Processing Blog

How I remember what I did. Many of the prescriptions or suggested usages are taken from the Keller video series, but some are taken from the PixInsight Forum or PixInsight documentation.


Functions and scripts


Requires prior use of
ImageSolver, which may require ManualImageSolver and CatalogStarGenerator.


I'm not entirely sure what this is used for. It seems to be used to enhance or sharpen structures at medium to large scales. The bias slider should be set between .05 and 4. Noise reduction should be left unchecked. Focus on the upper layers to avoid sharpening noise. Deringing is applied to all layers by default. Increase to reduce ringing. Example usage: Set layer 3 at .1, layer 4 at .05. Can use more layers for large scales. Use luminance or range mask.

BatchFormatConversion script (in BatchProcessing)

Canon CR2 files must be converted to .fit or .tif before they can be used in PixInsight processes and scripts.

Blink (in ImageInspection)

Preview a set of images. Open the images with the small folder in the Blink dialog window.

CatalogStarGenerator script (CSG) (in Render)

I was able to generate a star map at the coordinates for M58, but the map did not match the image of M58. Use this script to assist AperturePhotometry and ManualImageSolver. MIS is used to assist ImageSolver.

DarkStructureEnhancement (DSE)

DarkStructureEnhancement worked well on the emission nebula IC1396 using the defaults. It was applied to non-linear data after TGVDenoise. The actual improvement is more dramatic than we see in these .jpg images. Keller says the script is useful for embellishing the dust lanes of galaxies or the darker dust of nebulae.

Example usage, a la Keller video

  • layers to remove: 8
  • scaling: 5x5
  • iterations: 1
  • amount: experiment
Can darken with CurvesTransformation.
jpg image

Figure 1. Before DarkStructureEnhancement

 jpg image

Figure 2. After DarkStructureEnhancement

ExtractWaveletLayers script (EWL)

Extracts wavelet layers and creates new file for each layer. Useful for making masks to operate on layers.

HDRCompression (HDRC)

Reduces regions of high contrast. Most useful for non-linear data, but can be used on linear. The result is non-linear. Use with a mask.

HighDynamicRangeMultiscaleTransform (HDRMT)

Juan Conejero discusses the use of HDRMT with star masking and histogram transformation on the PixInsight Forum. It is for enhancing contrast in bright objects, such as galaxies and is not suitable for nebulae lacking bright cores. Use star mask on bright stars. Recover lost contrast in curves and histogram transformations.

ImageSolver script (in ImageAnalysis)

Theoretically, one should be able to obtain identities of stars and other deep sky objects. It requires correct parameters for camera specifications, RA, and DEC. I have not had much luck using it. See also, Pixinsight documention. The T1i has pixels 4.69 micrometers square. The WO FL 110 APO has a focal length of 770 mm, but what is it with the Flat-4 focal reducer set on 73.5? A message on an astronomy forum offered the figure of 616 mm with aperture f5.6. Neither 770 mm nor 616 mm worked in the ImageSolver.

ManualImageSolver script (MIS) (in Render)

Use this script to assist ImageSolver. This script requires use of star map generated in CatalogStarGenerator.


(MMT) Use to enhance or sharpen small scale structures. Example: Select No Ringing. Use on layer 2 with .5 bias.

RangeSelection (RS) in MaskGeneration

Range selection was used to isolate the compression ridges of IC1399A, B, and E.

Low 0.44
High 0.50
Fuzziness 0.0
Smoothness 0.0
Lightness checked

 jpg image

Figure 1. Crop from RangeSelection mask

jpg image

Figure 1. Before RangeSelection mask and CurvesTransformation

 jpg image

Figure 2. After RangeSelection mask and CurvesTransformation

Statistics (in ImageInspection)

Analyse a single image. Inspects R, G, and B for count (%), count (px), mean, median, aveDev, MAD, minimum, maximum.

SubframeSelector script (in BatchProcessing)

Measures weight, Full Width Half Median (FWHM), eccentricity, signal to noise weight (SNRWeight). Outputs tables and plots. Since it detects stars, it is time consuming.

TGVDenoise (TGVD)

in NoiseReduction

TGVDenoise worked well on a non-linear stack of emission nebula IC1396 following HistogramTransformation and Curves.

300 iterations
Local Support
Midtones .67
Shadows 0
Highlights .76.

That is, the shadows were given no protection. The midtones and highlights were given high levels of protection. Some contrast is lost, so some intensification with CurvesTransformation or HistogramTransformation is applied after usage. Aug 8, 2014

jpg image

Figure 1. Before TGVDenoise

 jpg image

Figure 2. After TGVDenoise

Gary Palmer, Friday, August 15, 2014.