Re: How does squinting defeat anonymity masking on television?

Dear Ken, 

I'll have to try this one myself next time I see it on TV.  I haven't 
actually seen this phenomenon, so anything I say is pure speculation, at 
this point, but I am familiar with several other phenomena that may be 
related.  Here goes nothing:

My guess is that this relates to the fact that our brains integrate visual 
information at any instant at a fixed scale.  Just like the optical 
illusions where you either see an old lady or a young one, but not both at 
the same time, it is hard for us to see things at two different scales 
simultaneously.  When the blocking is inserted into the picture, the 
broadcasters are basically averaging the pixels over a large square 
neighborhood.  This creates visual objects at a very different scale from 
the surrounding area that is not blocked.  By squinting, you are blurring, 
or filtering, the image so that the blocked area gets smoothed some and the 
more detailed areas not blocked are blurred some - actually bringing the 
scales of the two areas closer together.  When this happens, it is easier 
for your brain to integrate all of the contextual information and create a 
more vivid mental image of the blocked area.  

I have seen this a very similar effect in a number of cases.  
Example 1: If you look very closely at a picture (TV, Newspaper, etc) that 
is made up of dithered dots, you either see the dots, or the picture they 
create, but not both at the same time.  

Example 2:  I work in the imagery field and when you blow an image up to the 
point that the pixels become blocky, you really can't discern any 
information.  This form of "upsampling" is known as "Nearest Neighbor 
Resampling".  However, if a different form of upsampling, known as "Cubic 
Convolution Resampling" is used, the picture is much easier to view.  This 
can be explained mathematically through what is called "aliasing".  The NNR 
causes a lot of unnatural spatial frequencies to appear (kind of like the 
distortion of an electric guitar run through a fuzz box), but "CCR" 
suppresses many of these extra unnatural frequencies making it easier for 
our eyes to discern the patterns.  Mathematically, one can approximate the 
CCR resampled image from the NNR image by performing a low-pass filtering 
operation.  Squinting your eyes also performs a similar low-pass filter 
operation.  

Example 3:  Blow an image up on your computer until the pixels are square 
blocks (for example in Photoshop).  Then step back by about 6 feet.  The 
distance also lowers the resolution and you can actually see more of an 
image from a distance than you can up close.  This is because you are 
limiting the effects of the aliasing.  Perhaps squinting does the same?  You 
might try this, just for grins.  

Well, you've got my two cents on this effect.  I'll try it myself next time 
and see if you're right - though I'm pretty sure there is some truth to what 
you're saying.  

By the way, don't spend too much time in front of COPS - it'll warp your 
vision....

Cheers, 
Todd Jamison, Chief Scientist, Observera, Inc.