Cameras are already downsizing, DSLRs are long dead and mirrorless are not just the future, they're now the present. I don't think size is the issue any longer.
The purpose of using a "real" camera is quality and control. If you go on IG and look as all the top photography centric accounts they're using either film or mirrorless cameras, not iPhones. The ability to manipulate RAW photos, the outright quality of better sensors and lenses, and the ease of use of "real" cameras still makes them the top choice.
When you then throw in professional use into the mix the choice becomes more obvious. Just the ergonomics alone make traditional cameras the obvious choice. I'm not going to be mounting expensive shotgun mics, matte boxes, and filters to an iPhone.
As far as computational photography goes, it's the future and all manufacturers will need to start using it to get an edge on the competition. The thing with computational photography however is that it's a data driven process, just like anything else AI; data in is data out, so the better quality of dedicated cameras simply means better data for computational methods, thus better results in the end. There's no real substitution for better lenses and sensors, because any computational method applied to a phone can also be applied to a camera.
I think there's something to be said for the lack of control with computational methods too. my iPhone for example applies a lot of processing that makes skin to me look plasticy, but it's baked into the ISP. My Fujifilm when shooting in RAW or F-Log for example doesn't have this issue. So any computational methods applied to traditional cameras need to not impede on creative control; noise reduction is an area where it would be pretty universally helpful.
Cameras are already downsizing, DSLRs are long dead and mirrorless are not just the future, they're now the present. I don't think size is the issue any longer.
The purpose of using a "real" camera is quality and control. If you go on IG and look as all the top photography centric accounts they're using either film or mirrorless cameras, not iPhones. The ability to manipulate RAW photos, the outright quality of better sensors and lenses, and the ease of use of "real" cameras still makes them the top choice.
When you then throw in professional use into the mix the choice becomes more obvious. Just the ergonomics alone make traditional cameras the obvious choice. I'm not going to be mounting expensive shotgun mics, matte boxes, and filters to an iPhone.
As far as computational photography goes, it's the future and all manufacturers will need to start using it to get an edge on the competition. The thing with computational photography however is that it's a data driven process, just like anything else AI; data in is data out, so the better quality of dedicated cameras simply means better data for computational methods, thus better results in the end. There's no real substitution for better lenses and sensors, because any computational method applied to a phone can also be applied to a camera.
I think there's something to be said for the lack of control with computational methods too. my iPhone for example applies a lot of processing that makes skin to me look plasticy, but it's baked into the ISP. My Fujifilm when shooting in RAW or F-Log for example doesn't have this issue. So any computational methods applied to traditional cameras need to not impede on creative control; noise reduction is an area where it would be pretty universally helpful.