Re: [ccd-newastro] Re: Cooling question
- There is no way to remove noise.
You can remove instrument artifacts (e.g., dark current), but you can never remove noise.
And a dark frame especially does not remove read noise, which is the uncertainty in the readout. <g>
A dark frame removes dark current. The colder the chip, the lower the noise in the removal process, mostly because there is less dark current to remove.
On Apr 3, 2010, at 9:26 AM, Mike Dodd wrote:
> ancient.sull wrote:
>> I think the answer is that the shot noise of the 120 ADU dark frame
>> is SQRT 120 = 10 and the shot noise at 450 ADU is 21 so after
>> subtraction using the +10C dark frame you will have a little more
> Um.... I could be wrong, but I believe that shot noise is the
> uncertainty of the incoming SIGNAL, and the dark frame does nothing to
> remove the noise.
> As I understand it, the dark frame removes only the read noise. Since
> the shot noise comes in with the signal, there is no way to remove it
> with a dark frame.
> Mike Dodd
> Montpelier, VA USA
I've been predicting this for a few years now. Once Sony has exited there will be few affordable CCDs (Kodak / Truesense). Most surviving CCD will be expensive small runs (E2V). There will be significantly different dynamics when the amateur astro-imaging manufacturers and community come to terms with CMOS.
Astro-cam manufacturers may face a threat/challenge from the big camera companies (Canon, Nikon, Sony, etc.). But they might be safe due to the miniscule astro-cam market combined with the inertia and indifference of the big companies. Canon and Nikon have produced half-hearted "astro" models that simply omit the IR filter with a significant price increase. But these are not likely to go anywhere - the Canon is dead and the Nikon may die soon. But the real saving grace for astro-cams will be mono (non-Bayer) cams, which remain infuriatingly unavailable for DSLR/MICL (except for one very expensive Leica).
Astro practitioners will face new sensor dynamics and adopt new imaging practices. CMOS has much less read noise and shallower full well capacity than CCD. Both of these characteristics favor significantly shorter sub-exps than CCD practitioners are accustomed to. This will permit unguided imaging as viable sub-exps of only a few seconds become feasible. This of course also means many more sub-exps, which in turn affects processing in many ways. I have been exploring these dynamics for some time now (via intensified imaging) and they are real game changers. It could be an exciting new horizon...