I worked on camera drivers for mobile phone SoCs 5 years ago. Back then a strong constraint that separated "video" from "picture" was that we didn't have enough memory bandwidth to stream uncompressed "picture"-sized frames at "video" framerate.
Google's Camera2 API no longer presents that distinction.
Funny how things change, that's my personal experience with Moore's law.
Another problem we had is that we'd developed flexible OpenMAX-compliant modules to expose the hardware's capabilities, following Nokia's requirements. We tried pushing those to Google whose APIs back then were much more limited. They (as far as I knew back then) didn't care.
Google's Camera2 API no longer presents that distinction.
Funny how things change, that's my personal experience with Moore's law.
Another problem we had is that we'd developed flexible OpenMAX-compliant modules to expose the hardware's capabilities, following Nokia's requirements. We tried pushing those to Google whose APIs back then were much more limited. They (as far as I knew back then) didn't care.
Again, funny how things change...