Images Plus vs PixInsight ... image integration

[menardre]

Images Plus vs PixInsight
This evaluation covers creating a master image from lights, darks, flats, and biases. ‘Lights’ are the images you get directly from the camera. ‘Darks’ and ‘Biases’ are images taken with the lens cap on to help remove hot pixels. ‘Flats’ are images taken with even lighting to help remove ‘cold’ pixels or smudges on the lens.
Images Plus
• The process to combine lights, flats, darks, and biases is very straight forward. Select the ‘Image Set Operations’ tab.
• Then select ‘Automatic DSLR, OSC CCD image set processing. This results in a box with tabs for light frames, dark frames, flat frames, flat dark frames, and bias frames. You also select how the files will be debayer if they are one shot color.
• Selecting each frame tab you can then select your lights, flats, darks, biases, etc. If you have master darks or biases, you can select them. If you have the individual files, selecting all of the files will generate a master dark (or flat or bias).
• You then select the ‘Light frame processing parameters’ tab. Most of the defaults are very good .. I generally just change the combined file name … such as ‘M42 Master’.
• Then back on the automatic tab I check the ‘Minimize Hard Drive Use’, ‘Normalize’, ‘Grade’, ‘Align’ and ‘Combine’ boxes.
• This takes all of the lights, darks, flats, biases and processes them into a single aligned master file. It also deletes most of the un-needed supporting files.
• This process takes me about 45 minutes depending upon the number of light files.
• At this point you have a single master image in ‘FITS’ format.
• The next step is usually to stretch that image. There are several different types of stretch algorithms, but I generally use the ArcSinH stretch. I found this works really great for galaxies and nebula since it allows you to stretch the dim portions of the image (nebula) more than the rest of the image. This also works great for globular clusters since you can stretch the dense inner core less than the surrounding stars so you can bring out the star details without blowing out the central core.
• At this point you can use the rest of the vast capabilities of Images Plus to fine tune your image. I will cover this in the next review. As you can see, the stretching of the image is usually done prior to other functions such as background flattening, star compression, white balance, smoothing/sharpening, etc.
PixInsight
• The process to combine lights, flats, darks, and biases in Pixinsight is very different and more time consuming.
• PixInsight performs many of the color correction, noise reduction, etc prior to stretching. They call this linear processing (any processing done after stretching is non-linear processing).
• First you have to produce masters for darks, biases, and flats. Each uses a ‘ImageIntegration’ process which includes several selections. The darks and flats require to run an ImageCalibration’ process first (with it’s own settings), then an ‘ImageIntegration’ process.
• Next you have to calibrate the lights using ‘ImageCalibration’.
• At this point you can perform a ‘CosmeticCorrection’ process to look for hot or cold pixels. This step is not always necessary.
• It is now time to debayer the ‘lights’ to get the colors. This uses a ‘debayer’ process.
• Now you run a “SubframeSelector’ process to essentially grade each ‘light’ and assign weighting. (you can omit this process and hope that PixInsight’s automatic process does the job).
• Finally you are ready to align the lights using ‘StarAlignment’ process.
• Once the lights are aligned, you run the ‘LocalNormalization’ process.
• Finally we get to integrating all of the lights into a single ‘master’ light using the ‘ImageIntegration’ process. During all of these processes PixInsight utilizes its own file format which is much larger than FITS format (180meg vice 30meg).
• At this point you can also utilize the drizzle algorithm to remove pixilation. The ‘drizzle’ version master is really large. Where each of your FITS files might be 30meg, the drizzle may be about 800meg.
• It is at this point that you finally have a master non-stretched image. Expect to take several hours of mostly processing time to get to this point.
• Before you actually stretch (or in PixInsight terminology ‘transform’) the image you have to do the following:
• Dynamic crop
• Background Extraction
• Color calibration
• Noise reduction
• Many of these processes utilize specialized masks, Now you perform one of several ‘transform’ (stretch) processes. At this point you actually have done a lot of image processing and the image should be fairly good. Once again, at this point you can use the rest of the vast capabilities of PixInsight to fine tune the image.
• Along the way you have generated about 50 gig worth of files almost all of which can be deleted.
There is a 12 part video series which goes through each step in the process. This is called the very beginner’s series!
It should also be noted that PixInsight has ‘scripts’ to help automate some of the processes (I have not tried them yet since I want to understand each process).
These are my personal impressions of both software packages. I have used Images Plus for a few years but have only scratched the surface on what it can do. I have only had PixInsight for about a month and have just finished working completely through one image. I did read Warren Keller’s book on PixInsight, read several of the articles, and watched many of the on-line videos.

[Pete]

Thanks Rog. A good reminder as to why I dropped out of the LRGB party.