November 12, 2017 at 5:32AM


Alexa 3.4 K for 4K DCP

I read that Blade Runner 2049 was shot on Alexa open gate mode 3.4K and delivered for 4K DCP.
I just didn't understand how did they do it. Like they literally scaled a little bit&|? Or there was some anamorphic trick in the picture?
Unfortunately for now I havent work with alexa and don't have experienced DIT in my area, so i would glad to know your answers on this subject.


The ALEXA ST open-gate image format is 3414 x 2198 pixels, so scaling up to 4096 x 2160 would not be a huge difference, especially if they shot with anamorphic 1.5x lenses which would make the final image format 5121 x 2198 pixels.

November 13, 2017 at 12:26PM

Guy McLoughlin
Video Producer

the good is that Roger use Alexa XT in opengate, most of bluray and projection 4k are upsampled from 2k final master.
3,4 to 4k is a small step, 2k to 4k is big step, you double size in horizontal and vertical, 4 times original, less detail and often worst result.
If you project 2k dci with 4k projector you see a good result, be cause you have better and powerful light for all single pixel, when you see an upsampled 2k to 4k often there are used a unsharp mask to force contrast, then you can see better the aliasing aroung dark pixel on light background, and not only, often the postprocessing add the "clarity" the worst process you can see in an upsampling be cause change a good photography to soap opera photography.
Shooting Opengate is excellent way to grab all infos that you need from picture, to manage stabilization in post, a small reframing and more for postproduction.
If you shoot directly in 2k you risk every single postprocess blur your pixel.
Not only, if you shoot in raw or in prores you have different option of upsampling.
If you shoot raw you can upsampling during debayering process and picture are better then scaling in other point of workflow.
Algorithm of scaling and sharpening applied to raw or simple uncompressed shooting give you very different quality.
Scaling during debayering allow algorithm to work with full luminance info without limitation of partial color managment (bayer matrix is 4:2:2 color matrix), that mean sharpness is full, when you work on a picture that are debayered, also in 4:4:4, you have a picture where you have pixel of red and blu are rought then green (luminosity) when you upsampling it bleed aroud red and blu and you not have a so clear dectail, often software aware about this limits contrast only red and blu during scaling, but this cause aliasing and many other problems.
The choosing of final delivery before to shoot allow you to have more quality with same camera.

November 15, 2017 at 7:40AM

Carlo Macchiavello

Roger Deakins shot "In Time" in HD and upscaled it to 4k afterwards.... did you notice?

November 15, 2017 at 8:10AM

Torben Greve

Here is a great explanation of what you are asking about by Cooke Optics.

November 15, 2017 at 9:39AM


Just to tell you why 4k to cinema theather is a waste of time/bandwitch and more...

If you check the Snellen table to check the visual capacity to read detail of human's eye, you can discover that a 10/10 capable eye can see two different detail far 7,27mm at 5meters, but if you are in the middle of theather at 20 meters from a 24m screen, you eye can see only 29,09 mm detail difference, but...
a 2k dcp offer per line 9mm covering
a 4k dcp offer per line 4,5 mm covering
that mean you cannot see the difference, also if you have the eagle view....
Or you go at 2 meters from screen, and... not see the movie, only skin texture of actress :-D
obviously shooting higher res you can is good for reframing, stabilizing, optimizing of visual quality, vfx work and more, but the reason because Arri not think to produce an higher density chip for 4k, actually is clear.
More density, less light to single pixel, dinamic gamma suffer about it.

November 15, 2017 at 10:07AM

Carlo Macchiavello

January 1, 2018 at 3:14AM


Your Comment