My Astrophotography Processing Workflow
Unlike most other areas of photography, astrophotography typically requires a number of specialized tools and processing steps to fully realize an image. Below is my general workflow for processing images. It has evolved over the years as I have worked to continually improve and simplify the various steps wherever I can. It’s worth emphasizing that this is but one of many valid approaches to astrophotography image processing. It is not necessarily the best, simplest, or only method. That said, I hope it serves as a helpful reference for anyone looking to produce high-quality astrophotography images.
Downloading and Culling
After downloading the RAW images from my camera, I use Capture One to preview and cull any unusable frames. To better evaluate details, I typically increase the exposure by about 2 stops during review. I discard any images that lack sharpness—usually due to camera shake, atmospheric turbulence, or transient thin clouds. In my experience, retaining sub-optimal frames does not usually improve the final result. Any potential gain in signal-to-noise ratio (SNR) from using more images is often outweighed by the loss of fine detail introduced by including lower-quality exposures.
Raw Conversion
I use RawTherapee to convert camera RAW files into 16-bit TIFFs. It’s a powerful, feature-rich RAW converter that produces very good results – especially in terms of sharpness and noise control. While the interface is a bit unintuitive at first, it becomes fairly easy to navigate after some familiarization.
I have also used Nikon NX Studio, Capture One, and Adobe Camera Raw (ACR). None of them offer the same level of flexibility or detail as RawTherapee. Nikon NX Studio is notably slow and cumbersome. Capture One is easy to work with but poorly manages fixed pattern noise, often leaving significant artifacts in the background sky. ACR is perhaps the most viable alternative, coming close to but not quite to the level of RawTherapee in terms of sharpness. Differences in color rendition depending on image might cause me, however, to consider ARC. I hope to do a more detailed comparison of these raw converters soon.
I leave most of the setting in RawTherapee alone. Below are the settings that I typically change, grouped by tab:
Exposure Tab – I check Highlight Reconstruction and set method to: color propagation. This attempts to restore color information in small over-exposed areas (i.e. stars).
Detail Tab – I turn on noise reduction and defringe. I use the luminance and chrominance curve options at their default settings, which are set to bias noise reduction in the lower signal levels.
Under defringe I select the Curve: and Hue: option to adjust whichever bands best remove any star fringing. Zooming in on individual stars helps with this.
Raw Tab – I leave demosaicing at the default AMaZE method. I have also evaluated LMMSE, VN4, AMaZE + Bilinear, and AMaZE + VN4. They all seem to produce very similar results in files that I have tested. I also check the hot pixel filter and dead pixel filter boxes with a threshold level of 140. Lower levels risk eliminating stars.
Star Alignment and Stacking
I use DeepSkyStacker (DSS) for star alignment and stacking of light images. I don’t shoot flats or darks as they are really not necessary with today’s low noise cameras and post-processing software corrections. In some cases they can actually be detrimental to final image quality.
The only change in DSS that I make is in the Intermediate Files tab where I check Create a registered/calibrated file for each light frame. I leave the rest of the settings at their default values. The important settings in their respective tabs are shown below.
When processing images using a simple average, I’ve found that the results from the DSS Autosave.tif file are identical to those obtained by averaging the registered/calibrated frames externally. However, simple averaging can sometimes have trailing artifacts—such as those caused by airplanes, satellites, meteors, or stuck pixels. In those cases I do a Sigma Clipped average using astro-sca on the individual registered and calibrated light frames. I don’t use the Kappa Sigma clipping option in DSS, as it is significantly slower than using astro-sca.
Program astro-sca is available for free and is available here: Fast Sigma-Clipped-Averaging With astro-sca.
Image Stretching
For this step I use astro-color-stretch for image stretching and color correction. It is a one-step processing program where I can quickly and easily experiment with various stretches, curves, and color enhancement. I use an initial set of parameters for a baseline run to see how much stretch, curves adjustment, and color enhancement works best for the image. Program astro-color-stretch is available for free and is available here: Fast Image Stretching With astro-color-stretch.
Final Processing
After running astro-color-stretch I usually do a couple of minor adjustments in Capture One such as cropping and setting a final black point. The last step is noise reduction using Topaz Denoise AI. I usually go with the low-light mode and iterate a few times to get things right. It is a balance—too much noise reduction can soften fine details, while excessive sharpening can give the image an unnatural or artificial appearance.
Additional Resources:
Astrophotography Page