On Wed, Oct 09, 2013 at 03:00:10PM -0400, Postmaster wrote:
> Larry Colen wrote:
> 
> >It seems to me that if you had a rear illuminated sensor, with no space
> >between the pixels, and it had no bayer filter, then aliasing/moire would
> >not happen, because the light value would be averaged over the whole sample.
> >
> >It's the discontinuous aspect of what is effectively three overlayed photos
> >that is causing the aliasing.
> >
> >Is this correct?
> 
> Nope. Aliasing is possible whenever the (spatial) frequency of the
> image is higher than that of the sensor. The interesting thing is that

I suspect that there are several things going on.  I don't have
the math, or the time right now to work it out.  The Bayer array will
effectively reduce the sampling frequency by a factor of two or four.

There is also, as has been pointed out, a difference between instantaneous
sampling and integrated sampling.  Integrated sampling will greatly reduce
the magnitude of the aliasing, and I suspect the character of it.

The clever arrangement of color pixels on the fuji sensor probably addresses
the situation in much the way that film did.  In a pathological case of printing
out an image of the fuji color pattern and shifting the image fractions of 
pixels,
you would probably see cases of aliasing.  

> with 24 megapixels in an APS-C sensor you probably won't need an
> anti-aliasing filter in all but the most extreme cases (the absolute
> highest quality glass (at its optimum aperture), tripod mounting,

Not to mention shooting at an aperture below the diffraction limit, and
most importantly, taking a photo of something with a spatial pattern that
is above the nyquist frequency.

> mirror lock-up, high shutter speed). The Nikon D800 is notorious for
> demanding only top-shelf lenses and the pixel density of a 24MP APS-C
> sensor is even higher.

It's not like the images from lenses on a D800 would look any worse than
if those lenses were mounted on a D700.  

So, a Mazda Miata is 3950mm long, lets call it 4,000. A 24MP sensor would be
4,000x6000 pixels.  So, if you were to paint a miata in a checkerboard, of
1mm squares, and take a picture with it slightly less than filling the sensor,
then you would start to see moire.
Although, with the lower sampling frequency due to the bayer filter, those 
squares could be as large as 2mm or 4mm.  
Actually, what I think Doug should do is paint his racecar in a bayer filter 
of 2mm squares, just to screw with the trackside photographers.
 
> 
> 
> 
> -- 
> PDML Pentax-Discuss Mail List
> PDML@pdml.net
> http://pdml.net/mailman/listinfo/pdml_pdml.net
> to UNSUBSCRIBE from the PDML, please visit the link directly above and follow 
> the directions.

-- 
Larry Colen                  l...@red4est.com         http://red4est.com/lrc


-- 
PDML Pentax-Discuss Mail List
PDML@pdml.net
http://pdml.net/mailman/listinfo/pdml_pdml.net
to UNSUBSCRIBE from the PDML, please visit the link directly above and follow 
the directions.

Reply via email to