Hi John,
In the latest version of the ImageCalibration tool we have implemented a new feature that greatly improves dark frame optimization, especially when working with master dark frames of relatively poor quality. It is the
optimization threshold parameter in the Master Dark section of the tool. This parameter is also available on the BatchPreprocessing script since version 1.37 (on the Darks tabsheet). Unfortunately, we released these new versions just before Summer vacations, so we haven't had the time to document them yet. I'm going to explain how to use this feature briefly here, but we'll publish a more detailed description with examples when possible.
The idea behind the optimization threshold feature is conceptually simple. The master dark frame is always contaminated by readout noise to a smaller or larger extent, depending on the quality of the dark current data available. This random white noise degrades the performance of our dark frame optimization algorithm. A simple but very effective way to improve the behavior of this algorithm is to simply set to zero all master dark pixels below a specific threshold value. If the threshold is correctly determined, this simple method can remove a large portion of the readout noise, and the dark frame optimization algorithm works much better. The problem is hence how to find the correct threshold: too large of a threshold will remove too much true dark current data; too low of a threshold will have no effect. One way to find this critical value is as follows:
- Build master bias and master dark frames with the ImageIntegration tool. Follow the method described in
this tutorial.
- Open the master frames and subtract the master bias from the master dark frame with the PixelMath tool. The resulting calibrated master dark frame is just a temporary working image, so don't replace the master dark frame with it.
- Open the Binarize tool, select the calibrated master dark frame, and enable the Real Time Preview option. Now find an RGB/K binarizing threshold able to remove most of the darkest master dark pixels, leaving only the brightest ones. There's a critical point that sets to black about one half of the master dark. The correct threshold is above this value. Expect a value approximately in the range 0.00001 to 0.0001, but the actual threshold can vary greatly.
- You can make a few tests with the ImageCalibration tool, setting the dark optimization threshold parameter to the value found in the previous step. You should see a drastic improvement in your dark optimization factors in all cases, but especially if you work with a poor quality master dark frame.
- Once you've found a good threshold value, you can use it on the BatchPreprocessing script (
optimization threshold parameter on the Darks tabsheet). You can use the master bias and dark frames built in the first step. In such case, don't forget to check the
use master bias and
use master dark options on the BPP script.
This optimization threshold parameter is a first step towards a more sophisticated dark optimization method (multipoint optimization) that we plan on implementing during next Fall-Winter.
Also if I do image calibration in CCDStack (more convenient when having PA West and PA East Subs) and then move to Pixinsight for Image Alignment and Integration am I likely to lose anything
You can calibrate with meridian flips very easily with the BatchPreprocessing script. Just click the Add Custom button and use the Filter Name field to define a "pseudo-filter" to classify your frames as a function of their meridian side. The meridian position is only relevant to flat fielding.
Floating point images written by other applications (we call them
alien data) usually cause problems when loaded in PixInsight. This happens because the FITS format (also the TIFF format) provides no (standard) way to define the range to which the real-valued numerical data are referred. PixInsight always writes floating point images in the [0,1] range, where 0=black and 1=white. Unfortunately, other applications don't follow this rule and usually provide no information on the ranges used. Sometimes floating point images are stored in the 16-bit integer range [0.0,65535.0]. We have seen also calibrated data stored in 32-bit integer format using undocumented ranges that vary among images. When the black and white points are unknown, PixInsight has no way to interpret the alien data and the images lose their physical meaning when loaded.
