Great URSA 12K: but do not dismiss super 16
I wanted to compare the various resolutions of the URSA Mini Pro 12K as to what comes from the camera and no downsizing or down-sampling in Resolve. To do this I chose to make several time laps clips at full framing, 6 frames per second and Q0. (Exposures were all constant.) These comparisons concerned mostly the resolution as that is what was varied. Other aspects of image quality did not really show differences. I have also ignored any concern that that super16 mode has a big crop factor.
I first checked two lenses: Zeiss Otus 28mm and Canon 100-400 L IS II USM. The Otus clearly showed differences with 12K the most, then 8K, then 4K the least. Then Canon showed the 12K and 8K to be about the same and the 4K less. So I eliminated the Canon and only used the Otus for the comparison.
I then made clips as above at 12K, 8K, 4K, 6K super16 and 4K super16.
The clips were compared in Resolve at 1:1 using a 4K timeline and image scaling of center crop with no resizing. This was also done with a 12K timeline. No downsizing of anything except what the camera may have done.
The results showed the 12K clips to be the "best". The 8K and 6K super16 clips were next and very, very close to each other but clearly lesser than the 12K. The 4K and 4K super16 clips were the least and very close to each other.
What surprised me were the results of the super16 clips. I expected the 8K to show some real difference to the 6K super16. I also expected the 4K down-sampled off the whole sensor to be better than the 4K super16, but not so. I have yet to try down-sampling using resolve, but I now suspect that it may prove better then what the camera is giving. I also what to find if with a 4K deliverable what down-sampling might provide. At this point I don't see the camera doing this with it's down-sampling. The other image quality aspects like the great color may just be from the great sensor design alone. Might something be missing in the way this camera is down-sampling?
In seems though that super16 clips from the URSA 12K can hold their own quality wise and save some of that that precious file space. (or use it for 4K 240fps)
As another interest I have found that for a 4K deliverable the Otus can be zoomed 3X in Resolve without any loss when shot at 12K. (The Canon can be zoomed 2X.)
I wanted to compare the various resolutions of the URSA Mini Pro 12K as to what comes from the camera and no downsizing or down-sampling in Resolve. To do this I chose to make several time laps clips at full framing, 6 frames per second and Q0. (Exposures were all constant.) These comparisons concerned mostly the resolution as that is what was varied. Other aspects of image quality did not really show differences. I have also ignored any concern that that super16 mode has a big crop factor.
I first checked two lenses: Zeiss Otus 28mm and Canon 100-400 L IS II USM. The Otus clearly showed differences with 12K the most, then 8K, then 4K the least. Then Canon showed the 12K and 8K to be about the same and the 4K less. So I eliminated the Canon and only used the Otus for the comparison.
I then made clips as above at 12K, 8K, 4K, 6K super16 and 4K super16.
The clips were compared in Resolve at 1:1 using a 4K timeline and image scaling of center crop with no resizing. This was also done with a 12K timeline. No downsizing of anything except what the camera may have done.
The results showed the 12K clips to be the "best". The 8K and 6K super16 clips were next and very, very close to each other but clearly lesser than the 12K. The 4K and 4K super16 clips were the least and very close to each other.
What surprised me were the results of the super16 clips. I expected the 8K to show some real difference to the 6K super16. I also expected the 4K down-sampled off the whole sensor to be better than the 4K super16, but not so. I have yet to try down-sampling using resolve, but I now suspect that it may prove better then what the camera is giving. I also what to find if with a 4K deliverable what down-sampling might provide. At this point I don't see the camera doing this with it's down-sampling. The other image quality aspects like the great color may just be from the great sensor design alone. Might something be missing in the way this camera is down-sampling?
In seems though that super16 clips from the URSA 12K can hold their own quality wise and save some of that that precious file space. (or use it for 4K 240fps)
As another interest I have found that for a 4K deliverable the Otus can be zoomed 3X in Resolve without any loss when shot at 12K. (The Canon can be zoomed 2X.)
AMD Threadripper 1950x 16-core 3.4 GHz
96 GB Crucial DDR4 2666 ECC UDIMM RAM
AsRock Fatal1ty x399 motherboard
AMD Radeon Pro WX 8200 GPU
Windows 10 Pro 64-bit version 22H2, build 19045.3208
DeckLink 4K Extreme 12G
iPad Pro M2
96 GB Crucial DDR4 2666 ECC UDIMM RAM
AsRock Fatal1ty x399 motherboard
AMD Radeon Pro WX 8200 GPU
Windows 10 Pro 64-bit version 22H2, build 19045.3208
DeckLink 4K Extreme 12G
iPad Pro M2