Re: fisheye or roundshot?

[Willem-Jan Markerink <w.j.markerink@xxxxx>]

On 15 Apr 00 at 13:49, Glenn Barry wrote:

> The easy way to see this is to put the lens on look through the viewfinder
> or ground glass and watch the degree of image distortion as you rotate the
> camera. Fisheye mapping and compression requires that this is quite severe,
> more so at the edges agreed, but it is present throughout the frame
> including the centre. Just look at a straight line in the middle of the
> frame. It must begin to curve the moment you rotate the lens due to the
> fisheye projection. There will always be some blurring.

It is claimed that only a single-pixel wide digital 'slit' can overcome 
this blurring problem....in that case, no pixel is exposed at more 
than one rotational position, unlike any photo-chemical registration 
with a meaningful slit-width (a slit would have to be smaller than a 
hair to mimick the digital one-pixel concept, which would kill the 
image due to diffraction (remember that the digital slit is not a 
slit, but only a recording single row of pixels, so diffraction is 
not relevant there).

Not quite sure whether it is valid to explain the blur itself with 
the varying magnification of a fisheye....a rectangular lens has the 
same magnification from corner to center, a fisheye does not I 
believe (this even goes as far as the claim that a 14mm fisheye has 
less variety from corner to center than a 17mm fisheye with the same 
overall angle of view, hence the fisheye 'effekt' of the 17mm is 
larger).
And since film speed of a rotating camera is depending on 
focal length, this is exactly the problem....no speed can suit both 
center and edge at the same time.

--                 
Bye,

Willem-Jan Markerink


      The desire to understand 
is sometimes far less intelligent than
     the inability to understand


<w.j.markerink@xxxxx>
[note: 'a-one' & 'en-el'!]