Philgp: I'm not sure if I'm missing something here ( I probably am), but would't the noisiness be offset if you took photos at a lower resolution setting in-camera, say, 16mp (resolution of Nex-5N)?
Or, alternatively, would the fact that there would be pixel overlapping mean that there would be distortions, if decreasing to anything larger than quarter the pixel count (to c. 6mp, half the pixels in each dimension), which would screw with the image really badly when viewed at 1:1, on the pixel level?
Low Budget Dave,Thanks for the constructive comment and the link. I had a quick look through the page, and it more or less sums up what I was trying to say. Interesting that not all cameras do this.
Michael Barkowski,Regarding the probability that, had Sony developed this for NEX-7, they would have emphasised its inclusion a la Fuji, it certainly sounds plausible.
I suppose the best way to go about creating an ultra-low noise lo-res image from NEX-7 would be: working from the 24.3MP raw file, and developing to JPEG so it's exactly half the number of pixels in each dimension (ending up with c. 6MP JPEG. Quite lo-res, so wouldn't need much JPEG compression, result: few compresson artifacts).
I take it that, through that method, the image noise would be greatly reduced as all available data would be taken from the APS-C sensor (assuming the RAW>JPEG processing engine performed some description of pixel binning, which I take it is a sound assumption)
Does that book go into the mechanics of how a camera of a given resolution chooses and processes data that is fed into a lower-resolution image?
Or do you mean that at lower resolutions, the camera is instructing the sensor to take light data from fewer pixels into account? That would save on processing power (whether in-camera or in-computer) but it would be the same size file and the quality would be worse.
That doesn't make sense. If you have the camera set to take lower res, then the active area in the sensor is reduced? Not sure I follow your logic. Surely the active area of the sensor (the area the light falls on) will be the same, but there will be, say in the case of a 6mpx setting in a 24mpx camera, for simplicity's sake, a 2x2 area from which the imaging engine can take an average. Each 2x2 pixel area would be roughly as large as a pixel from a full-frame 8mpx camera, doesn't that scream "high image quality and noiselessness" to you?
I'm not sure if I'm missing something here ( I probably am), but would't the noisiness be offset if you took photos at a lower resolution setting in-camera, say, 16mp (resolution of Nex-5N)?