Thanks for the explanation. I have to question how reality-based that thinking is. I do not, of course expect you to defend it.

It seems to me that nearly all photographers who are particularly concerned with image quality shoot raw and use third-party processing software. Perhaps that's a decision not rooted firmly in reality, but it would take a massive effort focused on software UX to get very many to switch to first-party software.

> Raw files are where you can compensate for that, with the least impact on image quality. You can have ChrAb correction, applied after the demosaic, but it will be "lossy."

Are you saying that they're baking chromatic aberration corrections into the raw files themselves so that third-party software can't detect it? I know the trend lately is to tolerate more software-correctable flaws in lenses today because it allows for gains elsewhere (often sharpness or size, not just price), but I'm used to seeing those corrections as a step in the raw development pipeline which software can toggle.

I think we're getting into that stuff that I don't want to elaborate on. They would probably get cranky I have said what I've said, but that's pretty common knowledge.

If the third-party stuff has access to the raw Bayer format, they can do pretty much anything. They may not have the actual manufacturer data on lenses, but they may be able to do a lot.

Also, 50MP, lossless-compressed (or uncompressed) 16-bit-per-channel images tend to be big. It takes a lot to process them; especially if you have time constraints (like video). Remember that these devices have their own, low-power processors, and they need to handle the data. If we wrote host software to provide matching processing, we needed to mimic what the device firmware did. You don't necessarily have that issue, with third-party pipelines, as no one expects them to match.