Understand STF, Histograms, and Clipping

A deep-sky integration often looks almost black because most pixel values occupy a narrow dark range. ScreenTransferFunction, or STF, remaps those values for display so you can inspect the image without permanently stretching it.

MAS, GHS, and HistogramTransformation can change the actual pixel values. After a permanent nonlinear stretch, the image remains visibly stretched with STF reset. Preserve a linear checkpoint before making that change.

A bright display alone does not establish whether an image is linear. Check processing history and the source checkpoint. Conversely, a dark image is not automatically a healthy linear integration: inspect its values and acquisition history too.

Open a known linear integration and make a practice duplicate. Find ScreenTransferFunction through All Processes. Keep the original available and avoid applying a pixel-changing process during this first exercise.

Open HistogramTransformation and select the intended view without applying anything. The horizontal axis represents pixel intensity, from dark on the left to bright on the right. The vertical axis counts pixels in each intensity bin. A taller peak means more pixels in that range, not a brighter astronomical object.

For a typical linear integration, most pixels may cluster close to the left edge. Zoom the histogram’s dark range to inspect its shape rather than assuming the narrow peak is clipped. Changing graph zoom, bin display, or logarithmic count scaling does not stretch the target pixels.

Inspect individual RGB channels as well as the combined display. One channel can clip or behave differently even when the overall histogram appears acceptable. Check which view and channel are selected before interpreting the graph.

A histogram shows the distribution of values, not their location. It cannot tell you by itself whether an endpoint spike comes from invalid borders, saturated stars, or the object’s core. Use pixel readouts and image inspection alongside it.

Clipping maps a range of distinct input values to an endpoint so their differences are lost in that result. Raising a permanent black point can erase faint dust and sky variations; lowering a permanent white point can erase distinctions in bright cores and star color.

In the usual normalized display range, endpoints are 0 and 1. Floating-point processing can also contain values outside that range, so view the statistics and format context rather than treating every unusual value as proof of damage.

An endpoint pile-up is a reason to investigate, not automatic proof that your last stretch clipped the subject. Registration borders can contain zeros, and bright-star pixels may already be saturated in the raw exposures. Compare the same area in the earlier checkpoint.

Display clipping in STF affects what you can see without deleting the underlying values. Pixel clipping in a saved permanent transformation or limited-range export is different. Resetting STF cannot undo it; return to an earlier checkpoint when available.

A bright core with retained differences can be compressed or stretched more carefully. An originally saturated core needs other captured data, such as shorter exposures, to supply missing information. HDRMT cannot reconstruct clipped values.

Use a disposable duplicate of the known linear checkpoint. This exercise demonstrates the difference; it is not a requirement to use HistogramTransformation before MAS or GHS.

PixInsight can transfer an STF instance into HistogramTransformation to make that display mapping a pixel-changing transformation. It is useful to understand this common workflow, but Auto Stretch was designed for inspection and may be too aggressive for a finished image.

On a fresh duplicate, calculate the desired STF, reset HistogramTransformation, then drag STF’s New Instance triangle onto HistogramTransformation’s lower control bar, not onto the target image. Inspect the transferred channel settings and clipping before applying HistogramTransformation once.

Reset STF afterward so you do not view the permanent stretch through the old display stretch as well. Inspect the resulting pixels, histogram, faint structure, and star cores. Retain the linear original regardless of how pleasing the transferred preview looked.

For the recommended learning path, continue to MAS or GHS on a separate linear copy instead of treating STF transfer as a mandatory extra stage. The tools provide different ways to build a permanent stretch.

The exported TIFF or JPEG is almost black: confirm whether only STF was making the source visible. Reopen the export with STF reset and compare its actual state. Complete a permanent stretch on an appropriate source copy before display export.

The image suddenly looks washed out: reset STF and close or reset any process preview. Check whether you are seeing a double display treatment or a proposed repeated permanent transformation.

Resetting STF seems to do nothing: the target may already be permanently stretched. Check history and pixel data rather than assuming the reset failed.

The sky looks cleaner after a blackpoint change: compare faint structures and endpoint counts with the baseline. Darker is not automatically better, and an empty-looking part of the histogram is not permission to discard all values there.

Color shifts between views: compare linked/unlinked STF, selected channels, and permanent processing separately. A display balancing choice is not evidence that SPCC succeeded or failed.

Keep a known linear checkpoint and label later stretched versions clearly. Continue to MAS or GHS for the main stretch, then inspect every finishing step with STF off unless there is a deliberate reason to use a display transform.

This is an original beginner exercise. The linked material explains display transforms, permanent stretching, and image analysis; older screenshots are not current defaults. Installed process help governs current controls.