From: tobias@invalid.invalid   
      
   On 27.10.2015 00:13, RichD wrote:   
   > At an image processing seminar last week, I saw a startling presentation.   
   > The speaker noted that we cannot measure phase at optical wavelengths,   
      
   That's not entirely true, or at least not accurate. Interferometrically,   
   we can record complex-valued images, including phase. We just need   
   coherent illumination and a bit of that coherent light as a reference wave.   
      
   An extremely out-of-focus image can serve as an arbitrarily good   
   approximation of a uniform-phase illumination and serve as reference. I   
   don't have a citation for that kind of setup right now, but it's been   
   used for spatially phase-shifting interferometers which were effectively   
   self-homodyne.   
      
   So bascially, we can measure relative optical phases as long as   
   everything is monochromatic, even without access to the monochromatic   
   source..   
      
   If things aren't temporally coherent, what does "phase" even mean?   
      
   > To illustrate, he displayed two   
   > face photos, then swapped the phase, with no modification of the   
   > magnitudes.  The faces switched places, with slight distortion!   
      
   How is that surprising? With a 90° phase shift, you can have two   
   completely different intensity images in the real and imaginary   
   quadratures of a light field. In theory, there need not be any crosstalk   
   at all.   
      
   > Also, a theoretical question manifests - why is it, that most of the   
   > information is in the phase?    You'd expect 50-50.   
      
   Now this is the most interesting part of your post to me - could you   
   please elaborate what you mean by "most of the information is in the   
   phase" here?   
      
   --   
   Tobias   
   ...dragged out of lurking for the first time here. Hello.   
      
   --- SoupGate-Win32 v1.05   
    * Origin: you cannot sedate... all the things you hate (1:229/2)