>>3862840>You need 4x4 sections minimum to get regular RGB samples per section. Even 3x3s lead to huge gaps missing a single blue/red sample.Not exactly.
One way to look at it is, 2x2 of bayer is gonna be “equivalent” to 3x3 of the new design, but the new design will have double the amount of red information.
So colour-wise 12K of the new design will be roughly equivalent to 8K of bayer with better blue and potentially better DR or whatever they choose to do with the white pixels. Assuming they get their demosaïcing right and there’s software support etc etc. .
Not bad deal I’d say, and if I were to hazard a guess I’d say the output res would be 8K.
And if they decide to offer a higher res it’d be an upscaled thing, for “street cred” and to emphasise the new sensor and make a dig at bayer, kinda like what Foveon did with “le 45 mega pickles” or what Fuji did in the past with their SuperCCD sensors.
Which Btw seems like a very similar implementation with the “white” pixels and all, and if it works anything like it, it’ll have crazy DR when implemented with modern sensor tech. Fuji was a solid 5-10 years ahead vs the competition when it came to highlight DR and their SuperCCDs (as a reference, even those ancient mid 00’s cameras have ~13.5EVs DR).
Extended highlight DR would make sense on a cine camera.
