CaLIGHTs v3.1.1.0 Released
wols...@gmail.com
Latest Revision 3.1.1.00
-I updated CaLIGHTs to use version 0.20.2-Release of LibRAW which is the latest version. The LibRAW folks state that 1133 different DSLR cameras are now supported.
-New features added to Calibrate Preview screen. Left mouse button click &drag of the image is now possible. The mouse wheel can be used to zoom the image. If you position your mouse on a region of interest and then click the right mouse button a small green square will appear along with some statistics. The box is a 60 x 60 pixel square that is centred on your mouse position. The statistics are calculated for the red, green and blue pixels in the superpixel debayered image. Means, Standard Deviations and signal-to-Noise ratios are calculated. Here is an example…
N=900 Pedestal=600
Xmin=2224 Xmax=2284 Ymin=1632 Ymax=1692
Red Mean=151.2 SD=32.2 SNR=4.7
Green Mean=159.2 SD=30.4 SNR=5.2
Blue Mean=76.2 SD=28.7 SNR=2.7
I need to explain these values.
- The best way to use these statistics is for you to select a region of interest that is devoid of stars. These regions are where the noise in your LIGHT frames typically limits how much you can stretch your astrophotos when you are post processing. The goal is to make changes to how CaLIGHTs will calibrate your LIGHT frame and use these statistics to quantify any improvements. Increasing the SNR in an individual LIGHT frame has a cumulative improvement on your final astrophoto as you perform stacking and digital development.
-N=900 This is the number of data values used to generate the stats for each colour. The 60 x 60 pixel square of RAW values has been superpixel debayered to 30 x 30 RGB pixels. 30 x 30 = 900
-Pedestal=600 This is the pixel value that CaLIGHTs believes corresponds to total darkness. Depending on what settings you are using, this Pedestal value will equal either the Black ADU value extracted from your RAW DSLR file or FITS file, your custom Black ADU value or the mean value of the pixels in your master BIAS for LIGHT.
-Xmin=2224 Xmax=2284 Ymin=1632 Ymax=1692 These are the X and Y values that define the 60 x 60 pixel square. These values are scaled so that they are the correct values for your calibrated LIGHT frame.
-Red Mean=151.2 SD=32.2 SNR=4.7 These are the statistics calculated for the Red pixels. Mean is the mean value of the 900 red values. Note that the Pedestal value has already been subtracted. SD is the standard deviation of these 900 red values. Standard deviation is a good indicator of noise. SNR is the Signal-to-Noise ratio which equals Mean / SD. These values are similarly defined for the Blue pixels.
The statistics for the green pixels are calculated slightly differently. The superpixel debayer algorithm averaged the green pixels because there are twice as many green pixels as there are red or blue. This averaging causes the SD value for the green pixels to be smaller than the red or blue pixels by a factor of 1.414. This would make you think that the green pixels have much less noise which I believe is misleading. I decided, for the green pixels only, to increase the SD value by 1.414 so that it is straight forward to compare the SD and SNR values for all three colours.
-Fixed a bug in how CaLIGHTs writes calibrated fits files for dedicated astrocams. Astrocams typically generate LIGHT frames that are fits files and the fits headers in these files can become quite long. CaLIGHTs was not creating valid calibrated fits files when the fits headers were excessively long.
-I have also begun implementing multi-processor support in CaLIGHTs. The noise filter in version 3.1.0.92 had already been modified to support multi-processor CPUs. With my 2 core computer this significantly reduced the time required to perform noise filtering. This multi-processor support has been expanded to many aspects of how CaLIGHTs calibrates your LIGHT frames. The row noise filter showed the biggest reduction in execution time. As a simple comparison, I calibrated 28 LIGHT frames using master DARK, FLAT, DARKFLAT and BIAS frames. I also included the Noise Filter, the Bad Pixel Filter and the Row Noise Filter. No binning was performed so that the calibrated fits files were the largest. Version 3.1.0.92 took 172 seconds. Version 3.1.1.00 took 133 seconds. If binning is used these times actually go DOWN! This happens because the Noise Filter and the disk writing occur after binning.
-It worth noting that CaLIGHTs has been written to be frugal with RAM memory so that it can handle the biggest LIGHT frames without requiring the user to add lots of RAM. To perform the above mentioned calibrations required roughly 125Mb of RAM. Each LIGHT frame was 16 bit 11MPixel.
