Imaging session NGC 2304 intermediate spiral galaxy

 My secret indulgence – imaging from the back garden. SSh - I've said it out loud! 

Sometimes it is nice not to have to pack the car with all the gear to drive up onto Dartmoor or down to the coast. Not that it’s an onerous drive – twenty minutes’ drive north or south!

Occasionally, I get to image from my back garden. It isn't without its limitations:

• I can only see a northern sky - from NW around to NE
• The house cuts off the lower part of that sky range - so I can only target images higher in the sky
• Invariably I will have to do a meridian flip at some stage of a night shoot
• It is a Bortle 5 sky and so I need to use clip in filters

Last night I snatched a few hours targeting NGC 2403, an intermediate spiral galaxy in Camelopardolis, a mere 9.7 million light years from Earth.  The galaxy contains numerous HII star forming regions and is 90,000 light years in diameter. Its spiral arms are quite short and I was pushing my equipment to the maxium to get any image at all frankly.  

My equipment rig:

• Astromodded Canon 800D 
• Optolong L-Pro clip in filter
• Zenithstar star 61ii with optional 61A field flattener
• EQM35Pro mount
• Autoguiding system comprising: ASIair Mini, ZWO 120mm mini guide cam and RVO 32mm guide scope
• Celestron Lithium Pro power tank
• 3 power banks for dew heater bands and dslr power

Shooting details: 

• 240" x 40
• ISO 800
• 20 dark, bias and flat calibration frames 

Post editing: 

• stacked and edited in SIRIL using my normal workflow
• finishing in Affinity Photo

Here it is - an OK effort but needs more work on it - I dont shoot galaxies very often - so I need to develop a better galaxy workflow process. 

Not quite sure which version is the better one - I tend to 'lose the wood for the trees' when post editing! 

Post Processing the Rosette Nebula in Siril — Full Workflow Guide

 Blogger isn't always the most intuitive platform to navigate - but it is free and simple to use and manage. 

To help you find information quickly on this blog, you can

• use the search bar using key words e.g. Asiair mini
• use the blog menu list by date 
• use this page which will be regularly updated https://undersouthwestskies.blogspot.com/2025/12/want-to-find-something-quickly-on-my.html
• visit my home page https://undersouthwestskies.blogspot.com/p/welcome-to-my-astronomy-and.html

Post‑Processing the Rosette Nebula in Siril — Full Workflow Guide

Capturing the Rosette Nebula (NGC 2237) over two beautifully clear nights gave me one of my favourite datasets of the year. If you’d like to see how the imaging sessions unfolded - the gear I used, the shooting settings, and the changing conditions across both nights - you can read the field report here:

Imaging Session: Rosette Nebula 2244
https://undersouthwestskies.blogspot.com/2026/03/imaging-sessions-rosette-nebula-2244.html

It’s a useful overview before diving into the edit, especially if you’re learning astrophotography or working with similar equipment.

A Quick Note for Beginners

If you’re new to deep‑sky image processing, don’t worry - this post walks through the exact Siril workflow I used to turn two nights of raw data into a finished image. Siril is a powerful, free tool for astrophotography, and once you understand the order of operations, it becomes a reliable and enjoyable part of the hobby. It is ideal for beginners and if you’d like a gentler introduction first, my beginner‑friendly Siril guide is here:

Beginner Tutorial: Workflow for Using Siril
https://undersouthwestskies.blogspot.com/2025/12/beginner-tutorial-workflow-for-using.html

The Final Image

Before we get into the steps, here’s the final Rosette Nebula image produced from this dataset. It’s a big improvement over my first attempt three years ago - a reminder of how much progress comes from better data, better tools, and a clearer understanding of post‑processing. I’m pleased with the result, though there’s still room to refine colour palette, clarity, texture, and structural depth in future versions.

My most recent effort at NGC 2237 The Rosette Nebula

And my very first stab at it three years ago

My Rosette Nebula Processing Workflow (Siril + VeraLux + SyQon + Cosmic Clarity)

Below is the exact sequence I followed – I hope it is a clean, efficient workflow that works especially well for OSC dual‑band data from cameras like my astro‑modded Canon 800D paired with an Optolong L‑Enhance filter.

1. Pre‑Processing in Siril

•            Creating the folders directory – a SIRIL folder containing separate folders for Lights, Biases, Darks and Flat calibration frames

·               Opening Siril. Then selecting the ‘home’ directory. Then - OSC Preprocessing Script — stacking and calibration

•            Loading the ‘results’ FITs image and cropping it to remove stacking artefacts

•            Auto Background Extraction (Siril preprocessing script)  

 

2. Linear Stage Enhancements

•            VeraLux NOX — refined background extraction

·              Optional Cosmic Clarity Denoise or VeraLux SILENTIUM — linear noise reduction

 

·              Tools - Astrometry / Plate Solving for accurate framing and annotation

 

·              Colour calibration  - optional PCC or SPCC but NOTE: if doing narrowband imaging and planning to use veralux ALCHEMY  - DO NOT  do colour calibration

 

·              Optional - VeraLux ALCHEMY — colour calibration (L‑Enhance data benefits hugely from early ALCHEMY to correct colour cast)

•            Siril script – pre-processing – ABERRATION REMOVER

·              Optional Cosmic Clarity Denoise or VeraLux SILENTIUM — linear noise reduction

•            SyQon Starless — generate starless and star mask layers (Starless‑first workflow avoids star bloat during stretching)

·             Optional - VeraLux ALCHEMY — colour calibration (L‑Enhance data benefits hugely from early ALCHEMY to correct colour cast)

 

3. Stretching enhancements

•            VeraLux HYPERMETRIC STRETCH (HMS) — main stretch applied to the starless image - HMS gives a clean, colour‑safe stretch ideal for dual‑band OSC (Or SyQon autostretch)

 

4. Post‑Stretch Colour & Contrast enhancements

•            VeraLux CURVES — tone shaping, midtones, and global contrast

·               VeraLux REVELA — local contrast enhancement on the starless layer

•            VeraLux VECTRA — perceptual colour grading

(REVELA → Curves → VECTRA gives controlled, natural nebula shaping)

•            Cosmic Clarity Sharpen — applied lightly to the starless layer

 

5. Star Handling

•            VeraLux STAR RECOMPOSER — blend stars back cleanly with natural colour

·              Optional CC denoise and sharpening

 

6. Final Output

•            Export as 16‑bit TIFF for finishing touches in Affinity Photo

Let’s now go through the workflow in more detail and look at some of the processing images and settings I used, and why I used them. 

Workflow in depth commentary

1. Pre‑Processing in Siril

•            Creating the folders directory

·               Selecting the ‘home’ directory - OSC Preprocessing Script - stacking and calibration

•            Loading and cropping ‘results’ FITs file to remove stacking artefacts

•            Auto Background Extraction (Siril preprocessing script)  

 

What is Auto Background Extraction — What does it do and why does it Matter?

I find that Auto Background Extraction is a fast, beginner‑friendly way to remove large‑scale gradients caused by light pollution, moonlight, uneven flats, or optical vignetting. Automatically analysing the frame, modelling the broad background glow, and subtracting it to create a more even, neutral starting point for my edit - helps reveal faint nebula structure and gives me a cleaner foundation for colour calibration and stretching, later on.

Because ABE is fully automatic, it’s ideal for quick processing or simple gradients where I don’t need the precision of manual background extraction. However, it can sometimes misinterpret faint nebulosity as background, especially in wide‑field images with lots of emission. For complex or multi‑directional gradients, a manual approach is usually safer (I might then use Siril’s own background extraction or dip out into GraXpert) — but for most OSC dual‑band data, ABE seems a reliable first pass.

In nearly all cases, I just use the default settings that are shown in the op-up window. 

2. Linear Stage Enhancements

•            VeraLux NOX — refined background extraction

·              (Optional) CC Denoise or VeraLux SILENTIUM — linear noise reduction

·              Astrometry tools - Plate Solving

·              Colour calibration - optional PCC or SPCC but not if using ALCHEMY later

·              (Optional for narrowband)  - VeraLux ALCHEMY — colour calibration

•            Siril script – pre-processing – ABERRATION REMOVER

·              (Optional) Cosmic Clarity Denoise or VeraLux SILENTIUM

•            SyQon Starless - starless and star mask layers

·             (Optional on narrowband) - VeraLux ALCHEMY — colour calibration

 

 

Why do I do image plate solving early on in the post editing work flow?

Plate solving anchors the image to its true sky coordinates - the exact RA/Dec position, pixel scale, and field of view, so that Siril gains a precise map of where the Rosette Nebula sits in the night sky. This  information becomes the foundation for later tools, especially when working with deep‑sky objects.

Solving early also ensures the image is correctly oriented before any stretching, cropping, or star removal. If I wait until after heavy processing, plate solving often fails entirely – and no – I don’t know why this is. Running it at the start keeps the World Coordinate System (WCS) intact, which is essential for accurate annotation, star‑based colour calibration, and multi‑night alignment — all common tasks when editing targets like the Rosette Nebula.

Makes sure when completing plate solving that you enter the right focal length and camera sensor details into the op-up window – otherwise the plate solve is likely to fail. 

Why do I use Veralux Nox now to reconstruct the physical background?

I use Veralux Nox early in my work flow. Here is the science bit – most of which I don’t understand! It is one of those “just trust the process” scenarios! NOX uses a deterministic background‑modelling engine based on the Zenith Membrane concept — a method that rebuilds the sky background as a smooth, physical surface rather than relying on manually placed sample points. Instead of fitting curves like traditional ABE tools, it solves the Discrete Poisson Equation to create a natural, continuous background model. The result is cleaner gradients with fewer artefacts, especially in wide‑field DSLR and mirrorless astrophotography.

Phew! I hope you understand that better than I do.

I have found that running this process early, while the image is still linear, gives the most accurate reconstruction of the true sky background. NOX handles difficult gradients from light pollution, moonlight, and lens vignetting far better than simple automatic extraction, and it usually protects faint nebulosity more reliably. For workflow‑driven processing — especially when using the VeraLux suite — it’s fast, consistent, and requires almost no configuration.

Could I just use it instead of the Auto Background Extraction? Probably! Does it seem to do any harm to any of my images using it in conjunction with the latter? Not that I have seen thus far!

Like any global model, NOX has limits. Extremely complex gradients or very faint diffuse nebula can still challenge it, and you don’t get the fine control of manual DBE‑style tools. But for most OSC dual‑band or wide‑field data, this early background reconstruction provides a clean, trustworthy foundation for colour calibration and stretching in a modern Siril astrophotography workflow.

So, what settings did I use in the NOX interface?

• ·       PSF Auto‑Masking: ON – why? My Zenithstar 61II produces tight stars and the Rosette sits in a dense star field.  Dual‑band filters such as my clip in optolong L enhance filter slightly bloat brighter stars and PSF masking prevents NOX from touching star cores

• Auto‑Calculate: YES  - I run it first before touching anything else – why? It analyses my linear noise profile and adapts to my DSLR noise (which is higher than CMOS astro cams). It then compensates for the OIII noise from the L‑Enhance and sets a safe baseline for the rejection algorithm.  Think of Auto‑Calculate as “getting NOX in the right ballpark” after which I refine the values.

• Signal Rejection Power slider: This is the main “how hard should NOX work?” slider. My Canon 800D has visible linear noise and the L‑Enhance filter can produce noisy OIII. Given the Rosette has delicate Ha filaments that must not be blurred and I had over three hours of integration time – I opted for a slider setting of 35 – 40% - I think this range smooths the background without flattening the nebula. If I had shorter integration times, I’d have gone for 45 – 50%; or if the background had looked too noisy.

• Membrane Stiffness: 0.55–0.75 This is the “structure protection” slider and higher stiffness = more protection of dust lanes, the Rosette’s central cavity, the sculpted Ha edges and the faint outer petals. Initially I opted for 0.55–0.65 which gave me a balanced smoothing + structure but then because I had longer integration times, I tried 0.70–0.75 instead. It was a fine balancing act – when the background looked grainy, I lowered the stiffness slightly; when the nebula looked ‘plastic’ or ‘too soft’, I raised it.

When the image looked to over processed – I tried decreasing both sliders a little and that seemed to work. The key thing to remember is that NOX is subtle — it’s meant to refine, not transform.

At the end of the day, I think the settings worked for my data because:

My Canon 800D (astromodded) – has a strong Ha response, but noisy linear data, so moderate rejection needed.  The Optolong L‑Enhance occasionally causes noisy OIII.  My Zenithstar 61II is a wide field refractor scope, so PSF masking is essential to protect stars.  The Rosette Nebula has soft emission + fine dust structure, so high membrane stiffness is required.

you can find out more about Veralux NOX here: 

Next steps – optional Denoising and Aberration Remover

VeraLux SILENTIUM — Linear Noise Reduction for Clean, Smooth Data

VeraLux Silentium is a specialist linear‑stage denoising tool designed for Siril, created to clean up deep‑sky astrophotography data without damaging the delicate structures that make nebula images so compelling – before stretching takes place.

It works directly on the untouched linear signal, removing noise while preserving fine structure — an ideal approach for my DSLR and OSC dual‑band imaging where thermal noise and pattern noise can be really noticeable. Because Silentium operates before any stretching or contrast enhancement, it reduces noise at the stage where it’s easiest to control and least likely to damage detail.

Linear denoising is widely considered best practice in astrophotography because it treats noise when it is most mathematically manageable. Silentium excels here by gently smoothing the background while keeping star profiles tight and preserving the subtle gradients that define large emission nebulae. This is particularly important for the Rosette Nebula, where the outer petals and faint hydrogen structures can easily become blotchy or uneven if noise is left untreated until later stages.

Silentium doesn’t seem to soften my stars or nebula structure, so I find it great and useful when working with faint targets, short exposures, or data captured under light‑polluted skies like mine.

It’s also a strong alternative to AI‑based denoisers when you want a predictable, repeatable result that integrates smoothly into a Siril‑based workflow.

Silentium has limits. It can struggle with extremely noisy data, and it won’t fix banding, walking noise, or calibration issues that originate in the capture stage. It also offers less fine control than dedicated AI denoisers.

Once again, being new to astrophotography, and somewhat lazy at times, I tend to use recommended default settings.

 Find out more about Silentium here: 

What does Aberration remover do and why do I do it at this point in the workflow?

Aberration Remover is a corrective tool in Siril designed to fix optical distortions such as elongated stars, chromatic aberration, and mild field curvature that often appear in wide‑field or fast‑aperture astrophotography. It analyses star shapes across the frame and applies subtle geometric corrections to restore roundness without degrading fine detail. In a typical Siril workflow, Aberration Remover sits after stacking and background extraction but before colour calibration and stretching – something I am always forgetting!  This placement ensures the algorithm works on clean, linear data and prevents later steps from amplifying distortions. I keep forgetting to put a post it next to my screen to remind me of this!

Its main advantages are cleaner star profiles, improved edge‑to‑edge consistency, and a more polished final image - especially valuable for cameras like my DSLR and mirrorless and budget refractor. The trade‑offs are that heavy corrections can introduce artefacts, slightly soften stars, or mask underlying optical issues.  Used with restraint - Aberration Remover is a powerful way to elevate our deep‑sky images and streamline our Siril processing workflow.

More here at: 

What is the point of using Veralux Alchemy at this stage of the work flow?

VeraLux ALCHEMY — Colour Calibration & Narrowband Normalisation

VeraLux ALCHEMY is a fast global colour‑calibration and colour‑normalisation tool designed to bring astrophotography data into a physically meaningful, visually balanced colour space. It aligns colour channels, corrects colour casts, and produces a neutral, consistent baseline for later enhancement. Think of it as the “colour foundation” step in the VeraLux pipeline – it gives me consistent results and is ideal for repeatable workflows. This makes it especially effective for OSC dual‑band filters like the Optolong L‑Enhance, which often produce strong colour casts straight out of the camera. It also seems to work well with my astromodded Canon 800D.

Why Do I use it before star removal?

If I use ALCHEMY after star removal, I find the star extraction alters local intensities. The star mask no longer matches the original channel structure and the normalisation becomes biased with unpredictable colour balance shifts. Alchemy is designed to work on pure, linear, unmodified masters. In addition, I have found that star removal before using ALCHEMY introduces artifacts - some small residual halos, localised smoothing and slight structural distortions.

ALCHEMY also performs colour‑safe narrowband mixing, allowing SHO, HOO, or custom blends without the usual problems of washed‑out OIII, oversaturated Hα, or muddy yellows. Because it relies on the full, intact linear data, it must run before star removal. Running it early ensures clean gradients, accurate colour calibration, and a stable foundation for later modules like REVELA, CURVES, and VECTRA.

In practice, ALCHEMY is fast, automated, and highly consistent, making it ideal for repeatable workflows in Siril astrophotography. It excels with OSC and wide‑field DSLR data, correcting global colour imbalance and preparing the image for stretching. Its limitations are mostly tied to its global nature — it can’t fix local colour issues, may slightly mute very faint data, and isn’t a replacement for precision star‑based calibration in scientific workflows. But for most deep‑sky imaging, especially dual‑band nebula work, ALCHEMY provides a reliable, colour‑accurate starting point that simplifies the rest of the edit.

Again, I tend to gravitate towards using the default settings that appear in the pop-up window. However, I also experiment in moving the sliders one at a time. A double click on the slider returns it to its default position. 

So, what where my setting choices on this particular workflow session?

1.      Sensor setting - Canon 600D – the closest option I think to my Canon 800D sensor.  I think the Canon 600D and Canon 800D share the same fundamental Bayer pattern (RGGB) and extremely similar colour response curves. Even though the 800D is a newer sensor, the differences are small enough that Alchemy’s spectral‑unmixing and colour‑normalisation algorithms will behave correctly.

2.      Quantum Unmixing: ON - Why: The L‑Enhance filter blends Ha and OIII in the green channel but Quantum Unmixing mathematically separates them giving me cleaner teal OIII, deeper red Ha and preventing the “pink soup” look common in DSLR dual‑band data. For the Rosette Nebula, this is essential — the OIII halo is subtle and easily lost.

3.      Normalisation Section - Background Neutralisation: ON - Why: DSLR dual‑band data often has a red cast, made worse by my astromodded change.  The Rosette sits in a star‑dense region with uneven background and so neutralisation gives me a clean baseline before stretching which is almost always beneficial

4.      Auto Signal Fit: ON  - Why: Well, it balances Ha and OIII intensities, prevents Ha from overwhelming the palette, ensures the OIII halo around the Rosette remains visible and compensates for DSLR colour bias – which is especially important with the L‑Enhance, which strongly favours Ha. Or, so I am led to believe!

5.      OIII Boost slider: I played about with it between 0.20–0.30 as I had long integration time data. Remember my L EnHance filter produces weaker OIII but the nebula has a real but subtle OIII halo. Boosting too much creates fake teal rings whilst boosting too little makes the nebula look monochrome red. With shorter integration time data I’d have tried  0.30–0.40.   

6.      Palette Mixer (R / G / B sliders) - This is where we shape the final colour palette. I was trying to achieve strong Ha in red, OIII contributing to the teal and blue, and a minimal green cast.

a.      Red slider: I played with 85–100% Ha and setting the red slider close to the Ha side (left). Why? Because the Rosette is overwhelmingly Ha‑dominant, my astromodded 800D captures Ha extremely well and the L‑Enhance filter boosts Ha even further. Too much OIII in red makes the nebula look pink or salmon‑coloured

b.      Green Channel: 10–20% OIII – Why? Too much green muddies the Rosette but a small amount of green helps star colour look natural and also helps blend Ha and OIII smoothly. This all keeps the nebula clean while avoiding the “Christmas tree green” problem.

c.      Blue Channel: 40–60% OIII - blue slider toward the OIII side, moderately high. Why? A hunch based on little to no knowledge whatsoever, but if you want my justification -  OIII lives in the blue channel, the Rosette has a real but subtle OIII halo, L‑Enhance filter captures OIII, but weakly and so boosting blue helps reveal the teal/blue OIII without faking it.

Hopefully this gives me a natural dual‑band look: red Ha core + soft blue‑teal OIII halo.

 

So much of post editing is based on personal preference and for a beginner – it can be hard to decide whether an image you craft in ALCHEMY is truly representative of the actual colours in reality. Every time I process the same data set – I come out with a slightly different colour version!  At the end of the day, I go with the philosophy – if I like it – that’s fine.

As I have already said, I know that The Rosette Nebula images I see on Astrobin, often have more teal colours and an orangery almost copper tint at the edges – which I never seem to obtain. So be it! I will get there one day! 

Another great video from Rich - here explaining ALCHEMY: 

SyQon Starless — What it is and why do I use it here in my work flow?

I have only just started using SyQon Starless for the first time – on this actual image as it so happens! Before this, I always used Starnet++.

SyQon Starless is an AI‑powered star‑removal tool designed specifically for astrophotography workflows in Siril. It removes stars while preserving nebula structure, enabling clean starless processing and later recombination.

Using an AI model (Axiom v2) trained for astrophotography, I found it preserved the fine nebular detail and structure during removal, thus producing a clean starless layer for colour work, noise reduction, or contrast shaping late on. My starless image seemed to have minimal artifacts and good structure definition. It was fast processing, and once again I relied on recommended default settings. 

On the starmask image however, I did find some very minor residual halos around brighter stars. 

3. Stretching enhancements

•            VeraLux HYPERMETRIC STRETCH (HMS) — main stretch applied to the starless image - HMS gives a clean, colour‑safe stretch ideal for dual‑band OSC (Or SyQon autostretch)

Veralux Hypermetric Stretch

It is only recently that I have switched across to VHS – from using the existing siril image stretching tools.

VeraLux HyperMetric Stretch (HMS) is a next‑generation, colour‑safe stretching algorithm that transforms linear astrophotography data into a visually meaningful image while preserving photometric integrity. It replaces traditional curve‑based stretching with a physics‑informed, mathematically controlled approach. Non-destructive – it is a stretch that avoids the hue shifts common in other standard histogram or curves stretch tools.

So far, I have found it preserves star and nebula colour far better than the traditional curves approaches I was using – but this may also be because I didn’t fully understand what I was trying to do when using these. I am better able to bring out low signal structures without blowing out the highlights – so I’m getting cleaner faint-detail recovery. It is easy to, repeatable and consistent.

More experienced users of Siril may get frustrated using it – there is less artistic freedom than users might get using manual curve shaping. You do have to get your head around a learning curve - understanding the hypermetric philosophy – so that you get the best results.

 

So what settings do I use in the VHS pop-up window? 

Firstly, I choose either the recommended sensor setting as seen in above image – or canon 600D. This ensures the stretch engine models my sensor’s response curve correctly, especially important for dual‑band Ha/OIII data.

Then I tick ‘Ready to Move’ and leave ‘Scientific Preserve’ unticked. “Ready to Move” gives us a visually pleasing, contrast‑balanced stretch suitable for artistic astrophotography. I am told by copilot that “Scientific Preserve” is too conservative for dual‑band nebulae and will leave the Rosette flat and under‑stretched.

With regard to the Stretch Engine & Calibration section -

Target BG (background) to whatever default is presented – in my case it was 0.20. I am aiming for a clean, dark background without crushing the faint OIII halo around the Rosette.

Adaptive Anchor Selection is left at ON to help the algorithm protect the bright core while still lifting the faint outer petals.

Log‑D (checkbox + slider) was defaulted at 4.21 and once again – I went with the default recommendation.  Log‑D controls how aggressively midtones rise. Aiming to reveal the faint outer structure without blowing out the central star cluster, I auto-calculate these default settings first and from there decide whether I need to play about with this particular slider further to get the result I want.

The Protect BG is ticked to ON. As an L-enhance filter can cause noisy data in the background, I went with 6.0 on the slider to hopefully prevent the stretch from amplifying this noise further. I also hoped it would help keep any dust lanes natural.

I moved the Physics / Colour Engine slider to just over half way across, after a period of experimentation! I was trying to keep the Ha dominant but also give the OIII sufficient lift to show the inner teal regions, without turning the whole nebula cyan!

Star Core Recovery (checkbox) ON and set at default 3.50 - Essential for the Rosette — the central cluster is bright and easily saturates during stretching.

In reality, you need to play about with these settings to get your desired image – preview is great and holding the space bar allows you to track the changes between your current stretch and the original image.  

As always, I was aiming for a ‘balanced’ stretch which would give me a preserved bright core, visible outer petals, colour separation between Ha and OIII and minimal noise amplification.  

How do you think I did?

A screen grab from when I was in ALCHEMY mode. 

And let's remind ourselves of what I was achieving part way through this work flow

By now you will have gathered that I am a great fan of Deep Space Astro. He has done so much to help beginner SIRIL users get to grips with the software. You can find his channel here: https://www.youtube.com/@DeepSpaceAstro/featured

Ok, well done for staying the journey so far. We are heading into the homeward stretch: We are now into the toning and crafting phase of the work flow.

4. Post‑Stretch Colour & Contrast enhancements

•            VeraLux CURVES — tone shaping, midtones, and global contrast

·               VeraLux REVELA — local contrast enhancement on the starless layer

•            VeraLux VECTRA — perceptual colour grading

(REVELA → Curves → VECTRA gives controlled, natural nebula shaping)

•            (Cosmic Clarity Sharpen — applied lightly to the starless layer)

Let's start with Veralux CURVES

VeraLux CURVES is a precision tone‑mapping and contrast‑shaping system built around spline‑based photometric control. It allows you to adjust luminance and colour channels with extremely high fidelity, using interpolation methods designed to avoid artefacts and preserve astrophotographic structure. It will shape tonal curves smoothly and predictably, allow independent control in RGB, Lab, LCH, and HSV colour spaces and separates luminance structure from colour intensity, giving far more control than standard curves tools.

Essentially you can

·              Adjust luminance without shifting colour

•            Adjust colour intensity without affecting brightness

•            Target hue‑specific structures (e.g., OIII vs Ha regions)

·              Apply curves only to starless data

•            Protect highlights

•            Boost faint nebulosity without blowing out stars

The result is fine‑grained contrast shaping without introducing ringing, banding, or colour shifts. Great for astrophotography workflows where faint‑detail preservation is critical or for shaping nebula contrast without damaging star colour.  It avoids colour clipping, star bloat, background colour shifts and non‑linear distortions

This makes it far more reliable than generic photo‑editing curves.

So, the downsides?

It is more complex than a standard curves tool – well I think so! It requires an understanding of colour spaces that I don’t seem to possess. It isn’t a ‘quick-fix’ tool and has to be deliberately placed and used in a discrete workflow. 

I also think I still have much to learn about how to use curves as a tool and that's half the problem! 

Ok, so how do I use it?

I do RGB adjustments first – a series of shallow ‘S’ curves.

Is this the correct approach?  I genuinely have no idea but I’m sure I read somewhere that you have to do this first.

After that I start to work as follows:

Luminance – Now this is important – make sure you do this bit first before creating a curve for luminance – the ‘L’ box.

 Tick the ‘enable range limiting’ box and then tick the ‘show mask’ box. Use the two sliders to position the mask (the white area) over that part of the nebula extent you want to work on – then untick the ‘show mask’ box. You have now isolated the rosette nebula from the background – you can work on the nebula rea without affecting the background – although keep an eye on it.

Now you can create a gentle S‑curve for luminance – click ‘L’ on if you haven’t done so yet - add a point near the left (shadows) to anchor shadows and avoid crushing the dark dust lanes. I then lift the midtones by adding a similar point in the midtones, pulling it slightly up from the diagonal line to brighten the main Rosette body.

And then it is a similar process repeated to protect the highlights – adding a point near the right and keeping it close to the diagonal so the bright inner regions don’t blow out.

My Goal? To increase separation between the dark dust and glowing gas without making the background too bright. 

I found that it took a couple of ‘S’ curves on luminance – each time instigating the mask and using the slider to determine its extent. Slowly, a series of shallow ‘S’ curves were applied – lifting the midtones slightly, inserting a dip just below the midtones.  Contrast within the nebula filaments, petals and inner structure slowly emerged.

I used a similar approach with the C channel – chrominance  – boosting the nebula colouration carefully whilst ensuring the nebula mask was active. Key to it is to ensure the end points of the curves are not clipped.  As always, keep an eye on your background – if it starts to show changes in noise or colouration, reduce or tighten the effect of your nebula mask.  I was trying, at this stage, to increase the richness of the Ha/OIII mix.

I then repeated the same steps and approach to ‘S’ – saturation channel.  In the screen grabs here, they don’t show the nudge towards the teal colour I was trying to achieve.  I was also trying to work out how to bring out golden and yellow hues – something I failed at completely.

One last thing - you have to click apply after every change you make. And this is good - why? Because if you mess it up - click unapply and you go back a step! I cannot begin to calculate how many times 'undo' got hit! 

I should have also said earlier, at every stage I always resave the image I am working on – extending its file name so that it shows the last process completed on the image. So, I may end up with a file name something along the lines of ‘results_ABG_VSil_VAlch.fits’ and so on. When a file name starts to get too long – I rename it something completely different – starting with the last process just completed. 

Another explanatory video from rich here: https://www.youtube.com/watch?v=2Of-6qmgCOM

Moving on in my tones and colour sculpting phase I now use Veralux REVELA.

VeraLux REVELA

REVELA  is a signal‑aware local contrast enhancement tool that enhances fine detail in nebulae and galaxies without boosting noise, creating halos, or damaging stars.

Analysing the signal‑to‑noise characteristics of the image before applying contrast changes, REVELA enhances real structure while suppressing noise amplification. It also avoids common artefacts such as dark halos, crunchy backgrounds, and broken star profiles. It is a post‑stretch tool in the Veralux pipeline. I have used it after Veralux HMS and before curves. I have used it after curves.  

I particularly like the way it boosts detail only where the data supports it and in the way it avoids the dark rims you typically get with other local contrast tools. 

When using REVELA – use it gently! This is a tool that can amplify noise!

It is worth reminding ourselves at this stage that my image has been taken on an astromodded Canon 800D with a clip in optolong L-enHance filter. So I gained an image with strong Ha signal, slightly muted OIII, naturally soft stars and a tendency towards micro-noise in the red channel.

Let’s go through section by section.

The Enhancement Section

This is the “detail engine” of Revela. It works on local contrast, not global contrast.

Texture Micro‑Contrast – this slider boosts very fine detail and so accentuates dust lanes, filaments and subtle gradients. IT CAN ALSO AMPLIFY NOISE if pushed too far! My goal was to amplify the filaments without over exaggerating red channel noise.  Basically – you have to play about with the slider and use the space bar to see the changes made compared against the original.  Remember – the grey defaults button returns everything to the default settings.

My tip?

If you see “sandpaper” texture in the background, back off the slider towards the left-hand side.

 

The Structure Volume - Enhances larger‑scale structures, adds depth and dimensionality and makes the nebula look more 3D.  It is much less likely to create noise than the micro‑contrast slider above.  This is the slider that will bring out the “petal” shapes, and help define the central cavity.

Another tip or two?
If the nebula starts looking “inflated” or plasticky, reduce this slider back towards the left-hand side a little. Don’t over-enhance the central cavity. It is easy to make it look punched out – which I may be guilty of in my final image! I over did the structure volume slider!

 

The Protection Gate Section

This is where REVELA prevents damage to stars and shadows while you push the enhancement engine. The shadow authority/noise gate slider protects dark regions from being over-enhanced, stops noise amplification in the background and shields low signal areas. So, your background should remain smooth while the nebula gets detailed! It is basically an adjustable mask that you can determine the areal extent of.

Tip -
Increase this until the background stops looking gritty, then stop.

 

The Isolate Stars / Prevent Raccoon Eyes box

This is the star‑protection system which detects any stars, masks them before enhancement and prevents dark rings. The slider controls how strongly the stars are protected. The more you go to the right – the softer and rounder the stars.  I always tick this box to ON and I aim to keep any stars tiny – that is how they appear in this nebula naturally. If my stars are bloating – I increase star protection or reduce micro contast.

 

Finally a mega tip - Apply noise reduction before REVELA

REVELA enhances whatever is present — including noise.
A light denoise pass beforehand gives you more headroom.

The last tone and colour enhancement tool in my work flow before star re-composition – veralux VECTRA. 

This is the one I find trickiest to understand and use effectively and many times I blew things at this stage!

VeraLux VECTRA  is a perceptual vector‑based colour‑grading engine built on the LCH colour space (Lightness, Chroma, Hue). It is designed to let you manipulate colour in a way that aligns with how the human eye actually perceives colour, giving far more natural and controlled results than traditional RGB‑based adjustments.

When correctly applied, VECTRA should give us intuitive control over hue, saturation and brightness. I should be able to get targeted colour shaping without affecting star colour or introducing artifacts! It should be my final-polish colour tool.

An awful lot of ‘should’s’ there!

I think my problem is simple – it is a complex tool – and I don’t fully understand it – so I operate mainly on trial and error and guess work! Which I hate doing!

Anyway, exploring the pop-up panel – you will see two tabs – one for primary vectors and one for secondary.

Primary vectors are red, green and blue and each vector has two sliders – one for hue (which rotates that colour around the colour wheel) and one for saturation – that increases or decreases the intensity of that colour.  

Before we go any further – toggle on the vector wheel box in the top right-hand corner of the screen – it should then appear in the bottom left.

Going through each slider:

RED Vector (Hue + Saturation) - controls the hue and saturation of H‑alpha emission, affects the Rosette’s petals (the big red structures) and also influences red channel noise.

Hue slider

• Moves red toward orange (clockwise) or magenta (counter‑clockwise)
• Small adjustments only — red is dominant in L‑Enhance data

Saturation slider

• Boosts or reduces the intensity of the H‑alpha petals

 

GREEN Vector (Hue + Saturation) - Controls green channel contamination. The L‑Enhance filter often produces a slight green cast in OIII regions so adjusting green helps balance the core colour.

Hue slider

• Shifts green toward cyan or yellow
• Useful for removing greenish tint in the core

Saturation slider

• Reducing green saturation helps reveal the true teal OIII colour

 

BLUE Vector (Hue + Saturation) - Controls the OIII component, and affects the central cavity and faint outer OIII shell

Hue slider

• Moves blue toward teal or purple
• Teal is more natural for L‑Enhance OIII

Saturation slider

• Boosts the OIII signal, which is weaker than H‑alpha

 

SECONDARY VECTORS

These affect Yellow, Cyan, and Magenta — the combinations of the primaries.

They are extremely useful for fine‑tuning nebula colour separation.

 

YELLOW Vector (Hue + Saturation) - Controls the red+green overlap, helps remove muddy brown tones in the nebula and is useful for cleaning up dust lanes

 

CYAN Vector (Hue + Saturation) - Controls the OIII‑dominant areas which is very important for the Rosette’s core; and also helps refine the teal colour

 

MAGENTA Vector (Hue + Saturation) - Controls the red+blue overlap, helps prevent the nebula from turning pink and is useful for balancing the transition between H‑alpha and OIII

 

PROTECTION: Neutrality Lock + Shadow Authority + White Star Integrity

These are essential for keeping the image natural.

Neutrality Lock — Shadow Authority Slider - Protects dark regions from colour distortion, prevents noise in the background from becoming tinted and keeps shadows neutral

White Star Integrity (Tick Box) - Prevents stars from picking up colour shifts, keeps stars white or naturally tinted and avoids teal, magenta, or red halos caused by vector adjustments

 

Well, if that’s the theory, what is the reality in use for me?

I suck at it! I know what I want to achieve:

·       Subtle OIII with a soft teal, not bright blue colouration.

·       Avoiding oversaturating the Ha

·       Protect the stars to stop colour shifts

Up to now, I can just about use the primary vectors tab. The moment I go into the secondary one – I mess up big time!

I know you don’t want to hear this if you are beginning. But, this is important – knowing what you get wrong is important – because it’s the area you have to target in order to improve! Simple! 

And finally, at the end of what could well be the longest post I have written for this blog - we get towards the end: 

Veralux STAR RECOMPOSER 

VeraLux STAR RECOMPOSER (often referred to as StarComposer) is a precision star‑reintegration tool designed to blend stars back into a starless image without colour shifts, halos, or “pasted‑on” artefacts. It uses the same recomposition engine as HyperMetric Stretch (HMS) for physically accurate colour preservation.

It recombines our star mask with our starless layer using a mathematically controlled blending model, thus preserving true star colour, avoiding the common problem of stars becoming too white or too saturated and maintaining natural transitions so stars sit correctly in the stretched image. No blown-out cores and harsh edges.

It isn’t, however, a star reduction tool. It only recomposes and not reshapes.

I use it when I’ve processed a starless image and need to add stars back in cleanly without losing their colour accuracy.

So here is what I did in each section on the pop up menu.

1. Load Star Mask

• This is our stars-only image from when we did Starnett++ separation earlier in the work flow.  VeraLux uses it to rebuild the star field with physically‑based brightness and colour behaviour.

2. Load Starless

• This is our starless nebula/galaxy image – the one we have done a whole amount of work on.  It becomes the base layer onto which the recomposed stars are added.

3. Composition Mode

Screen Safe (checkbox)

• Uses a screen‑style blend that avoids blowing out highlights and is good for natural‑looking stars and protecting bright nebula cores. So I always tick this to ON.

Linear Add (checkbox)

• Adds star brightness mathematically with no protection and produces stronger, brighter stars but can clip easily.  Use only if you want a punchier star field. I have yet to use it.

4. Sensor Profile (dropdown)

• Applies a camera‑specific response curve so star colour and brightness behave more like real sensor physics. It helps avoid colour shifts or unnatural star halos. If unsure, pick the profile that matches your camera brand or leave on default.  I either use default or opt for Canon 600D which is the nearest sensor to my 800D DSLR.

5. VeraLux Stretch Section

This is where you control how the stars look.

 Star Intensity Log (slider)

• Controls how bright the stars become after recomposition. Low = subtle, natural stars. High = bright, punchy stars.  This is the main slider you’ll adjust. I play about with it until I gain what I think is a pleasing image.

Profile Hardness (slider)

• Controls how tight or soft the star profiles are. Low = softer, more bloomy stars. High = tighter, more defined stars. Useful for avoiding “fat” stars after recomposition, it is another one I play about with by trial and error until I get something I am pleased with.

Adaptive Anchor (checkbox)

• Automatically adjusts the star stretch so the brightest stars don’t blow out and helps maintain colour in bright stars. I always it leave ON.

6. Physics Section

These controls refine realism and hide artefacts.

Colour Grip Blend (slider)

• Controls how strongly VeraLux preserves true star colour. Low = softer, more blended colours. High = stronger, more saturated star colours. Great for restoring natural yellows, blues, and reds. I normally tend to go with the default setting.

Shadow Conv Hide Artifacts (slider)

• Reduces dark halos or extraction artefacts around stars. Higher values hide more artefacts but can slightly soften the star edges. Adjust only if you see halos or weird dark rings.

Show Star Surgery (checkbox)

• Visual debugging mode. Highlights where VeraLux is repairing or reconstructing stars. Not for final output — just for checking problem areas. Normally I just go with default settings or leave it be.

 

I hope this workflow walk through has helped you. it is aimed at beginners new to SIRIL but maybe there are one or tgwo points that might help intermediates as well. As always, I may have some things wrong - in which case - drop me a line in the comment box below and I will make immediate corrections. 

Clear skies, stay safe, have fun out there under the cosmos

Steve  aka Plymouthastroboy. 

My March 2026 versions above and below - the result of this workflow post

My first two attempts at the Rosette nebula back in 2022 and 2023 (above and below) 

My late 2025 attempt