From: dmichel@prosound.de   
      
   Hi Eric,   
      
   > Is it possible that an image is correctly sampled   
    > according to Nyquist [...]but nevertheless show some   
    > pixelization effect when sufficiently zoomed ?   
      
   to correctly spatially discretize and reconstruct an   
   image you need both an analog anti-aliasing-filter in   
   front of the sampling process and an analog reconstruction   
   filter when viewing the image.   
      
   In a properly designed system either the limited resolution   
   of the taking lens (objective) or a dedicated anti-aliasing   
   filter will limit the spatial bandwidth before the sampling   
   takes place.   
      
   At the viewer's end, there is normally no dedicated reconstruction   
   filter (e.g. in an LCD), but the limited resolution of the human   
   eye will serve as such. As soon as the pixels are large enough   
   to be resolved by the human eye, you can see them ;-)   
      
   If cannot resolve the physical pixel of e.g. an LCD screen,   
   but zoom into the picture, you will enlarge the image pixels   
   no matter if the image was sampled correctly.   
   In that case, you would either have to use a dedicated reconstruction   
   filter (spatial low-pass) such as an opal glass pane or you need to to   
   oversampling.   
      
   Oversampling converts the spatial sampling frequency of the   
   original image to e.g. that of the LCD monitor. The oversampling   
   process includes a spatial lowpass-filter so that - then again -   
   you won't see pixelation. If you omit proper reconstruction,   
   you can see pixelation even with a correctly sampled image.   
      
   Best regards,   
      
   Dieter Michel   
      
   --- SoupGate-Win32 v1.05   
    * Origin: you cannot sedate... all the things you hate (1:229/2)