As explained in earlier posts, two prominent types of noise in astronomical CCD images are read noise and shot noise. Unlike read noise which is constant for any given image, shot noise varies directly with number of photons recorded by the CCD. Since the number of photons recorded is a function of the length of an exposure, the shot noise will increase as the exposure time is increased.
Generally, sky limit is the the zone (or range of exposure times) where the sky (shot) noise overpowers or buries the read noise. Typically, the shot noise is considered to overwhelm the read noise when the read noise contributes no more than 5% of the total noise contributed by these two sources.
Example: For my SBIG ST-8 camera, I have measured the read noise to be around 20e- and the gain for this camera is near 2.6 e- per ADU. Suppose I take a five minute exposure and record a average reading of 1000 ADUs from the darkest part of the resulting image (i.e. the sky background of the image). For this particular image, the shot noise would be equal to sqrt((adu – pedistal)* gain) or sqrt((1000-100)*2.6) or sqrt(2340) or 48e-. Since the total noise is calculated by adding these two terms quadratically, total noise = sqrt(sn*sn + rn*rn) which is equal to sqrt(48*48 + 20*20) = 52e-. In this case, the total noise of the read noise and the shot noise is about 52 electrons. If we consider only the shot noise, the total noise is 48 electrons. Since 52/48 equals 1.083 then in this case the read noise contributes about 8.3% to the total noise.
In general, the shot noise should be about 3x the read noise so that the read noise contributes no more that 5% to the total noise. Since my camera exhibits a read noise of 20e- I need to calculate a sky background count that gives me a sky background shot noise of about 60e-.
Calculating this general case: background target adu = ((sn*sn)/gain + pedistal) which in this case is target adu = ((60*60)/2.6 + 100) = 1485 ADU.
Conclusion: In order to minimize the effects of read noise, I should try to make my exposure length long enough such that the background sky of the image attains an average value of around 1500 ADUs.
But how practical is this method to minimize read noise? Well, for an observing site that suffers from light pollution it is very easy to reach this level of sky background count. At my home site, my images are sky limited at about 5 minutes through my 4? refractor and slightly longer through my SCT. However, at my dark site in Chiefland, I can not practically sky limit my exposures due to the extremely dark skies (since it would take an exposure of several hours to sky limit the image). For these images, the read noise is an unavoidably significant contributor to the total noise.
In the process of recently moving all my amateur radio QSOs to DXKeeper database, I’ve come up with a breakdown of all 10,709 QSOs by band and mode.
By Band:
By Mode:
After an initial tryout using Blogger, I’ve decided to switch to Wordpress for my blog software. Previously, with a blog on my old website, I also used Wordpress for blogging software but this time I wanted to give Blogger a try first. It turns out that Wordpress requires a bit more initial setup on the website’s server (in terms of setting up a SQL database), but I find Wordpress to be easier to customize and adapt for my particular use. Also, Wordpress seems to have a larger selection of ready-to-use themes and, like Blogger, has great documentation and a large, active, and helpful user community.
In conclusion, for the beginning blog, Blogger is a great way to start but for the more experienced user who requires more flexibility and adaptability, Wordpress is the way to go.
Here is my current DX Century Club award status:
Mixed (135 Confirmed) : 106 Awarded, 29 Pending (15 LotW, 14 Card)
CW (100 Confirmed) : 79 Awarded, 21 Pending (11 LotW, 10 Card)
Phone (118 Confirmed) : 79 Awarded, 39 Pending (24 LotW, 15 Card)
RTTY (66 Confirmed) : 44 Awarded, 22 Pending (12 LotW, 10 Card)
80m (18 Confirmed) : 13 Awarded, 5 Pending (2 LotW, 3 Card)
40m (69 Confirmed) : 51 Awarded, 18 Pending (11 LotW, 7 Card)
20m (102 Confirmed) : 76 Awarded, 26 Pending (15 LotW, 11 Card)
15m (111 Confirmed) : 81 Awarded, 30 Pending (17 LotW, 13 Card)
10m (93 Confirmed) : 70 Awarded, 23 Pending (13 LotW, 10 Card)
A second source of noise in astronomical CCD images is the noise or uncertainty generated from the object itself and is commonly referred to as shot noise. This uncertainty arises because the detection of photons by a CCD is a statistical process. If the the same object is imaged for the same length of time under the same conditions, the number of photons detected will vary slightly from image to image. If enough images are taken, the pixel values will be seen to follow a Poisson distribution. The shot noise is defined as the uncertainty (plus or minus) around the mean value of the pixel values and is usually defined in terms of one standard deviation from the mean. Since a Poisson distribution’s standard deviation is equal to the square root of the intensity (SQRT(i)) this is also the definition of shot noise. For example, if a CCD records an intensity of 100 adu from a certain object, the shot noise is equal to +/- 10 adu. Likewise, if a CCD records an intensity of 1000 adu from a certain object, then the shot noise is about +/- 31 adu.
In conclusion, shot noise, unlike readout noise, is dependent on the intensity (photon flux) of the object and is therefore, dependent on the length of the exposure. Brighter objects, therefore, have better SNR (10 in the first example and 31 in the second example.)