[Author Prev][Author Next][Thread Prev][Thread Next][Author Index][Thread Index]

Re: Re: [APML] OT: CCD sensitivity [was: first CCD images]



Radu,

OK now we're are really asking for it <g>. The reason I tend to lump photon shot noise as a bias gradient is because I think the noise granularity is too fine and the DC bias component too large. But I will stop now before I am way out of my league.

Loke



> 
> From: Radu Corlan <rcorlan@pcnet.ro>
> Date: 2002/03/26 Tue AM 02:19:14 EST
> To: astro-photo@seds.org
> Subject: Re: [APML] OT: CCD sensitivity [was: first CCD images]
> 
> Matt, 
> 
> In fact, light pollution _is_ noise; The fact that light is discrete in
> nature introduces the so-called 'photon shot noise'; If you had a perfect
> sensor that would give a count of zero when 'exposed' to darkness and 
> exposed it to a stationary flat light source that gives a count
> corresponding to N electrons (absorbed photons), you will not get a count
> of N each time, but rather a count which has a rms error of sqrt(N). This
> applies both to light-generated electrons, and to thermally generated
> electrons (dark current). Even if you substract the 'N', you will still be
> left with the sqrt(N) noise.
> 
> Radu
> 
> 
> On Mon, 25 Mar 2002, Matt BenDaniel wrote:
> 
> > Loke,
> > 
> > One of the most widely misunderstood concepts in astrophotography is noise. 
> > Light pollution is NOT noise in an engineering sense. Noise is a component 
> > that introduces a statistical variation to the signal. Light pollution is 
> > unwanted bias. It can be eliminated in a CCD image by a flat field. A flat 
> > field cannot eliminate noise. Gradients asides, light pollution reduces the 
> > dynamic range (S/N ratio) of an image. A light pollution gradient can be 
> > subtracted if you model it correctly (and we can do it intuitively Photoshop).
> > 
> > The way to compensate for the loss of S/N due to light pollution is to 
> > collect more photon flux to increase the S/N. CCD's higher QE enables the 
> > capture of more photon flux per unit time, regaining more S/N lost by light 
> > pollution.
> > 
> > You can fight light pollution with film by taking lots of images and 
> > stacking. But it takes longer and is more work.
> > 
> > As an aside, a dark current is not noise either, because it can be 
> > subtracted. The source of noise for CCD is primarily readout noise. For 
> > film, the noise is mainly the grain.
> > 
> > Biases include: light pollution, dark current, film fogging, and of course 
> > CCD bias.
> > 
> > Matt
> > 
> > At 02:46 PM 3/25/2002, lkt wrote:
> > 
> > >         The QE part would only make a difference in exposure lengths, that is
> > >the higher the QE the shorter the exposures for the same amount of
> > >exposure depth. The main question is, what makes CCDs seem almost immune
> > >to light pollution vs. film. Of course, if you campare only CCD images
> > >between themselves, one with light pollution and one without, you can
> > >see the linear effect of degraded SNR. There must be something about
> > >film which makes is much more susceptible to light pollution. Maybe its
> > >the toe of the response or the logarithmic sensitivity (your 2nd point).
> > >Your 3rd point about lower noise levels in CCDs is also not an issue. I
> > >am talking about a case where the dominant noise component is light
> > >pollution such as the moon etc and not any noise due to film grain or
> > >CCD electronics.
> > >
> > >Loke
> > >
> > >
> > >
> > >Matt BenDaniel wrote:
> > > >
> > > > Loke,
> > > >
> > > > I'm newer at CCD than you are, but I have been reading a lot. Here's my
> > > > understanding:
> > > >
> > > > CCD's have:
> > > > * much higher quantum efficiency
> > > > * linear response and essentially no toe
> > > > * lower noise levels (basically readout noise) yielding higher dynamic 
> > > range
> > > > * a more automatable exposure process, making stacking easier
> > > >
> > > > The noise that we fight in film is mostly grain size and uneven 
> > > distribution.
> > > >
> > > > Matt
> > > >
> > > > At 01:36 PM 3/25/2002, lkt wrote:
> > > >
> > > > >Yes, CCDs are quite impressive at recording light aren't they? I have
> > > > >often wondered what the explanation of this effect is. That is, can this
> > > > >be due to the linearity of CCDs in recording light versus film's
> > > > >logarithmic response? Or possibly that film records light at a non
> > > > >uniform monotonic rate as a function of image intensity.
> > > > >
> > > > >Loke
> > > > >
> > > > >
> > > > >
> > > > > > Tony Hallas wrote:
> > > > > >
> > > > > > Hi,
> > > > > >
> > > > > >     Being a real "film guy" I thought it would be interesting for you
> > > > > > to hear my impressions of the CCD camera I have on loan from SBIG...
> > > > > >
> > > > > >      Picture this first:   Cold front has just cleared out...
> > > > > > everything is dripping wet... 3/4 moon is lighting up the haze in the
> > > > > > air (aka as major dew event) to the extent that all I can see is the
> > > > > > Big Dipper and a few other bright stars... in the finder scope I can
> > > > > > just see the triangle that defines where M-51 is... no sign of the
> > > > > > galaxy.
> > > > > >
> > > > > >      The ST-10 takes a one minute exposure... one minute!
> > > > > > As in 60 seconds... to my utter surprise most of the galaxy is
> > > > > > there... including some of the faint tendrils coming off the
> > > > > > companion... it is very noisy from low S/N.
> > > > > >
> > > > > >      I took 3 - 10 minute shots with dark frames but no bias and no
> > > > > > flat field just to see what it would record out of the haze...
> > > > > > although there was information in the core when it was a native file,
> > > > > > when I brought it over to PS as a Tiff file it got lost so the cores
> > > > > > are all burned out.
> > > > > >
> > > > > >      The first image is the 60 second image, the second is the 3 - 10
> > > > > > minute shots run through PW and a little clean up because I didn't
> > > > > > take a flat field. These are my first CCD images in many years so
> > > > > > don't be too critical!  : - )  What is amazing is the conditions under
> > > > > > which they were taken... bright, hazy, moonlit night.
> > > > > >
> > > > > >         http://www.astrophoto.com/CCD.htm
> > > > > >
> > > > > >        For what it's worth,
> > > > > >
> > > > > >              Tony
> > > > > >
> > > > > >
> > > > > >
> > > > > >
> > > > >
> > > > >--  APML Archives at <http://astro.umsystem.edu/apml/>  ---
> > > > >              Unsubscribe at <majordomo@seds.org>
> > > >
> > > > --
> > > > 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>
> > 
> > --
> > Matt BenDaniel
> > matt@starmatt.com
> > http://starmatt.com
> > 
> > 
> > --  APML Archives at <http://astro.umsystem.edu/apml/>  ---
> >              Unsubscribe at <majordomo@seds.org>
> > 
> > 
> 
> -------------
> Radu Corlan           Snail Mail: Bucuresti sect. 1, 
> rcorlan@pcnet.ro             str. Argentina nr. 28, 71206 Romania
> 
>    You can still escape the "Gates" of Hell!   
>                  Use Linux!                       
> -------------
> 
> 
> --  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>