Seems people missed the OP's exact question. 4K 16:9 is *NOT* 4096x2160, it's actually 4096x2304 and is a 16:9 aspect ratio.
RED One "4K" modes are as follows:
4.5K = 4480x1920 | 2.33:1
4K 16:9 = 4096x2304 | 16:9
4KHD = 3840x2160 | 16:9 (this is commonly known as quad-HD or 4xHD 1080p)
4K2:1 = 4096x2048 | 2:1
The differences between 4K 16:9 and 4KHD are that the 4K 16:9 frame is actually "4K" in that it's 4096 pixels across. Being slightly larger, it is closer in size to the full Super35 frame, but not by much -- it's very close to 35mm Academy format. The 4KHD mode has some advantages as it provides quicker scaling to 1080p HD since it is even 2X scale in both linear directions or a total of 4 times the number of pixels.
for 1080P delivery... it should also be noted that 4K HD transcodes much faster than 4K 16:9 - because the math is perfect - no rounding. either and faster with our without Red Rocket
A wider field of view will look different if filmed by a 35mm lens at 5K than it will from a 24 mm lens at 4K. Not just sharper, but different focus falloff, not to mention signal to noise on final image is much higher at 5K - you are using more light to create your image. AT 4K you are zooming in on noise, however nominal that noise may be on MX. Also, side-by-side, even on Red One, 4K 16:9 looks creamier and less digital than 4KHD.
This kind of simplistic, vaguely brutalist advice bugs me - it's like when people say "just adjust your iris if you want less light".
There is more to all this than counting pixels...
Seriously though, if there's no realistic route to a piece being seen theatrically, then it will live its life as a piece of 1080P rec709 media (at best) destined for compression to less than 8mb/s in MPEG-2. For that pathway I think the slightly crunchier edges in the "brute force" full integer downscale might actually provide more usable metrics for the encoder's algorithms. If so, that would typically (some high contrast or extreme exposure samples excepted) improve the IQ of the encoded media the viewer will see.
On the flip side, for any narrative or "glossy" piece that lends itself to a creamier look I'd shoot the biggest frame I can (love that 5,120 by 2,700 on the Epic) with a 16x9 guide for reference and manage the scaling filter type, contrast distribution, saturation, curves and other values during grading to reveal as much or as little of the true resolution of the base image as suits creative intent/cosmetic imperatives/etc.
Cheers - #19
Simplistic? I believe I stated exactly why 4KHD came to be, it's benefits, and pointed out that I could also get that little bit of extra width if I were up against a wall and needed it.
Brutalist? Are you serious? What planet are you from where switching between 4K 16:9 to 4KHD is some kind of violent behavior?
I hardly think that a few wide 4K shots are going to jump out of a film that is otherwise 4KHD. As someone who is focused on the entire production process (pre-prod, prod, and post) I stand by my advice. Not everyone has a RedRocket card and all the wonderful support hardware in their workstation, and there is a significant difference between the two when editing and rendering.
There are reasons that Red created the 4KHD format. One is the math gets easier for the computer and oh, BTW: it's more accurate too. Scaleing in any other fashion creates "creamyness" or as others would call it "softness" due to pixels being unevenly divided and added to each other.. If you want softness you can add it in post easy enough (oh dear!). I have yet to see a sharpening algorithm that actually worked.
Hey Rob, I actually have 4K 2:1, 3K 2:1 and (horrors!) 2K 2:1 in the same film!!! I had to because I needed slow motion... Had to do it. Guess it's not worth watching... Too brutal...
|« Previous Thread | Next Thread »|