3D and long lenses

[David Mullen ASC]

Well, if you aren't interested in that lifelike immersive viewing experience and sit at the back of the movie theater, then why are you interested in 3D? Why ask about matching human vision when you want to sit so far back, look at a small image, that the experience is nothing like human vision?

 

[Tom Lowe]

Tom,

No, those won't work unless you have one of the so called interlaced monitors, but alas, that image Michael posted has very, very convincing 3D effect, even though is shown in the anaglyphic (red & blue) format, which is the oldest and worst of all formats. Michael shot the shark footage with one of his housings, but I am the one that did all the 3D work on it. And let me tell you, since you were being discouraged the other day on another thread that posting 3D is really a very complex endeavor, that was my first ever attempt at posting 3D, and although not the best, the resulting images are 3D, plain and simple. Get yourself a pair of red/blue glasses from 3dglassesonline.com and take a look at that frame. Then send me a PM if you're interested in finding out how feasible it is to start working on 3D, I'll hook you up brotha. As a matter of fact, once you start shooting 3D, I will post your first few projects for you free of charge until you get yourself set up. Don't listen to the naysayers, start playing with 3D, because uncle Cameron is right, it is the way of things to come...

Rudi, thank you! I will take you up on that.

It's almost certain in my mind that my 2nd film will be 3D. That would be starting around June 2011, I think, so I will have some time to figure all of this 3D stuff out and wait for the 3D rigs to come along to support the new DSMC cameras and DSLRs I might use for 3D timelapse.

BTW, of course it was your epic, prescient post about AVATAR back in June that got me so fired up about AVATAR... and you turned out to be right! http://reduser.net/forum/showpost.php?p=441160&postcount=68 :cheers2:

 

[Michael Hastings]

Tom:

Basic 3D has never been that hard to achieve - the trick is to get two cameras of high quality, small size, and reasonable price. That's the big convergence that is happening now.

I (and now Rudi) have been playing with two Canon HFS100 camcorders. These shoot true 1080P at 24Mbps AVCHD. The cameras are $899 each and since they record to SD cards they are very compact allowing an intraocular of about 2.75". These are not quite as good as SI2K - but amazingly close - and they have the capability of recording to something like the nanoflash for much better recording/compression quality.

We are using a simple side by side bracket with a piece I machine on my CNC to hold the cameras perfectly parallel and solid since any movement or misalignment between the cameras causes problems.

I would suggest you look into working with something like this. While it isn't as sophisticated as the beamsplitter systems it is very useful for the "normal" ranges and will give you a good feel for what works in 3D.

 

[Tom Lowe]

Well what I'm thinking is two 2/3" Scarlets for a 2K 3D finish. Are the 2/3 Scarlet bodies small enough?

 

[Roberto Lequeux]

Well, if you aren't interested in that lifelike immersive viewing experience and sit at the back of the movie theater, then why are you interested in 3D? Why ask about matching human vision when you want to sit so far back, look at a small image, that the experience is nothing like human vision?

All the way back only when the room is small and the screen big David. I sit as close as I possibly can without having to move my head, try to overlap the edge of the frame with the line between peripheral and not. I want to get the most without missing anything, or adding anything of my own. I try to plug the art directly into my brain.

What I (personally) don't like is when some of the art is used for peripheral, or even intended for it. I don't like that sort of immersive experience, but see heaps of purpose in making the art more realistic in that area I am interested in. With friends and family I am often confused with shallow, overly rational, cold, and that is because I like to rationalize and dissect everything. It has no baring on my actual degree of emotion and those that know well tell me I am the opposite. But I do like to watch and experience the world actively, in a rational sense. I bathe in pressure waves and photons when I watch something and I believe my placement in the theater comes from that personal trait I tried to describe... I guess what I am fumbling with in trying express here is that I don't want to miss anything that the artist put in there, since understanding something is only achieved by looking at the whole.

Understanding human vision can only help with a visual art aimed at humans, and I see 3D giving me the ability to make something that much more realistic and present when it calls for it, whether you chose to immerse yourself in the screen or not.

 

[Roberto Lequeux]

Well what I'm thinking is two 2/3" Scarlets for a 2K 3D finish. Are the 2/3 Scarlet bodies small enough?

Am I the only one that still has this image (even if just a bold statement of intention, which I trust) very present in their mind?

And whether you are able to shoot 3k 2/3" or 6k S35 seems to be all about the lenses used and inter-ocular limitations due to their front element diameters. Without going beam-splitter of course.

 

[David Mullen ASC]

It just seems contradictory to be interested in matching human field of vision but say that you aren't interested in peripheral vision, which is a major aspect of human vision.

If you are OK with not taking advantage of peripheral vision by keeping the viewed image smaller, fine, but forget then trying to replicate human field of view in a realistic-looking way - your image is going to be way too wide-angle to look natural if you squeeze all of human field of view into a projected frame that does not come close to filling the viewer's field of view -- get it? You've just crammed a wider view into a narrower view. It's like I said about an IMAX image on a 50' tall screen. If you sit close enough that a lot of the image encroaches your peripheral vision, then wide-angle photography feels less wide-angle and distorted because your eyes are concentrating on the narrower area inside the wider area. But sit farther back so you see the whole frame clearly in the center of your vision and the image is very obviously wide-angle.

Look at IMAX movies on DVD and notice the odd framing, with a lot of excess headroom, since your eyes are supposed to go to the lower half of the frame, and the use of very wide-angle lenses at times.

Large screen presentation that fills the viewers vision, engaging their peripheral vision, is the centerpiece of all immersive formats starting with Cinerama.

Now 3D has been a bit different in that earlier renditions did not involve hyper-clear photography that could be enlarged greatly on a huge screen, other than 3D IMAX (which is quite spectacular) and 3D 65mm ala "Captain Eo".

But I think that 3D and higher-resolution photography destined for large screen presentation are converging trends designed to make the theatrical experience more impressive.

Personally, I've been of the belief that you make the most impressive theatrical movie you can and the TV versions often take care of themselves, especially since most movies are released in their theatrical aspect ratio on DVD and Blu-Ray. In fact, one could make the argument that some of the greatest theatrical experiences naturally do not translate as well to home video, such as "2001", if you do it right, make something that literally screams out to be seen on a large screen. You look at some of the most demanded Blu-Ray titles and they are movies like "Star Wars" and the "Raiders" trilogy, movies shot in 2.40 35mm anamorphic for theatrical presentation, a format not optimal for TV presentation.

What I (personally) don't like is when some of the art is used for peripheral, or even intended for it. I don't like that sort of immersive experience, but see heaps of purpose in making the art more realistic in that area I am interested in. With friends and family I am often confused with shallow, overly rational, cold, and that is because I like to rationalize and dissect everything. It has no baring on my actual degree of emotion and those that know me do tell me I am as deep as they know, maybe annoyingly so at times. But I do like to watch and experience the world actively, in a rational sense. I bathe in pressure waves and photons when I watch something and I believe my placement in the theater comes from that personal trait I tried to describe... I guess what I am fumbling with in trying express here is that I don't want to miss anything that the artist put in there, since understanding something is only achieved by looking at the whole.

Understanding human vision can only help with a visual art aimed at humans, and I see 3D giving me the ability to make something that much more realistic and present when it calls for it, whether you chose to immerse yourself in the screen or not.

 

[Dan Hudgins]

Various focal lengths

Various focal lengths

For the most part you should NOT use more than one focal length when shooting 3D and you should only sit in one seat in the theatre.

But thats not what happens, and you can use more than one focal length and people can sit in more than one seat in the theatre if you keep the forground close to the subject and you keep the most distant object about in the same position in both cameras.

You can see some depth in distant subjects its just that the accuracy of that perception is in ratio to the distance.

When you look through binoculars you see exaggerated stereo, but not in ratio to the magnification since the prisms are not making the front lenses as wide as they would be to be 10x or 20x apart.

I think it would be better to shoot stereo for distant nature subjects than to print the same 2D image for both eyes, but you should not compensate the spacing of the cameras in ratio to the increase in focal length, rather make the spacing larger than 65mm but not larger than maybe 650mm.

Also you need to keep the forground out of the image since any foreground will cause too much disparity to be fused by stereo vision.

As for the background you need to keep the image points from being reversed more than 65mm on the movie screen (which would make the viewer's eyes diverge). Because the screen is not close to the subject it is best to keep objects at infinity to superimpose on the screen, although you might want things to seem more distant than the screen the small 65mm movement of the image points is hard to control or predict and any divergence of the viewers eyes will cause eye strain, so having the most distant subject points align on the screen plane is not unreasonable for use in a theatre setting.

That means that when you use longer lenses ALL points in the subject will be adjusted for the eyes to cross (converge) at a point in front of the screen, but to be limited to no closer than two meters or so.

 

[Roberto Lequeux]

I apologize if I am not making myself clear David. BTW, I always sit close enough to have to move my eyes a lot, and thinking about it I do move my head, a very little bit, but mostly to avoid straining my eyes. You will never find me at the very back of, a large, room either. I'd rather wait for an empty showing. Maybe I can bring a tape roll next time so I can give you the exact numbers for the screen and distance. :thumbsup:

What I am not currently interested in is approximating a "virtual reality" of sorts. Instead I just want to understand how human vision works because it can only help me become more able to generate the responses I am after. I just wish I could make sure I could present a very large painting which you can easily lose yourself in, if you choose to, and not because you have no option. People chose their theaters of course, but I digress. My goal then is to successfully make the material seductive enough that you would willingly accept the invitation. I would feel I succeeded when I manage to strike a balance that leaves the audience feeling both respected and stimulated. 3D gives you other buttons to press, but the art is still in the performance behind the keys.

I wish I could give you a technical quantification of the range of visual stimulation that I would personally wish to have available... but even if I could all humans are different, would be in different moods, have different personalities... Either way, I don't think I'd be able to convince all theaters and consumers to build their rooms to my specifications, hehe. Also, I am a bit daunted by the myriad of screen sizes, and technologies that are out there at the same time. I'll limit myself to continue to vigorously study cinematography and the rest of my craft, and hire the best DP I can.

 

[Dan Hudgins]

Distance to screen

Distance to screen

There is a very big issue with shooting 3D when you try to view the images on a monitor.

Because the distance between your eyes is fixed and the distance between you and the monitor is much closer than the screen in a theatre the spacing of points on the screen for INF on a monitor is LARGER in ratio to the image width than it would be on the screen for your eyes being parallel.

That means in shooting or Post the way the images are lined up cannot be the same for viewing on a monitor with 3D glasses and on a movie screen that is use in a theatre, nor can the same image line up used for 3D Blu-Ray or broadcast be used in a theatre, and if people have a 12" set or a 40" set they cannot see INF if the images are formated the same way, the distance between points on the screen needs to be adjusted depending on the size of the screen for your eyes to be parallel, the spacing of the points needs to be about 65mm no mater what size screen is used, but 65mm is more of the image width on a 12" monitor than on a 12meter movie screen.

If you show a stereo pair setup for a 12" mointor on a 12 yard movie screen then points that would be 65mm apart on the 12" monitor would be about 2340mm apart, causing the viewer's eyes to diverge and eyes never diverge in natural vision so that causes eye strain and in some there not being able to see the images as stereo.

So how do you go about making a different stereo version for each screen size?

In the old two projector 3D days, the setup films let the projectionist adjust the way the align the INF point on the screen, almost no spread if done right, but with digital release, home versions would need some sizeable spread to get INF, but that spread depends on the screen size since you never want the eyes to diverge. If you keep INF at zero disparity then you cannot see INF on a small monitor, but that is a safer way to go with a single alignment since eyes can converge at the screen no mater what size or distance it is.

The only way to monitor on the set or in Post is to use a viewing hood or VR glasses etc, where the eye lenses set the screen at a far point like the movie screen, so the left right shift for INF points can be near zero displacement (unlike on a monitor viewed with glasses where some displacement is needed for objects to seem at INF). But that does not help with viewing on a lenticular TV set, which is why it seems the Sony 3D camera uses a single large lens, but that has its own issues.

 

[Rudi Herbert]

Well,
I have to agree with Mike, all you need is a pair of reasonably cheap and high quality cameras, there's no need to wait for Scarlet or Epic to come out. What I have realized is that, like Mike said, achieving basic 3D is not all difficult, and a side by side rig with two cameras properly aligned and running identical settings is all that's needed for this. However, after spending 3-4 days posting the 3 minute short with the sharks that Mike posted a photo of, I also realized that, as every other technology with its own set of specificities and whims, 3D must be practiced a LOT to improve. Lots of different things make a big difference, not just the convergence of the two images shot by the cameras. I was so surprised how the color correction made a difference, as did contrast and saturation. Altering each of those made a profound difference in the stereo effect. Too much saturation caused edge softening and loss of 3D, contrast improved it, but too much of it led to ghosting, selectively tweaking some colors made the 3D more subtle or very in your face, etc, etc, etc. My point is that for those who are seriously interested in 3D and can afford a small investment, like the Canon cameras Mike and I are using, it is well worth it. Learn the ropes now, learn what works and what doesn't, how to stage scenes for a particular 3D look, etc, etc, but this teething period will take time, so getting it out of the way now before you dump a substantial amount on RED gear to start from scratch is not something I'd recommend...

 

[Elsie N]

Just was thinking that if your 3-D project is an important one, maybe it would be a good idea to use the greater cameras, such as the Scarlet. That way, if the 3-D didn't workout, you could always use one camera's output and finish in a higher quality 2-D output.

 

[Rudi Herbert]

That way, if the 3-D didn't workout, you could always use one camera's output and finish in a higher quality 2-D output.

Well, if I promise a client 3D, I better deliver 3D, so it "not working out" is not, and never has been an option for me. I just come across so many people nowadays that have not paid their dues, never shot anything on Beta, DV, HDV, AVCHD, whatever, just because is not "good enough", even though thousands of us have made a living out of using those formats. They're waiting and waiting for the dream of RED to become affordable enough, and when that happens, they step out there with a professional camera and fail miserably, since all they have is empirical knowledge and never, ever shot anything in their lives....and of course, I'm not referring to anybody in this thread or forum, although such characters do abound here and elsewhere...

 

[Michael Hastings]

Well what I'm thinking is two 2/3" Scarlets for a 2K 3D finish. Are the 2/3 Scarlet bodies small enough?

Certainly, in fact the cheapest scarlet with a couple of mini-primes will be an excellent candidate for side by side 3D.

Well,
I have to agree with Mike, all you need is a pair of reasonably cheap and high quality cameras, there's no need to wait for Scarlet or Epic to come out. What I have realized is that, like Mike said, achieving basic 3D is not all difficult, and a side by side rig with two cameras properly aligned and running identical settings is all that's needed for this. However, after spending 3-4 days posting the 3 minute short with the sharks that Mike posted a photo of, I also realized that, as every other technology with its own set of specificities and whims, 3D must be practiced a LOT to improve. Lots of different things make a big difference, not just the convergence of the two images shot by the cameras. I was so surprised how the color correction made a difference, as did contrast and saturation. Altering each of those made a profound difference in the stereo effect. Too much saturation caused edge softening and loss of 3D, contrast improved it, but too much of it led to ghosting, selectively tweaking some colors made the 3D more subtle or very in your face, etc, etc, etc. My point is that for those who are seriously interested in 3D and can afford a small investment, like the Canon cameras Mike and I are using, it is well worth it. Learn the ropes now, learn what works and what doesn't, how to stage scenes for a particular 3D look, etc, etc, but this teething period will take time, so getting it out of the way now before you dump a substantial amount on RED gear to start from scratch is not something I'd recommend...

This is my whole point. These days with what I call "ebay-rental" where you buy a couple cameras for $899 and have the ability to instantly sell them seven or eight months later for probably $600 or so means for an "ebay rental" cost of $600 you can spend the next 7 or 8 months shooting and editing 3D material, and if your heart is set on scarlet simply sell the canons when it is available. I would suspect you can easily make back that $600 (3D is a great gimmick for trade show product demos for instance) but if not just the learning experience should be worth it.

 

[Elsie N]

Well, if I promise a client 3D, I better deliver 3D, so it "not working out" is not, and never has been an option for me. I just come across so many people nowadays that have not paid their dues, never shot anything on Beta, DV, HDV, AVCHD, whatever, just because is not "good enough", even though thousands of us have made a living out of using those formats. They're waiting and waiting for the dream of RED to become affordable enough, and when that happens, they step out there with a professional camera and fail miserably, since all they have is empirical knowledge and never, ever shot anything in their lives....and of course, I'm not referring to anybody in this thread or forum, although such characters do abound here and elsewhere...

My point is that RED is about "not paying your dues" because that puts you at the mercy of someone else's ideas of whether or not you've paid enough. Besides, if you think an independent movie-maker hasn't paid their dues, I submit you are unfamiliar with the new normal. By using the better quality camera on your independent project, you have the option of selling your work as a 2-D deliverable if no one is interested in it as a 3-D deliverable. Of course the youtubers and such need not use the higher quality camera. But I think Tom's aim is to film unscripted actors (wild life) and if he has two high quality cameras and one isn't quite in sync with the other or just fails for some reason, then at least he has the shot. I wasn't slamming you for using sub standard cameras, just saying that the redundancy of the higher quality ones is a safety net.

 

[Michael Hastings]

My point is that RED is about "not paying your dues" because that puts you at the mercy of someone else's ideas of whether or not you've paid enough. Besides, if you think an independent movie-maker hasn't paid their dues, I submit you are unfamiliar with the new normal. By using the better quality camera on your independent project, you have the option of selling your work as a 2-D deliverable if no one is interested in it as a 3-D deliverable. Of course the youtubers and such need not use the higher quality camera. But I think Tom's aim is to film unscripted actors (wild life) and if he has two high quality cameras and one isn't quite in sync with the other or just fails for some reason, then at least he has the shot. I wasn't slamming you for using sub standard cameras, just saying that the redundancy of the higher quality ones is a safety net.

I may not be understanding this so don't take it wrong. The first point is that these cameras exist now. Second is that you have redundancy/backup with either type of camera. Third is that you might be surprised at how good the quality is. These are modern full 1920 x 1080p chips, so 66% of the resolution of scarlet with the main difference probably being lower sensitivity due to smaller pixel size. But this often isn't an issue when shooting outdoor nature stuff where you shoot at f8 and up (in fact with most modern video cameras you often need to use ND so you don't have to stop down past f11 and lose resolution due to diffraction effects). And the increased depth of field with the smaller chip may actually help with with telephoto on many nature subjects. Further, they also shoot 8MP stills so time lapse may actually be better than scarlet.

But mainly this isn't about HFS100 vs scarlet, it is about the fact that not much has changed in actual 3d techniques in a long time - it just wasn't very practical to get into it. But the convergence of affordable and capable editing tools plus camcorders that are extremely compact (giving suitable intraocular distances); extremely high quality (true 1080P HD for 1/4 of the price we paid for a fixed lens one focal length no zoom 8mm camcorder with half of standard definition resolution 20 years ago); and extremely affordable; means you can start producing and learning those 3D techniques now with a minimal trade up cost when scarlet is available.

 

[Tom Lowe]

For me, the point is well taken that it's a good idea to do tests and experiments well in advance of starting a 3D project. Since I will be busy with principle photography on my first film all of next year, that will allow me time to wait for Scarlet 2/3" and the 3D rigs that will follow its release -- and to start doing tests in the meantime. The fact that the 2/3" Scarlet will create beautiful, quality 2K with overcranking up to 150fps is a big selling point for me.

Then again, if my plan turns out to be literally flying the 3D rig on a Dragonfly-style RC helo for 50% of the film, maybe it will turn out that CCD sensors are a better idea than CMOS. I think the 2/3" Scarlet should be good on RC choppers, but with 3D, I have to imagine that jello could be a deal breaker, since matching the two stereo images together with significant jello and vibration skew sounds very difficult. I've seen footage shot with vDSLRs on RC choppers, and it aint pretty.

 

[Elsie N]

I may not be understanding this so don't take it wrong. The first point is that these cameras exist now. Second is that you have redundancy/backup with either type of camera. Third is that you might be surprised at how good the quality is. These are modern full 1920 x 1080p chips, so 66% of the resolution of scarlet with the main difference probably being lower sensitivity due to smaller pixel size. But this often isn't an issue when shooting outdoor nature stuff where you shoot at f8 and up (in fact with most modern video cameras you often need to use ND so you don't have to stop down past f11 and lose resolution due to diffraction effects). And the increased depth of field with the smaller chip may actually help with with telephoto on many nature subjects. Further, they also shoot 8MP stills so time lapse may actually be better than scarlet.

But mainly this isn't about HFS100 vs scarlet, it is about the fact that not much has changed in actual 3d techniques in a long time - it just wasn't very practical to get into it. But the convergence of affordable and capable editing tools plus camcorders that are extremely compact (giving suitable intraocular distances); extremely high quality (true 1080P HD for 1/4 of the price we paid for a fixed lens one focal length no zoom 8mm camcorder with half of standard definition resolution 20 years ago); and extremely affordable; means you can start producing and learning those 3D techniques now with a minimal trade up cost when scarlet is available.

No quarrels with anything said here. Was just addressing Rudi's response to my post that a 2-D backup option from using higher quality footage for a movie deliverable (Tom's stated purpose) was not in the equation. My first post:
"Just was thinking that if your 3-D project is an important one, maybe it would be a good idea to use the greater cameras, such as the Scarlet. That way, if the 3-D didn't workout, you could always use one camera's output and finish in a higher quality 2-D output."

was simply being dismissed because everyone hasn't "paid their dues." People are paying their dues right now... the younger ones are only just starting paying them at a different spot on the timeline. Learning currently obsolete or soon-to-be obsolete ways of doing things are a waste of valuable time IMHO. I think Tom said 2011 would be when shooting would begin. By then, it may very well be possible to set up a Scarlet 3-D rig that will rival point and shoot technology of today.

 

[Ketch Rossi]

For me, the point is well taken that it's a good idea to do tests and experiments well in advance of starting a 3D project.

Well Tom, that pretty much says everything one can say, to me testing, rehearsal, and then some more testing, and experimenting is all you will ever need, off course it is always nice to ask questions and get advises and info related before doing the testing and experimenting, as to have more of an idea of what can be expected.

Personally I have been taking everything in my hands as half the info out there is completely and incredibly WRONG.

S3D has some very fundamentals that must be absolutely obeyed, as far as not giving your viewers some very bad experience, but you can use any lens at any focal, zoom or Prime, even so I prefer primes, and just test it out, and see what look pleases you the most, just as far as you avoid some classical mistakes.

As far as the Scarlet 2/3 with Mini Primes, that would surely also be a camera set up of choice, do to size and flexibility, since we have seen that the 2/3 imagers from the Sony cameras used in some great S3D blockbusters do more then in off cover the needs, however we have to see when the Scarlet and Epic get released and tested, which one at the ned could be the real winner for S3D use.

IMO, they both will serve a very distinct place even in S3D shooting needs, then again, there is wryly not that Huge of a difference form one Brain and the other, but there is a Huge difference in the lenses.