p.5 #3 · Sony A900/850 users jumped ship to Nikon D 800/E?
AhamB wrote:
I can't imagine that would be the case; it wouldn't make sense at all because the cut-off frequency would be way too low.
I also went from the Canon 5D to the a900 and my experience led me to strongly suspect that the 5D "classic" used a very weak AA filter. Per pixel detail was very high and moiré was very easy to induce. While I much preferred the files from the a900, the per pixel detail was not at the 5D "classic" level and moire was pretty much non existent by comparison. 5DII files do not have the per pixel detail of the 5D "classic" which would lead me to believe the AA filter is stronger.
p.5 #4 · Sony A900/850 users jumped ship to Nikon D 800/E?
Aehm if you stick same strenght of AA on same size of sensor with very different size of cells, dont you think effect will be much less on bigger cells?
p.5 #5 · Sony A900/850 users jumped ship to Nikon D 800/E?
Mescalamba wrote:
Aehm if you stick same strenght of AA on same size of sensor with very different size of cells, dont you think effect will be much less on bigger cells?
It will be indeed as your Nyquist rate on the lower-resolution sensor will be lower
p.5 #6 · Sony A900/850 users jumped ship to Nikon D 800/E?
AA filter strength is for practical use measured as "perpendicular AA spread width divided by pixel pitch". That is, the amount of spread a ray hitting the sensor head on gets (in µm) divided the pixel size (also in µm). And BTW - there is no such thing as a sharp cutoff at Nyquist, especially not regarding AA filters. Firstly, no point in object space (the target) is infinitely small - and secondly, we're looking at a lossy optical system here, not a mathematical model. If the pixels in modern cameras were 30-40µm big, then the mathematical model might have been applicable. With 4-5-6µm pixels, it's not. The losses in the other parts of the system is to big.
Mescalamba might be partly right about the 5D vs 5Dmk2, the pixel pitch difference isn't that big.
sqrt(21MP) / sqrt (13MP) = 1.27 >>> If you use the 5Dmk1 filter on the 5Dmk2, the pixel array (and the raw interpolation) would see the filter as "27% stronger per pixel". Not a VERY big difference, though I actually think the difference in reality might be slightly less than that.
