> >To: [email protected] >Sent: Mon, 8 February, >2010 10:58:02 AM >Subject: Re: >stereoscopic imagery > >>With the shutter glasses approach you can use "regular" frame rates. >>Dolby 3D and RealD 3D, use higher refresh rates showing the same image >>for each eye several (three) times before moving on to the next actual >>frame, but if you aren't too fussy about it, a 60 Hz refresh rate, >>just showing each frame once, works quite well. At least for still >>images. Of course, even for shutters, higher refresh rates would make >>it look better, but if you are moving up that ladder, then the >>polarisation or phase filtering methods might be more suitable. > >>/ Fredrik > > >>On 7 February 2010 22:23, Mark Heuymans <[email protected]> >wrote: >>> Op 7-2-2010 20:19, Neil Cooke schreef: >>> >>> Two cents worth ... >>> Stereoscopic entertainment history shows it as a fad type of >thing... It comes >>> along gets all hot and fired then fades to zero. >>> >>> Yes, like quadraphony. Maybe this time it will break through, hard >to tell. >>> But it will double render times and make post processing much more >>> complicated. Cameras, both real and virtual are much more >complicated >>> because they need to replicate every movement of the human eyes. >(nice >>> project to set up in RS, also for mono renders!) >>> >>> >>> Only one of the seven or so ways the brain deciphers depth in a >view is >>> dependent on the stereoscopic event ... the fact that the left >view is 64mm >>> offset from the right view .... and as Jen-Sebastien points out, >actual near >>> and distant eye lens shift for focus is absent. The other factors >are >>> intellectual ....... place in the field, overlap, known and >relative sizes, >>> colour tone strengths, etc. >>> >>> Yes, but I once tried to play tennis with one eye shut - totally >impossible! >>> Sure, it doesn't add much to maybe 99% of all productions, it's >worthwhile >>> only for visual spectacles like Avatar and Coraline (great stop >motion movie >>> btw!). >>> >>> Nice and quick replies guys, I'm checking out the tips! Shutter >glasses are >>> maybe an option but I doubt if my LCD monitors are fast enough. >>> I was just wondering if I could view RS stereo animations in >stereo. I have >>> two identical LCD monitors, maybe I can set up something with >mirrors. >>> >>> Thanks, >>> Mark H >>> >
