IMAX 2x monstro test.

[Björn Benckert]

As the title says. I did a little test, have not seen it on the max screen yet but looks promising.

Basically shooting two monstro´s on top of each other where the top camera is upside done and slightly uptilted both with the same lens in this case a 35mm FD.

Stitching together the two plates gives more than a full width 8k 4:3 which is wats wanted for IMAX.


To me monstro just beat the shit of all other cameras, noise levles is so low its insane, looks like glas even at high iso, some of this shots I wrote 800 iso by mistake they where considerably higher at some of them.


possible to download in glorious 8k 4:3 prores to look between the pixels...


https://vimeo.com/251675921

 

[Ty Maier]

That’s crazy epic.....

 

[Karim D. Ghantous]

That's impressive that you were able to seamlessly match the cameras like that! I really liked the long shot of the traffic on the bridge.

Personally I would just use an anamorphic lens, mounted 90 degrees. Thoughts?

 

[Björn Benckert]

Anas dont have the needed sharpnes especially not along the edges. As it works you got the IMAX theatre horizon on less than 1/4 from the bottom of the screen, the rest is pretty much just a sky dome. So you really need to strive to get sharpness on the bottom edge of the frame. That you will not get with a vertical ana and especially not with the wider ones.

A large format fisheye that is offcentered, so center is at horizon would be theoretically ideal... Just need to find / make such lens which is not easy.

Thelast frame, the interior of the office is a shitty fisheye with an offset. Not sharp etc but such lens really create the VR feel to the Imax dome.

 

[Ketch Rossi]

Insane image quality!

 

[Kwan Khan]

Wow... Nicely done

 

[Björn Benckert]

Wow... Nicely done

Not much of a trick you can easily do it by filming straight forward and then tilt up so you have a little overlap, like 20% is enough. Then project those in a VR dome with same setting for the projectors as lens used when shot, in this case a 35mm gives a 33 deg vertical projector, and then tilt one up until the match and make a feather mask in between.

Thos shots are not much but it´s nuts see what the monstro delivers in term of low noise and color fidelity, those white skys there have full coverage just as the dark blacks of the grills of the cars... find that amazing.

Screendump from resolve with only iso changed drastically.

[/url]PastedGraphic-5-1 by Björn Benckert, on Flickr[/IMG]

 

[Peter Lyons Collister ASC]

pretty cool

 

[Karim D. Ghantous]

The Zeiss Master Anamorphics are apparently *too good* for some DPs. Apparently: no distortion, well controlled flares, and consistent sharpness. I would look at those - with a 2x TC, if that can work. I suspect that nobody has yet done that sort of experiment.

 

[Chad Lancaster]

very cool bjorn, I've been wanting to do a test like this to test its limits with various lenses and how close you can get to the subject.

 

[Björn Benckert]

The Zeiss Master Anamorphics are apparently *too good* for some DPs. Apparently: no distortion, well controlled flares, and consistent sharpness. I would look at those - with a 2x TC, if that can work. I suspect that nobody has yet done that sort of experiment.

We used those a lot. "no distortion" is pretty far from where they are. They actually distort like mad, the wider once actually go from very much pincushing the image (like a negative fisheye) to quite a bit of fisheye when pulling focus. Possibly so that the range of distortion is neutral on a few meters distance but when going from infinity to close focus things very much change. And no they are way far from sharp corner to corner compared even to a shitty spherical lens.

The thing with anas even the master anas is that they are far from even barrelled, in other words when you project them into a CG sphere things dont align, and when you undistor them you need to do so with a mesh not a normal lens distort tool and the distortion is very much uneven across the picture and its not on a gradual scale but more like a wobbly one. So for image stitching round lenses is far better and that also a fact when it comes to corner / edge sharpnes.


The thing with anamorphic and especially the wide ones is they do not look good when paning tilting when making the image perfectly straight. It looks way better when they ar a bit fishy. As if you make the image flat it travels way faster on the corners, and if its a bit of fisheye you dont feel that as much.

very cool bjorn, I've been wanting to do a test like this to test its limits with various lenses and how close you can get to the subject.

The distance to subject is relative to the distance between the nodal points of the lenses so if you can get the two lenses very close to each other then that helps. One camera upside down on the other camera is as close as you can get. Then the wider lenses you use the more tilt in between and that increase the distance.

But best of all is to shoot trough a mirror rig. Then the nodals can be perfectly aligned no matter how much tilt and they will stitch perfect even though the person comes right up to the mirror box. But then yes, you got all the size, dust and other problems that comes with the mirror...