Upscaling 2.5K ArriRaw?

[Max Field]

Forgive me if this question is completely off base.

 

Have any of you ever recorded the output of an EV camera to a 7Q and captured/upscaled in 4k?

Was the client able to notice? Would the portable recorder even give you the option?

 

Thanks.

[Stuart Brereton]

Arri Alexa material is upscaled to 4k all the time, and no one can ever tell, unless they are pixel peepers.

[Max Field]

Arri Alexa material is upscaled to 4k all the time, and no one can ever tell, unless they are pixel peepers.

Now is this solely because ARRI footage comes with such a great quality, or could virtually any 1080p/2.5k camera image be upscaled to 4K without anyone really noticing?

[David Mullen ASC]

Depends on the quality of the original -- ARRIRAW is uncompressed so there wouldn't be the smoothing and softening that comes with compression, which is one reason why 2.8K ARRIRAW footage upscales to 4K better than, let's say, 1080P from a DSLR. And it's 2.8K, not 1080P!

 

Amazon shows (like the one I'm shooting) are often using 3.2K 16x9 ProRes 4444 on the Alexa and upscaling to UHD (3840 x 2160) with excellent results.

 

But sharpness is often a perceptual issue too -- footage shot on a sunny day under hard light will look sharper than footage shot on an overcast day. And shallow focus photography often gives the illusion that the subject is sharper simply because it pops out from a soft background more clearly.

[Stuart Brereton]

Have a look at Steve Yedlin's resolution tests. He compares upscaled 2k material with native 4k and 11k downscaled material. When it's all viewed at 4k, the differences can only be seen by pixel peeping. 4k is really the sweet spot of human perception for most purposes.

[M Joel W]

On 7/22/2018 at 9:36 PM, David Mullen ASC said:

Depends on the quality of the original -- ARRIRAW is uncompressed so there wouldn't be the smoothing and softening that comes with compression, which is one reason why 2.8K ARRIRAW footage upscales to 4K better than, let's say, 1080P from a DSLR. And it's 2.8K, not 1080P!

 

Amazon shows (like the one I'm shooting) are often using 3.2K 16x9 ProRes 4444 on the Alexa and upscaling to UHD (3840 x 2160) with excellent results.

 

But sharpness is often a perceptual issue too -- footage shot on a sunny day under hard light will look sharper than footage shot on an overcast day. And shallow focus photography often gives the illusion that the subject is sharper simply because it pops out from a soft background more clearly.

Sorry to revive a dead thread but have you compared 2.8K ARRIRAW with 2K ProRes4444? 

I've been working a lot recently with 2k, 3.2k, and 4k (UHD) and 4K+ LF open gate footage. And the difference in sharpness is less than I'd expect although LF footage is clearly sharper. But I haven't had the chance to compare side-by-side (of the same scene, that is) directly. 

Asking myself if the huge amount of effort that would go into shooting ARRIRAW is worth it. My preference is for a "softer" image but I also tend to use soft lenses (would shoot with standard speeds, cine-xenons, K35s, Lomo standard speed or LOMO anamorphic even though Rokinons look nice on Alexa to me too).

Edited January 7, 2022 by M Joel W