Now the final buffer specs are out, I'm left with one question about performance.
In Standard DR mode, the FPS is 3. Not fast, not too slow considering it's not a sports cam, etc..
In Wide DR mode the FPS is 1.5. This IS slow.
What I am wondering about, though, is why the framerate drops for wide DR shots. The only conclusion I can reach myself is that in wide DR shots, all 12 million sites are used and thus the amount of data on the memory bus is doubled.
However this would imply that in Standard DR mode, the extra photosites are not being read off,meaning that only 6mp are actually in use.
This makes the leaping assumption that the FPS is sensor-> memory bandwidth limited. I'm not sure of the computational overhead in the JPEG engine but I would have thought that if computation was the bottleneck that you would see better FPS in RAW mode (No JPG DR calculation.)
SO I'm left thinking that getting the data off the CCD is where the bottleneck ocurrs and amd thus left wondering if all 12 million sites are put to use in Standard DR mode.
If they're not, then we only receive 3fps when using the camera in a crippled mode, as the whole point of the S5 is the S and R pixel combination.
Does anyone out there know how this fits together?
In Standard DR mode, the FPS is 3. Not fast, not too slow considering it's not a sports cam, etc..
In Wide DR mode the FPS is 1.5. This IS slow.
What I am wondering about, though, is why the framerate drops for wide DR shots. The only conclusion I can reach myself is that in wide DR shots, all 12 million sites are used and thus the amount of data on the memory bus is doubled.
However this would imply that in Standard DR mode, the extra photosites are not being read off,meaning that only 6mp are actually in use.
This makes the leaping assumption that the FPS is sensor-> memory bandwidth limited. I'm not sure of the computational overhead in the JPEG engine but I would have thought that if computation was the bottleneck that you would see better FPS in RAW mode (No JPG DR calculation.)
SO I'm left thinking that getting the data off the CCD is where the bottleneck ocurrs and amd thus left wondering if all 12 million sites are put to use in Standard DR mode.
If they're not, then we only receive 3fps when using the camera in a crippled mode, as the whole point of the S5 is the S and R pixel combination.
Does anyone out there know how this fits together?