PROCESSING GUIDE
Standard Pixinsight Deep-Sky Processing Workflow
From calibrated subframes to a finished image
This is the processing workflow I currently use for many one-shot-color deep-sky images in PixInsight. It isn't intended as a fixed recipe or a list of universal settings. Different targets and datasets require different decisions. The goal is to provide a repeatable backbone, explain what each stage is intended to accomplish, and identify the points where I normally stop, inspect the image, and decide what comes next.
Workflow at a Glance
The Workflow
The sequence below is the backbone I currently use for many one-shot-color deep-sky images. The individual settings change from image to image, but the overall order is intentional. I generally complete calibration, color calibration, gradient correction, blur correction, and initial noise reduction while the image is still linear. I then separate the stars and process the starless image and stars independently before recombining them for final adjustments. Each stage includes points where I stop, inspect the image, and decide whether the next step is appropriate for that particular dataset.
1. Calibrate and Integrate — WBPP
I begin in PixInsight with WeightedBatchPreprocessing (WBPP), using it to calibrate, register, evaluate, and integrate the individual subframes into a master image.
​
This is more than a housekeeping step. Problems introduced here can follow the image through the entire processing workflow, so I inspect the resulting master before moving forward. I look for registration problems, satellite or aircraft trails that survived rejection, unusual edge artifacts, gradients, clipping, and anything else that suggests the integration should be revisited.
​
I don't use one universal set of WBPP parameters for every dataset. The number of imaging sessions, calibration data available, image quality, and characteristics of the target can all affect the choices I make.
What I'm looking for:Â Â A clean, properly registered master with good rejection and no obvious integration problems before beginning image processing.
2. Establish Astrometry — ImageSolver
Before performing color calibration, I use ImageSolver to establish an accurate astrometric solution for the integrated image. This identifies exactly where the image is located in the sky and provides the coordinate and scale information needed by processes such as SpectrophotometricColorCalibration (SPCC).
​
If the image has already been successfully plate solved during acquisition or integration and contains a valid astrometric solution, this step may not need to be repeated. I verify that the solution is present and correct before moving on.
What I'm looking for: A successful astrometric solution with the correct image coordinates, orientation, and scale.
3. Establish Initial Color — SPCC
With a valid astrometric solution in place, I use SpectrophotometricColorCalibration (SPCC) to establish a physically based starting point for color. SPCC uses the stars in the image, along with catalog data and the characteristics of the imaging system, to calibrate the relationship between the red, green, and blue channels.
​
I treat SPCC as a starting point rather than the final artistic color treatment. At this stage, my goal is to establish a reliable color balance while the image is still linear. More subjective color adjustments can come much later in the workflow.
What I'm looking for: A believable, neutral starting color balance with natural-looking star colors and no obvious color cast introduced by the calibration.
4. Correct Gradients — GraXpert
After color calibration, I evaluate the image for gradients and uneven background illumination. I typically use GraXpert for gradient correction, particularly with data captured from my light-polluted home skies.
​
I don't treat gradient removal as an automatic cleanup step. I inspect the correction carefully to make sure real astronomical signal—especially faint nebulosity—isn't being mistaken for background. The appropriate correction strength depends heavily on the target and the character of the gradient, so I generally favor the least aggressive correction that produces a clean, natural-looking background.
​
This is also a point where I may decide that the image doesn't need additional gradient correction at all. The goal is to remove unwanted background structure, not to make the background artificially uniform.
What I'm looking for: A more even background without removing faint target signal, creating artifacts, or flattening legitimate large-scale structure.
5. Correct Optical Blur — BlurXTerminator
While the image is still linear, I use BlurXTerminator (BX) to correct optical blur and improve the definition of both stars and deep-sky structure. Applied at this stage, BX can recover detail affected by atmospheric seeing, tracking, and the optical system before the image is stretched.
​
I approach BlurXTerminator primarily as a correction tool rather than simply as a way to make the image look sharper. I inspect stars closely for changes in size and shape and examine fine structures within the target to make sure the correction remains natural.
​
The appropriate settings vary with the quality and image scale of the data. I would rather leave a little softness in the image than push the correction far enough to create hard edges, unnatural small-scale structure, or artifacts around stars.
What I'm looking for: Better-defined stars and improved structural detail without halos, ringing, distorted stars, or an artificially sharpened appearance.
6. Reduce Noise — NoiseXTerminator
Before stretching, I use NoiseXTerminator (NX) for an initial reduction of noise while the image is still linear. At this stage, the goal is not to eliminate every trace of noise, but to give the image a cleaner foundation before stretching makes the remaining noise more visible.
​
I generally favor a controlled application rather than aggressive smoothing. I compare the preview carefully, paying particular attention to faint nebulosity and small-scale detail that can be softened along with the noise.
​
Additional noise reduction may be appropriate later in processing, particularly after stretching, but I prefer to preserve real structure rather than chase a perfectly smooth background.
What I'm looking for: Noticeably cleaner data with faint structure and fine detail preserved, rather than a completely smooth or artificial-looking image.
7. Separate Stars — StarXTerminator
Before stretching, I use StarXTerminator (StarX) to separate the linear image into a starless image and a stars-only image. From this point forward, I can process the nebula, galaxy, or other deep-sky structure independently from the stars.
​
This separation gives me much greater control during stretching and color adjustment. The starless image can be stretched aggressively enough to reveal faint structure without simultaneously enlarging or overwhelming the stars, while the stars can receive a more restrained treatment of their own.
​
I inspect the starless image carefully after separation, particularly around bright stars and dense star fields, for residual stars, holes, halos, or other artifacts that may need attention before continuing.
What I'm looking for: A clean separation with the target structure preserved, minimal star remnants or artifacts, and a stars-only image suitable for independent processing.
8. Process the Starless Image
With the stars removed, I begin the main nonlinear processing of the deep-sky target. I typically use HistogramTransformation (HT) and/or Generalized Hyperbolic Stretch (GHS) to move the image from linear to nonlinear while controlling the balance between faint structure, brighter regions, and the background.
​
From this point, the workflow becomes much less prescriptive. Depending on the target, I may use CurvesTransformation, GHS, masks, contrast adjustments, and selective color work to develop structure and color while protecting highlights and the background.
​
I usually work in several restrained passes rather than trying to achieve the final appearance with one aggressive adjustment. I stop frequently to evaluate the image at different scales and watch for clipped blacks, blown highlights, excessive saturation, harsh contrast, or processing artifacts.
What I'm looking for: Faint structure becoming visible while brighter regions remain controlled, with natural transitions, a believable background, and enough processing headroom for further refinement.
9. Process the Stars
I process the stars separately from the starless image, generally using a more restrained stretch than I use for the deep-sky target. HistogramTransformation (HT) is usually sufficient to bring the stars into the nonlinear state while maintaining control over their size, brightness, and color.
​
My goal is usually to preserve natural-looking star color without allowing the stars to dominate the finished image. Depending on the dataset, I may make modest adjustments to saturation, brightness, or contrast before recombination, but I try to keep this processing minimal.
​
Processing the stars independently also gives me the flexibility to balance their prominence against the target rather than accepting whatever star intensity results from stretching the entire image together.
What I'm looking for: Natural star color, controlled brightness and size, and enough presence to complement the target without overwhelming it.
10. Recombine Stars and Starless Image
Once the starless image and stars have been processed independently, I recombine them to create the finished composite. I typically use ImageBlend for this step, which allows the separately processed components to be brought back together while retaining the advantages of the starless workflow.
​
Recombination is another decision point rather than simply a mechanical step. I evaluate whether the stars feel appropriately balanced with the target and may adjust their intensity before settling on the final blend. Stars that looked appropriate in isolation can sometimes appear too prominent once they are returned to the processed nebula or galaxy.
​
I also inspect the recombined image closely for halos, unusual transitions around bright stars, color shifts, or artifacts that may have been introduced during star separation or independent processing.
What I'm looking for: A seamless recombination in which the stars look like a natural part of the image, complement the deep-sky target, and show no obvious signs of having been processed separately.
11. Final Finishing
After recombination, I evaluate the image as a whole and make any final adjustments needed to bring the processing together. These may include modest changes to color, saturation, contrast, brightness, or the background using tools such as CurvesTransformation, GHS, or carefully constructed masks.
​
I try to keep this stage restrained. By this point, most of the important processing decisions have already been made, so final finishing should refine the image rather than fundamentally change it. I pay particular attention to the relationship between the target, stars, and background and to whether the image still looks natural at both full-frame and closer viewing scales.
​
I may also make the final crop or framing decision at this stage, particularly if registration or stacking has left edge artifacts or if a tighter composition improves the presentation of the target.
What I'm looking for: A balanced final image with controlled color and contrast, a clean background, natural stars, and a composition that directs attention toward the target without looking overprocessed.
12. Final Quality Check
Before considering an image finished, I step away from individual processes and evaluate the result as a complete image. I inspect it both at full-frame scale and at higher magnification, looking for problems that may have been easy to miss while concentrating on individual processing steps.
​
I check the background for clipping, gradients, excessive smoothing, or color casts; examine bright and faint stars for halos, distortion, or unnatural color; and look through the target for oversharpening, excessive saturation, blown highlights, or processing artifacts. I also reconsider the crop and overall composition.
​
Most importantly, I compare the final image with the data I started with. The objective isn't to extract every possible bit of contrast or color, but to produce an image that reveals the available structure while still looking balanced and believable. If something draws attention because of the processing rather than the astronomical subject, I usually go back and reconsider that adjustment.
What I'm looking for: A finished image that holds up at both normal viewing size and closer inspection, with no obvious technical distractions and a balance of detail, color, stars, and background that serves the astronomical target.