From: into@oblivion.nothing.com   
      
   "anorton"  wrote in   
   news:g4SdnZkrFNyqkHTQnZ2dnUVZ_gmdnZ2d@earthlink.com:   
      
   > I think you are coming to the right conclusions.   
   >   
   > I looked at your original link. A big problem with the test outlined on   
   > that page is that it does not measure both horizontal and vertical   
   > resolution at the edges.   
      
   Well, here's the funny thing, I kinda thought of that and made a   
   test board with multiple copies of that test chart both vertical   
   and horizontal, both in the center and at the outside corner. I   
   also put a couple in between at a 45 angle.   
      
      
      
      
   > Also keep in mind that your resolution measurement result is   
   > the combination of your lens and sensor.   
      
   Yes, I do realize that one. Further, I also realize that since I   
   am using a different camera and sensor than DPReview, my results   
   aren't directly comparable. But I'm just looking for a ballpark   
   idea of where my lens resides.   
      
      
   > I believe the slanted-edge test that dpreview   
   > uses is less sensitive to the effect of pixelation.   
      
   Yes, I found out that DPReview uses Imatest, whose website has some   
   information that was useful to me. But it's not worth paying for it   
   since I'm only testing one lens. But the site is still chock full of   
   good info.   
      
   I have no delusions of making as highly detailed and accurate a   
   test as on DPReview.   
      
   I'm still a bit concerned that I'm doing the test right. I've only   
   done limited test shots to find the sweet spot f/ratio for my   
   lens at 28mm (it's a 28-300 superzoom). The resolution I came up   
   with was around 1000 lw/ph. That's higher than I expected.   
      
   But then, in reviewing that first website I think I had the camera   
   too far away. There are formulae for determining the film plane to   
   chart distance based on the focal length. The thing is, I was   
   multiplying by 1.5 since I have an APS-C sensor, thus using 42mm in   
   my calculations. But then other info I found leads me to believe I   
   shouldn't do that, and that therefore I was 1.5x too far away, which   
   means my results needs to be multiplied by 1.5x, which means my leans   
   was doing 1500 lw/ph. This doesn't make sense as that's up there with   
   the $2000+ lenses and this was less than $500.   
      
   The whole reason I'm even doing this is that I seem to have a   
   terribly soft focus. No matter how much I fuss, my images are just   
   a tad too blurry. And, I'm wondering if I need to invest in more   
   expensive glass.   
      
   BTW, am I right in that if my vertical resolution is 2592, my nyquist   
   frequency would be half that? 1296 line widths/picture height? If so,   
   then that definately means I'm doing something wrong as there's no   
   way I could be measuring better resolution than the nyquist frequency.   
      
   I think I need to go back and read everything from the beginning again.   
   I surely missed something.   
      
   Brian   
   --   
   http://www.skywise711.com - Lasers, Seismology, Astronomy, Skepticism   
   Sed quis custodiet ipsos Custodes?   
      
   --- SoupGate-Win32 v1.05   
    * Origin: you cannot sedate... all the things you hate (1:229/2)