01-13-2012, 09:49 AM
[quote name='waardij' timestamp='1326447251' post='14826']
Can I conclude from the above that you tested the M9 yourself and observed symmetrical resolution? Sorry if I am a bit persistent on this, but I am just wondering why Sony would use only a single AA filter. NEX cameras are well known for aliasing. I observe this a lot myself (it is not something I mind, I like the way a NEX can render fine detail in landscape photography). The thing is, when I look at a picture from a Siemens star (for lack of a better name), I see no asymmetry at all. When looking at resolution plots where vertical and horizontal are split up (the sets of lines of which the distance gets smaller and smaller), I do see a difference, although there is aliasing on both the horizontal and the vertical axis. What I am asking is, if it is certain this is caused by an AA filter (a single filter, only one direction), or that it is a result of the used demosaicing algorithm. That algorithm is part of the chain true which we observe the resolution. Or did I miss something altogether and are there brands that specify they have used an asymmetrical AA filter?
[/quote]
Can't comment on the M9 but regarding the NEX 7 it is very obvious in the lab analysis. The horizontal max. LW/PHs are about 30% higher than the vertical ones (based on the current parameter set). This is also visible if you look hard.
However, even with AA you will run into moirees all the time - just check out the RAW analysis over at some of dpreview camera reviews.
As someone already mentioned it is probably just a matter of costs. It is probably hard to design AA filters for very small photodiodes. Olympus struggled ages to come up with a solution here for instances.
Markus once send me a convertered M9-RAW via CaptureOne - which features an Anti-Moiree option - it looked quite Ok to me (from a real world perspective).
Can I conclude from the above that you tested the M9 yourself and observed symmetrical resolution? Sorry if I am a bit persistent on this, but I am just wondering why Sony would use only a single AA filter. NEX cameras are well known for aliasing. I observe this a lot myself (it is not something I mind, I like the way a NEX can render fine detail in landscape photography). The thing is, when I look at a picture from a Siemens star (for lack of a better name), I see no asymmetry at all. When looking at resolution plots where vertical and horizontal are split up (the sets of lines of which the distance gets smaller and smaller), I do see a difference, although there is aliasing on both the horizontal and the vertical axis. What I am asking is, if it is certain this is caused by an AA filter (a single filter, only one direction), or that it is a result of the used demosaicing algorithm. That algorithm is part of the chain true which we observe the resolution. Or did I miss something altogether and are there brands that specify they have used an asymmetrical AA filter?
[/quote]
Can't comment on the M9 but regarding the NEX 7 it is very obvious in the lab analysis. The horizontal max. LW/PHs are about 30% higher than the vertical ones (based on the current parameter set). This is also visible if you look hard.
However, even with AA you will run into moirees all the time - just check out the RAW analysis over at some of dpreview camera reviews.
As someone already mentioned it is probably just a matter of costs. It is probably hard to design AA filters for very small photodiodes. Olympus struggled ages to come up with a solution here for instances.
Markus once send me a convertered M9-RAW via CaptureOne - which features an Anti-Moiree option - it looked quite Ok to me (from a real world perspective).