Eric Lange
Well-known member
Yup absolutely…
This is a bit long… but explains and lays out my Symetric/non-retrofocal technical philosophy.
We are doing high precision s-3d/photogrammteric to VR/CG type application on high res MF sensors. Basically the eyes for the software we have been developing for over six-seven years. But the technology has many cinemagraphic applications. So testing with Scarlet will be eminently useful.
So when you (RON) say “view-finder” camera we are building around 40MP sensors+ but using the Schneider –KREUZNACH Apo-Digitar5.6/47 XL-100 (degree) MC. This (as you probably know) is symmetric layout like the Biogons. The Schneider “Digitars” have been designed specifically for digital imaging i.e. smaller image circles than what one would need on regular 4” x 5” frame. So I try to twist the arm of very seasoned and notable professional photographers to use this lens (for example on an ALPA) and usually with reluctant skepticism they use it as they don’t believe that $1200.00 lens could possibly be better than the lenses that they use that are between $4K to $10k++. The last comment I got back from a professional MF shooter was “Eric it is rumored that your brain must be connected and powered by the core of the planet…”. They are totally blown away by the resolving power of these lenses (which not always typical of a Biogon style layout). So for Architectural work etc. From the ground, one camera was set up in an old English church, and the photographer could show the resident priest in amazing detail near real time tiny pin dots of wood worm… across the whole frame of a 500 hundred year old carved wood paneled ceiling (100 feet above their head) . Remember this is a wide angle lens and its maximum radial distortion is in a few microns rather than greater than a few thousand microns as would be the case for a lens that has a more typical 2% distortion. Admittedly the sweet spot of the lens is towards f-8 and f-11.
One of the reason we pursue Symetric lenses is so that we can calibrate and engineer far more accurately to extract 3d geometry and special stereo textures to build new types of extremely life like 3d/VR models (that are extremely efficient to render real time). Precision is important if you have to join together 3d geometries from multiple stereo camera positions so that you can capture more of the subject or scene without occluded geometries from only one single (stereo) view.
We tend to use wider stereo base separations to capture fine grained 3d surface texture as we are able (as by product) of the processes we use to re-sample the ranges of parallaxes into more comfortable total ranges of parallax WITHOUT flattening the fine 3d surface texture (i.e. no cardboard cut out effect) [we do have granted patents for this]...
I also go for the low distortion Biogons also as most “practitioners” of the art of s-3d are not aware that radial distortion causes heterogeneous Y parallaxes across the field of view. We have some Math Cad simulations that show a complex rippling pattern of good and bad Y parallaxes; basically a vertical lozenge shape from the centre of the stereo frame from the stereo overlap and the radial distortion of the left and right lenses overlaid. So I/we like the idea that we can get the S-3d as near perfect as anyone can IN CAMERA… rather than to have to degrade the imagery to compensate for a whole set of concatenated image transformations or worse multistep separately applied spatial transformations. I’m not sure in the “industry” how many folks are seriously re-sampling their stereo imagery to correct for radial distortion induced Y parallax. Most retro focal lenses have at least 2% radial distortion unless the lens is particularly long. In other words you can get a much better stereo effect for the audience as all the fine detail resolves properly without weird rippling patterns of Y parallax and less stereo fatigue (of course). [I like the fact that RED has a 4k laser projector in the works that will do justice to well executed S-3d, like nothing else].
Given than we don’t have the mirror box/DSLR/mirror/ shutter… for modern digital systems these cleaner optical approaches [As you say RON] can yield very beautiful results through the use of these types of symmetric lenses you mention. Would be great if these types of approaches can be fully realized on the EPIC/Scarlet especially for the wider angle lenses we can’t use right now.
For a lot of S-3d based applications (and I have been saying this for years) I think wide angle low distortion symmetric layouts are the way to go, and exactly as you say (RON) in some cases the retro-focal cinematic layouts introduce a ton of extra lense elements that ultimately add weight, expense and complexity AND take down to some degree the “cleanness” of the image. You can only munge through so many optical elements until you’re your basic MTF start to get a bit soft.
On the other hand the advantage of the primes is the complex mechanics and floating elements to compensate for breathing. I saw a few months ago a Stanley Kubrick movie (I don’t remember which one) but half of the pull in’s breath like hell. So I’m not 100% sure that breathing in s-3d is a huge problem or not as it may enhance the effect? And of course the cinematic primes are designed to be pretty sharp wide open. So it seems that some of the Leica range of lenses could be a good compromise and the simplicity of these lenses means potentially much higher precision complimentary mechanics can be engineered.
Lot to think about and DO about…
Eric
(sorry I didn’t mean for this to turn into “what I did last summer” (school essay) but seemed a good opportunity to explain WHY…?)
![ArriStand[1].jpg](/data/attachments/5/5810-db65c5465de1cbc068ae0a9257601010.jpg?hash=22XFRl3hy8)
![NewStyle[1].jpg](/data/attachments/5/5814-1987dbc4f68324052928245de671f01f.jpg?hash=GYfbxPaDJA)
