Re: [APML] Histograms

["Eddie Guscott" <eddie.guscott@blueyonder.co.uk>]

Hi Matt

You may have noticed me lurking behing Alan's emails and following this
thread closely. A lot of
whats been posted in the last 24 hours regarding Histograms has been very
very helpful. I've tried
rescanning at 16 bit and doing some levels adjustments at the scanning stage
and my histograms
are no longer picket fence posts.

Thanks to your reference on curves and clipping of histograms, I visited
your web site article (this
helps a lot).

I downloaded your raw scan of North America Nebula and tried following your
steps (I know the
image is a JPEG and a lot of info is lost etc...) but I could follow your
steps up to curve adjustments.

How can you take 30 minutes to do curve adjustments - I seemed to do it an
about a minute!!!!!
This is obviously one of the areas where I (and I suspect Alan Voetsch) am
falling down. How can
it take 30 minutes - what exactly are you doing with curves that takes so
long. I know it must
be important or else you wouldn't spend so long doing it, and as Tony Hallas
said "Curves, curves, curves".

Would appreciate a little light shed on this if possible.

Thanks

Eddie Guscott


----- Original Message -----
From: "Matt BenDaniel" <matt@starmatt.com>
To: <astro-photo@seds.org>; <astro-photo@seds.org>
Sent: Monday, April 29, 2002 8:30 PM
Subject: Re: [APML] Histograms


> At 03:28 PM 4/29/2002, Alan Voetsch wrote:
> > >as
> > > > long as you know how to configure the histogram
> > > correctly before
> > > > switching into 8-bit mode and as long as you use
> > > layers.
> >
> >Ok. Now can anyone answer me this: How does one learn
> >to recognize when a histogram is properly configured?
> >What are we looking for? How do we learn this?
>
> Alan,
>
> The shape of the histogram depends on what's in the FOV and the how it is
> processed. There is discretion in the processing. One thing that is easy
to
> tell is whether the histogram is clipped. You generally do not want a
> clipped histogram in the early processing steps, because clipping throws
> away data that may be useful. To see what clipping looks like see:
> http://world.std.com/~mattb/articles/curves.htm
> --
> Matt BenDaniel
> matt@starmatt.com
> http://starmatt.com
>
>
> --  APML Archives at <http://astro.umsystem.edu/apml/>  ---
>              Unsubscribe at <majordomo@seds.org>


--  APML Archives at <http://astro.umsystem.edu/apml/>  ---
             Unsubscribe at <majordomo@seds.org>