Re: [APML] a short note on film flat fielding

[Wei-Hao Wang <wang@ifa.hawaii.edu>]

On Wed, 25 Feb 2004, Jerry Lodriguss wrote:

> I would have thought that you would need to use the same film to model the 
> film's reciprocity failure...
> 
> Also, wouldn't you need a flat field for each exposure length for the same 
> reason?
> 

Hi Jerry,

Very good questions.  You really get to the point.  I have to use 20 
times more words to answer the questions.  :)

One of the reason is, there is no hope to model reciprocity failure,
even with the same film.  RF is a function of strength of incident light,
air presure, humidity, and perhaps many other unknown factors.  Unless 
the experiment is carried in well controled environments such as those 
used for photographic plates 10s years ago in big observatories, I don't 
think I can handle RF directly.

However, I believe the effect of RF is taken into accout indirectly.  
All RF does is to modify the shape of the HD curve.  And such modification
is a function of incident light strength, i.e., the illumination pattern.
If all I did is to derive the fim response under certain illumination 
pattern, the effect of RF should be already included.

I would also like to emphasize.  The film is not linear.  The idea 
of dividing a film image with a flat is already wrong (but is perhaps the 
best I can do at this stage).  Comparing to the nonlinearity, perhaps RF 
is such a minor effect that it is not even worth thinking about it.

Furthermore, many people think that RF is a function of exposure time.
This is not quite correct.  

## RF is a function of incident light strenght. ##

For many people, especially for those who use meters on cameras, 
incident light strength does mean exposure time.  However, for 
astrophotography, only incident light strength matters.  This means,
even with well controlled air pressure and humidity, to well handle RF, 
I have to take flats under the same dark sky, not just with the same
exposure time.  This is almost impossible in practice.  

So, in short, I just give up, forget about RF, and pray that RF is
really a minor effect.

> Why not create a synthetic flat field from the original image then?

In most of my cases, the images are always full of stars and 
nebulas.  There are very few real background regions.  A flat fild
image just cannot be created from these very few regions.  The only 
way is to assume some flat field models and use these very few 
background regions to make the models closely represent the real sky 
background.

I have an old note about vignetting correction in my website:
http://www.ifa.hawaii.edu/~wang/gallery/random_notes/vig_correction.htm
Perhaps this can also help the discussion.

I hope I can post some images later this week.

Cheers,

Wei-Hao 


_______________________________________________
Astro-Photo mailing list
Astro-Photo@seds.org
http://seds.org/mailman/listinfo/astro-photo