The Problems Of Shooting S3-D
Posted 26 March 2012 - 10:56 AM
She doesn't quite expalin passive (pollarized light) 3D properly though becuase the D-Cinema passive REAL D 3D systems run at 144Hz or 72Hz per eye and the home system runs at 120Hz or 60Hz per eye. Whilst the passive glasses of D-Cinema don't have shutters built in, the device infront of the projector is a big shutter that also features the mentioned polarizing light filters'. So there is still left/right shuttering going on, it is just that it is done at the projector, not on the eye glasses. The problem with systems like REAL D is that the cinema screen must be 'silver' to keep the light pollarized where we can use a standard white screen at home with active shutters.
Edited by MarkTecher, 26 March 2012 - 11:06 AM.
Posted 26 March 2012 - 07:09 PM
Home systems in America may work at 60 or 120 Hz but in Australia it will be either 50 or 100 Hz or for Blu-ray 48 or 96 Hz
Posted 26 March 2012 - 07:34 PM
Posted 26 March 2012 - 07:54 PM
Posted 27 March 2012 - 12:58 AM
Home systems in America may work at 60 or 120 Hz but in Australia it will be either 50 or 100 Hz or for Blu-ray 48 or 96 Hz
This compares with 72 flashes per second Left eye, and 72 flashes per second Right eye, at a RealD public cinema; which is usually described as 144Hz. [No judder factor (jitter still present: 24fps is a slow frame rate, particularly for 3D). A reduced phase difference compared with a home 120Hz setup that uses alternation.]
Thanks for this, alanh. The first few minutes of the video present some of the basics of viewing, and capturing, stereoscopic 3D, very nicely. I think those first few minutes could be successfully used as educational material. Some oversimplifications (e.g. the diagram suggests cinema projectors use linear polarisation, rather than the much more common circular polarisation).
I'll be on the lookout for reviews of the finished surfing movie. I expect the 3D photography will be spectacular.
Edited by MLXXX, 27 March 2012 - 01:56 AM.
Posted 27 March 2012 - 08:55 AM
Posted 27 March 2012 - 09:36 AM
I'd really prefer 192Hz [96Hz per eye] for The Hobbit (48fps) when it comes out and assuming it is actually released at that frame-rate in Full HD 3D, using an upgraded form of Blu-ray, or other form of media.
Edited by MLXXX, 29 March 2012 - 12:27 AM.
Posted 27 March 2012 - 11:18 AM
It wil be interesting to see where D-Cinema goes and if indeed ultra fast refresh like 192Hz is even possible with the existing technology.
Posted 29 March 2012 - 12:16 AM
Shutter glasses models
1. Number of flashes per eye per second with different frame rate sources
1a. "Phase lag", related to 1, for example, "Right lags left by 8.33mS in 24p mode".
2. Left to Right and Right to Left crosstalk at a standard video contrast setting, at a standard distance, using shutter glasses supplied. (This may require development of an international measurement standard, if not already developed.)
3. Drop in brightness of screen in 3D mode, and total drop in brightness when using supplied glasses.
LCD panel with fixed alternate line alternate polarisation
1. Native screen crosstalk
2. Measured crosstalk viewing set with supplied passive glasses
3. Drop in brightness when using supplied passive glasses
Projectors that rely on polarisation have the screen as a potential variable. Perhaps cross-talk measurements could be taken at the screen, prior to reflection. And then the screen manufacturers could publish cross-talk figures for circular polarization (I think some screen manufacturers may do this already).
I currently am very satisfied with the 3D performance of small cheap LCD passive set* with its zero flicker, and zero phase lag, and good brightness level. However the cross-talk is disturbing at times. As a consumer, I would have appreciated a published figure on cross-talk.
I also think the review magazines could lift their game, and equip themselves to carry out measurements of crosstalk; and verify "flashes per second", and Left Right phase mismatch, and performance of the supplied glasses.
There is the question of the effect of discolouration by 3D glasses and whether that is compensated for by the display device when operating in 3D mode.
The other day I watched some of the 2010 FIFA World Cup Soccer at 720p50 3D.** It was very smooth. Watching a 1080p24 3D movie immediately afterwards underlined for me how inadequate 24p is as a frame rate for 3D. What our eyes are reasonably content with for 2D does not really suffice for stereoscopic 3D, as James Cameron and others have pointed out. I don't think a higher frame rate than 24fps will prove to be a mere fad for 3D cinema. I think using greater than 24fps for stereoscopic 3D will become entrenched.
* A Soniq 42" L42D11A.
** A Blu-ray acquired cheaply (40% discount) from a WOW store (trading under administration).
Posted 29 March 2012 - 02:00 AM
What does this have to do with shooting S3-D? A bit more complex than you have proposed in the past.
Your post is all about the display end.
Put it in a new post or another post about 3-D displays.
Posted 29 March 2012 - 02:47 AM
In any case:
1. The first few minutes of the video the OP linked to starts with a general explanation of 3D technology including use of shutter glasses, and it was that that prompted comments on display technology.
2. This thread is in a part of the forum that is a subcategory of the category Flat Panel TV Screens.
3. DTV forum threads are notorious for going off topic.
4. Only a handful of forum members contribute to 3D related threads, so if a thread does pique the interest of some of these people it is reasonable to allow the thread to meander somewhat, following the proclivities of the participants.
5. Your own contribution so far has been to post an external link (post #1), and provide somewhat misleading information (post #3). If you wish to provide something positive about challenges in shooting 3D and how those challenges are being met, you are free to do so. Please proceed.
Posted 29 March 2012 - 06:04 AM
Posted 31 March 2012 - 08:38 PM
It lasts 10m 45s. The blog describes a number of 3D issues and techniques in making the two Hobbit movies. The first movie (The Hobbit: An Unexpected Journey), due for release in December 2012, when completed promises to be the most technically satisfying 3D feature length production to that future date, as not only will great attention have been paid to the stereoscopic camerawork, such as toe-in adjustment, but the use of 48fps for stereoscopic 3D will surely prove to be a "sea change" advance over 24fps. No special need for a surfboard camera "air knife", though.
The use of 48fps should provide a palpable boost in realism. Or, in the words of James Cameron (quoted at the foot of http://www.christied...es/default.aspx ):-
“If watching a 3D movie is like looking through a window, then [with HFR] we’ve taken the glass out of the window and we’re staring at reality.”
“Avatar” director, James Cameron, on high frame rate movies
Posted 01 April 2012 - 10:56 AM
When a 3-D image is captured the left and right images are captured simultaneously so the capture frame rate has nothing to do with the quality of the 3-D image. In fact a single image can be captured as a still and is not affected by the frame rate.
The display can be at any multiple of the captured frame rate. The current methods of display are sequential left and right causing a delay. This would not be the case if laser holograms were used.
I recommend 48 frame/s for the illusion of motion to become smoother but this has nothing to do with S3-D After all the selection of the frame rate is upto the producer and their financiers. The display is upto the viewer.
Posted 01 April 2012 - 12:57 PM
Using the traditional 180 degree camera shutter exposure time, there is image blur and jitter that has been found tolerable at 24fps, for 2d. However, empirically it has been found that 24p is a very significant compromise for stereoscopic 3D. I think that is because the brain is actively engaged in comparing the Left and Right views, and that that level of activity requires a frame rate closer to 60fps than 24fps for the sensation of stereoscopic realism to be reasonably maintained.
A higher frame rate than 24fps may require props that involve spatial movement to be of a higher standard. I recall that the production video I linked to referred to the need for a large wig to be made of real hair. This was because the movement of the hair in the wind would be subjected to inspection by a stereoscopic pair of cameras (exposing the spatial movement to scrutiny by the audience) and at a faster frame rate (permitting even greater scrutiny of the spatial movement!). Long strands of human hair in a close-up acquire a spatial reality that demands a material that actually moves with the springiness of human hair.
On the related question of challenges of simulated shooting in 3D, falling rain in nature involves individual droplets being visible at different distances. Water splashes similarly. Such scenes provide a litmus test of whether the material is or is not genuine stereoscopic 3D. Productions involving continual close-ups of water splashes do not lend themselves to simulation: e.g. Grand Canyon Adventure - River at Risk 3D, or Sanctum 3D. The 3D simulation of Titanic from the original 2D footage will be challenging wherever there are sprays of water in a close-up (one solution being a blurring to hide the lack of 3D differentiation of the hundreds or thousands of moving droplets).
In 2010, sequential was dominant in the home. In 2011, there were sales of hundreds of thousands of LG "cinema" 3D LED TV displays that display Left and Right simultaneously. In 2012 Panasonic are bringing out such models. There are other brands too, that display Left and Right simultaneously, such as my own humble Soniq 3D TV, that I use as a 3D computer monitor.
There is an expensive consumer level projector of which I am aware that uses separate lamps and separate polarising filters for simultaneous projection of Left and Right.
Although most digital commercial cinemas use the RealD projection method of frame alternation at 144Hz, some commercial digital cinema projectors use simultaneous projection. For many years, IMAX in its non-digital form has used separate reels of film, one for Left and one for right. These are synchronised for simultaneous display of Left and Right for viewing with linearly polarised glasses; though shutter glasses have also been used with IMAX.
I'd reiterate that use of 48fps or 60fps is not merely for better fluidity of motion, it is a more fundamental question than that. The brain's processing of stereopsis is significantly impaired by a frame rate as low as 24fps. One early scene in Avatar exemplifies this. The cameras show the boots of the commanding officer as he strides towards the dais to deliver an induction address to the new recruits. I was astonished at the cinema to see a staccato stroboscopic succession of stills. I had never noticed such severe jitter with 2D. Even when viewed with my Soniq 3D TV with its simultaneous display of Left and Right, this scene, and many others in Avatar and other 3D 24fps movies are distractingly jittery, far more so than watching a 2D 24fps Blu-ray.
When watching stereoscopic 3D I find my eyes craving a higher frame rate to maintain a mental sense of where everything is spatially located, i.e. something more closely comparable with real life vision.
Also, and I mention this for the first time in this thread, when watching in stereoscopic 3D I become more conscious than I do when watching 2D, of the inadequacy of 1920x1080 as a sufficiently dense sampling grid. What I see looks as though I am wearing glasses of a slightly wrong prescription: it is slightly blurred. 1920x1080 was a huge advance over the 720x576 sampling of PAL DVDs, but we really could do with an even denser sampling than 1920x1080, for an even more realistic visual experience in the home.
The GoPro cameras used for certain action scenes in the surfing movie are only 1080p and in my opinion barely prosumer grade (see the HERO2 range, which is generously described as "professional", http://gopro.com/hd-...CFWFNpgodIgKSxw ). They are probably adequate for very fast moving surfing scenes in terms of raw video resolution but their overall video performance even for that specialised application will be compromised by the limited bitrate of the compressed video.
The Red cameras being used for the two Hobbit movies have state of the art video resolution. I can't specifically recall reading what sample grid is being used for the digital intermediate but presumably it will be around 4k or 5k, rather than the often used 2k, so these Hobbit movies could be shown in advanced public cinemas and in a future improved home consumer format at not only a higher frame rate than we are used to, but with superior resolution. It's quite exciting what we may be seeing in the not too distant future.
Edited by MLXXX, 06 April 2012 - 08:00 PM.
Posted 02 April 2012 - 02:32 AM
With still photography if a camera is set for 1/24 to shoot motion, the only way to use this is to pan the camera in the direction of motion. This produces a clear image of the moving object, and the background has motion blur making the speed of the object to appear higher. So if a 180 degree shutter is used that is a image every is exposed for 20.83 milliseconds every 41.7 ms. So if there is plenty of light exposure could be reduced to 1 ms (1/1000) where the movement blur will be neglible but the jerkiness will be most obvious. If 48 frame/s is used the shortest exposure is a little less than 20.8 ms, and again if the camera operator and the lighting and permit it the 1 ms exposure can be used again. If these short exposures are used at night, it is an expensive requirement for lighting and the power consumed.
If you wish to compare the jerkyness of 2D to 3-D you have to do it very close chin time otherwise your view becomes coloured by expectations. This is a well known psychological effect. Its just like trying to memorise the colour of a shade of white.
Your interpretation of wigs is to increase the depth of the hair to emphasise the 3-D. When colour started the directors required strong colours to emphasise the fact it was in colour. Later after the novelty wore off, natural colours were used. So extra large wigs will not appear natural and this practice will die.
I don't care about conversions from 2-d to 3D unless the vision is re-created which is what has happened to Titanic 3-D where many mgraphics designers were employed by James Cameron to put in the desired depth values.
I don't care about displays in this section other than to say that Frame compatible 3-D would have had the image captured simultaneously and the left image is followed by the right image. If the non frame compatible distribution of the signal means that signals are near simultanous. It is then upto the the designer to how it's displayed.
Currently any single display device must use a sequential left and then right displays.
Posted 02 April 2012 - 03:28 AM
Regarding single displays, the LED passive 3D TV models have a fixed polarisation for odd lines, and a fixed opposite sense polarisation for even lines of the display panel. So both Left and Right content is updated at the same time. A drawback of this technique is a fine venetian blind look if a viewer moves close to the display.
Posted 02 April 2012 - 05:02 PM
Reproduction of fine detail has always been a test for 2-D TV particularly when the signal is being used for chroma keying. So the registration of the pair of images in 3-D becomes particularly critical due to the fine detail. It is also extremely important to have the camera exactly as level.
The only broadcast 3-D has reduced the resolution because;
the frame compatible resolution of Full HD is reduced to 960 x (540+)
The adoption of broadcast transmission standards for 3-D has not been standardised because you need DVB-T2 & MPEG-4 to transmit S3-D in Full HD. Also this will enable the transmission of 48 and 50 frame/s images. When S3-D full HD @ 50 frame/s is not being transmitted that data capacity can be used for other programs requiring less data rate.
Posted 06 April 2012 - 07:57 PM
A very competent, monumental, effort! The first shot of the ship about to depart England (the port of Southhampton) almost had me gasping. The ship had a slight animated look at times but I did not find it too difficult to suspend disbelief.
The footage early in the movie of the exploring of the wreck, rather drab as I recall seeing it in 2D, 14 years ago, was more interesting in 3D.
In close-ups, Kate Winslet's prominent eyes appeared to me and my partner to be in 3D. And certainly the faces and heads of a number of the actors were clearly in stereoscopic 3D, indicating very detailed use of depth maps, and the likely painting in of missing view detail (as was done in the conversion of Harry Potter and the Deathly Hallows: Part 2). A tremendous effort must have gone into this simulation. It reportedly cost $18m.
The engine room with its huge, fast moving, piston rods, was spectacular.
Fourteen years ago I found the 2D version of this three and a quarter hour movie a little tedious, and some of the characters a bit two dimensional (no pun intended). The conversion to 3D can do nothing to overcome any deficiencies of the screenplay script (deficiencies I felt again today). However the use of 3D, albeit only simulated, in my opinion has been a very worthwhile endeavour from an artistic viewpoint. It allows the spectacle of the luxuriously appointed interior of the ocean liner, her ultimate foundering, and of so many individual scenes, to take on added dramatic effect.
I'm sure much celebrated movie director Cameron is onto a winner here. The $18m was money very well spent. It will raise the stature of this 'poor boy meets rich girl in doomed luxury liner' 1997 blockbuster, in the pantheon of movie classics. I note that early box office indications are very strong.
And what of droplets of water? I recall one scene where a few drops appeared very much in the foreground. On the whole, though, the turbulent water was accorded a flat, slightly out of focus look, albeit at an appropriately staggered apparent distance away. That is about all that can be expected with post-production 3D without attempting to overlay a fresh animation on top of the original 2D footage. [I don't find simulated 3D as fascinating (or intoxicating) as I find real life movie scenes captured using separate Left and Right cameras. I wouldn't be alone in this.]
I thought there may possibly have been some additions of electrical sparks (appearing to descend from a glass dome) not present in the original movie, while the room below the dome was filling with water. These sparks were obviously coded for stereoscopic depth. On the other hand those very sparks may have been special effects added into the original 2D footage, that I simply don't remember, after 14 years. No doubt some people will be doing minute comparisons, particularly when a 3D Blu-ray version is released.
Still not my favourite movie, but the 3D treatment has given it a substantial boost in dramatic effect, IMHO. Recommended.
Edited by MLXXX, 07 April 2012 - 07:20 PM.
Posted 07 April 2012 - 03:30 AM
Not one mention of motion blur for a movie shot at 24 frame/s!
I agree that 48 frame/s allows directors more latitude for fast moving objects.
Posted 07 April 2012 - 01:53 PM
I was able to sit near the front of the cinema, and in the middle of the row. I was conscious of jitter right from the start. I noticed it particularly during the first 1912 view of the enormous ship leaving the dock. My initial near gasp of excitement on seeing a full view of the ship in stereoscopic 3D was followed by a strong awareness of continual jitter as the camera view of the ship dollied and panned. The low frame rate detracted from the realism. The detrimental effect of the jitter seemed significantly worse for my eyes than with 2D.
After a while my brain was able to ignore the jitter, most of the time.
However jitter reasserted itself perniciously in part of the fantasy epilogue stage of the movie. In the imagination of the elderly survivor she and her lover were being toasted by the assembled passengers back in 1912. The camerawork starts with a brisk progress of the camera view through the corridors that lead to a large room with a formal staircase, a mezzanine level, and a translucent dome in the ceiling. This brisk progress of the camera view is accompanied by severe jitter.
The scene itself provides a positve and romantic touch, an excellent conclusion to the story telling, but the 24fps frame rate technically compromises the beginning of the scene.
I'd also mention that with the simulated 3D movies I have seen to date, my brain tends not to be quite as actively engaged as with the true stereoscopic 3D movies I've seen. I think it is because I concentrate less with simulated 3D, that I do not notice jitter, or lack of visual resolution, as much as I do with true stereoscopic 3D.
Something else Titanic 3D tests is the ability of an audience to watch stereoscopic 3D for an extended period: three and a quater hours. Neither my partner nor I experienced any sensation of viewing fatigue. I think it may have helped that the strength of the 3D effect was 'mild' and 'lifelike' most of the time. There will be a percentage of viewers who will complain that the 3D effect should have been made stronger, i.e. they would prefer to view an exaggeration of the stereopsis we perceive in real life.
Edited by MLXXX, 08 April 2012 - 10:47 AM.
Posted 07 April 2012 - 09:02 PM
Posted 07 April 2012 - 09:48 PM
My vision is strongly binocular. Walking down the street with one eye closed is a totally different experience for me than walking down the street with both eyes open. With both eyes my brain creates a powerful spatial analysis, much more powerful a feeling for me than the difference between stereo audio and mono audio.
With only one eye open, the outlook is flat, uninteresting, and it is not as obvious how far away different objects are, looking in the range 50m to 300m ahead. If I became blind in one eye, my driving of a car would need to be much more conservative. My ability to judge distance would be significantly impaired. Whether given time and experience my monocular vision would improve for the purpose of driving a car I don't know.
I rely little on focus. Objects more than about 10 metres away are virtually at infinity. Only quite close objects differ in their focussing requirements for my eyes. On the other hand, I can see parallax differences quite distinctly at 200 metres.
I have discovered spatially complex scenes on 3D Blu-ray discs that look coherent when watched in 3D: flocks of birds, schools of fish; but look like a jumbled mess of shapes when viewed in 2D with both [my] eyes. Rain in 3D looks completely different to me to rain in 2D. People's faces look very different in 3D.
For my vision, stereoscopic video is a huge advance over 2D video. It is not an effect that boosting the brightness or contrast of the photography, or the display of it, could substitute for. Not for my vision.
Edited by MLXXX, 07 April 2012 - 10:23 PM.
Posted 07 April 2012 - 11:13 PM
As far as I know the displays you have owned dont qualify as high contrast, dont dismiss the advantages till you have experienced them.