>>3862092>>3862761Oh, well since it isn't Bayer, then it's actually even worse. That "12K" camera can only give you 3072x1620 4:4:4 RGB information after downsampling unless they're using some kind of patterned off-white white where some whites offer some blues and reds, but then that would defeat the purpose of deviating from bayer using white in the first place.
Image related.
>Left = Bayer.One or more RGB samples per 2x2 section.
>Middle = This cancer.You need 4x4 sections minimum to get regular RGB samples per section. Even 3x3s lead to huge gaps missing a single blue/red sample. It's just a whole lotta white. Worse than X-Trans. (Full color at just 3x3)
>Right = The only proper way to do RGWB.Every 2x2 chunk yields a pure luma sample and one of every color.
Remember we're being objective here, these people clearly want to minimize luma noise and just interpolate from shitty color and do "color grading" (faking colors completely) to compensate. This is a flawed tool but for their use it's likely good. That doesn't mean their sensor design is anywhere near objectively ideal because it's definitely not. They are very clearly prioritizing getting lower noise at the cost of chroma resolution and there's no way around that. That's 12K res, but severely compromised chroma.
An 8K bayer sensor would have more chroma information than this. The light gathering will be less on the 8K bayer since it lacks white but it would certainly capture more accurate colors since it'd be capable of 4:4:4 4096x2160 vs 4:4:4 3072x1620 from this 12K half panchromatic weirdo.
