[Tig] Is there anybody out there

[Bob Friesenhahn]

On Tue, 1 Aug 2006, Peter Swinson wrote:
>
> Where is everyone.
>
> Not all doing Bubble Rise Rate Experiments I hope.

It is the day after heavy Bubble Rise Rate Experiments. :-)

> Let me put this up to be knocked down.
>
> Granular image sampling is the main reason film outstrips any HD or other
> electronic capture medium. Or is it.

It seems to me that the mean reason why film is still considered 
"better" than electronic capture is due to the wider dynamic range of 
film.  But high end electronic capture provides sufficient dynamic 
range and usually looks cleaner than film originated material.  "HD" 
usually signifies Rec.709 colorspace which implies a specifically 
limited gamut most suitable for unaltered reproduction and not for 
substantial post-production.

Another issue is that electronic capture based on a Bayer pattern may 
not provide as much actual resolution as a film transfer since RGB 
values are computed from a dithered spacial matrix.  The real 
resolution obtained from a Bayer sensor depends quite a bit on the 
honesty of the camera vendor.

The nyquist sampling theorem 
(http://en.wikipedia.org/wiki/Nyquist-Shannon_sampling_theorem) which 
is often used to explain sampling resolution actually only applies to 
*sine* waves so if the image data (or harmonic thereof) does not 
perfectly fit a sine wave of 1/2 the sampling frequency, then the 
actual supported resolution must be lower.  According to fourier 
analysis, any waveform may be recreated using sine waves of particular 
frequency and amplitude.  In order to be able to replicate the 
original, one must be able to replicate the sine waves.  The output 
replicated from the samples must be perfectly low-pass filtered to 
allow only sine waves up to 1/2 the sampling frequency to be passed. 
Film is reputed to support as much as 6K resolution in 35mm (evaluated 
by projecting in "analog" form on a screen) so it has a spectral 
advantage over "HD".

One last issue is that most HD capture sensors are smaller than 35mm 
so it is more difficult to produce the high Modulation Transfer 
Function (MTF) lenses (back to fourier theory!) required to apply the 
same resolution image on these smaller sensors.

BRRE. QED.

Bob
======================================
Bob Friesenhahn
bfriesen at simple.dallas.tx.us, http://www.simplesystems.org/users/bfriesen/
GraphicsMagick Maintainer,    http://www.GraphicsMagick.org/