Getting More From Your Data with Free Software
The image your telescope hands you is a starting point, not a finished result. Two free tools do most of the remaining work, and neither requires you to become a retoucher.
Smart telescopes produce a finished-looking image in the app, and for many people that is enough. But the stacked file usually contains noticeably more than the app chooses to show, and recovering it does not require expensive software or a long apprenticeship.
Two free, open-source tools cover the great majority of what is worth doing, and the whole workflow can be learned in an evening.
What the app is doing, and what it holds back
Your instrument stacks frames and applies an automatic stretch, a transformation that maps the enormous dynamic range of the raw data into something a screen can display. That stretch is conservative by necessity: it has to look reasonable on every target without supervision.
Faint outer structure often survives in the data while sitting below the level the automatic stretch reveals. Gradients from light pollution are usually left in place. Colour balance is approximate.
None of this is a criticism. It is the sensible default for an unattended process, and it leaves room for improvement by anyone willing to spend twenty minutes.
Siril: the free stacking and processing suite
Siril is an astronomical image processing application, free and open-source, that handles calibration, stacking and the core processing steps.[1]
The three functions worth learning first:
Background extraction removes the gradient caused by light pollution and moonlight, the uneven brightening across the frame that makes an image look like it is sitting on a slope. This single operation improves more smart-telescope images than anything else on this page.
Colour calibration, particularly photometric colour calibration, uses star colours measured against a catalogue to set a defensible colour balance rather than an eyeballed one. It converts "that looks a bit green" into a solved problem.
Histogram stretching gives you manual control over the transformation the app applied automatically, so you can decide how much faint material to bring up and how much noise you are willing to accept alongside it.
GraXpert: gradient removal made simple
GraXpert does one job and does it well: removing background gradients, with a simpler interface than Siril's equivalent and an AI-assisted mode that requires almost no decisions.[2]
For a smart-telescope owner working from a light-polluted garden, this may be the highest-value fifteen minutes of learning available. The gradient is usually the single most obvious thing separating your image from a good one, and it is invisible until you remove it and compare.
It also includes denoising, which is worth using sparingly. Noise reduction always costs some real detail, and the temptation is to overdo it on under-integrated data. The better fix for noise is more integration time; denoising is for the last 10%.
The order matters
One rule prevents most beginner frustration: remove the gradient before you stretch.
Stretching amplifies everything, including the gradient. Once amplified, a gradient becomes far harder to model and remove cleanly, and attempts to do so leave patchy artefacts. Work on the linear data first.
The same applies to colour calibration, which works far better on linear data where the relationship between star colours and their catalogue values is still straightforward.
What not to do
Do not chase other people's results. Published images often represent many hours of integration from dark sites, and sometimes considerable artistic licence. Processing your ninety-minute suburban data to look like that produces something garish and noisy.
Do not over-saturate. The most common tell of an over-processed astrophoto is colour pushed past what the data supports, particularly magenta stars and neon nebulae.
Keep the original. Always process on a copy. The point of learning this is to try things, and that requires the ability to go back.
A reasonable first attempt
Take your best existing image, the one you were pleased with. Run it through gradient removal, then photometric colour calibration, then a gentle manual stretch. Compare against the original.
For most people the difference is larger than expected, and it is free. That is the entire argument for spending an evening on this.
Notes & sources
- Siril, free astronomical image processing software: calibration, stacking, background extraction and photometric colour calibration. Siril project documentation ↩
- GraXpert, open-source background extraction and denoising for astronomical images. GraXpert project ↩