The RED SCARLET and Sony F3 Trade Blows in a Head-to-Head Test and... Both Look Good
DP Timur Civan, who lensed my RED SCARLET test short and now has a SCARLET of his own, has done an apples-to-apples comparison with two of the top similarly-priced Super35 motion picture cameras currently on the market: the Sony F3 and the RED SCARLET. Timur has the luxury of owning both -- and some flawless Cooke Panchro primes -- and has set up a nice skin tone and still life test. Here it is:
My conclusion from watching Timur's test: I don't know. They both look great. I can't help but think that digital motion picture cameras are at a point where, if you can't tell your story with either one of these... the camera is not the problem. Plus, because both cameras give you a very gradeable image (the F3 with the now-standard S-Log option and the SCARLET using REDRAW), you can push the images in a lot of different directions.
Timur's own conclusions can and should be read in full using the link below (where he compares not just the images you see in the clip but also workflow considerations). To pull a short quote:
This real issue is that it comes down to image quality. I find myself shooting films and documentary on the F3, but commercials, music videos, and corporate spots on the RED. When it needs to look natural, I use the F3. When it needs to look crisp and slick, RED.
One interesting thing about this comparison: the F3 is more sensitive and has greater dynamic range, but is "only" 1080p. It's rumored Sony will announce an F5 with 4K output (or at least an upgrade path to 4K output) at NAB, addressing this concern. Meanwhile the SCARLET already has 4K resolution but is noisier/less sensitive. RED's forthcoming Dragon sensor upgrade for the SCARLET will attempt to address this concern. Both companies are already targeting the chief complaint about their respective cameras... and they already make great images. Not a bad time to be shooting!
Finally, as with all camera tests: different tools for different jobs. There is no such thing as "one camera to rule them all," which is why it often makes sense to rent (but not always). Any observations of your own from watching Timur's test?