Super Structure Module Use (v1.7)

Notes from users, documentation addendums.
Post Reply
Posts: 158
Joined: Thu Feb 19, 2015 8:35 am

Super Structure Module Use (v1.7)

Post by Guy »

Here are some notes relating to using this module. It is not the only way to use the module - please experiment and provide feedback.
They relate to StarTools version 1.7.430 and later.
Please let me know if anyone sees any errors or has any additional advice they think helpful so this post can be improved.
I will update this post as needed.
To see a full alphabetical list of module topics click here.
For an index of similar notes on the other StarTools modules see StarTools Main Window Use.

Super Structure Module

  • To process large scale structures in the image (e.g. galaxies & large scale nebulosity) independent of the rest of the image to allow effects such as de-emphasising busy star fields, emphasising large structures or enhancing their colour.
  • To put back the light diffraction effects in the image that the processing has distorted and suppressed.
  • To bring back 'life' into an image by remodelling uniform light diffraction. Removes the flatness that heavy processing can cause bringing back some of the 3d effects.
  • To lift objects from very noisy data or dense star fields.
The Super Structure module restores the uniform light diffraction of the image that may have been distorted and suppressed by the stretching and other processing that the image has undergone.
This light diffraction - the Airy Disc around stars - shows as a 'glow' and adds 'life' to the image. The Super Structure module restores the light diffraction.
The Super Structure Module supersedes the Life Module - but has a similar purpose and a number of controls in common.

Useful Sources
For a module description see Super Structure: Global Light Diffraction Remodelling of Large Scale Structures.
The Official StarTools English Manual. It relates to StarTools version 1.6+.
The Unofficial StarTools English Manual is another good general source of help. It relates to StarTools version 1.7.
The notes below relate to StarTools version 1.7+.
There are a number of useful links in StarTools Links and Tutorials.

When to use:
  • Towards the end of the workflow - usually after using the Color module and before the final Denoise.
Example Workflow (v1.7):
AutoDev-{Lens}-Bin-Crop-Wipe-AutoDev (or FilmDev)-{Contrast/HDR/Sharp/Decon/Flux}-Color-{Shrink/Filter/Entropy/SuperStr}-Track/NR-{Layer/Heal/Repair/Synth/Stereo 3D}
Key: {...} optional modules

Example Workflow (v1.8):
{Compose}-AutoDev-{Lens}-Bin-Crop-Wipe-AutoDev (or FilmDev)-{Contrast/HDR/Sharp/SVDecon}-Color-{Shrink/Filter/Entropy/SuperStr/NBAccent/}-Track/NR(Unified-Denoise)-{Flux/Repair/Heal/Layer/Synth}
Key: {...} optional modules

This is a way of using the module which should give good results in most cases:
Isolate preset - to push back noise/starfield.
  1. Select Isolate preset - this sets most of the controls to their optimum.
  2. Set Strength - to adjust the impact of the Super Structure module. Normally somewhere between 30-100% - use your judgement.
  3. Set Saturation - Change this to adjust the amount of 'glow'.
Other Presets - to add 'life' to the image - increased glow and a more 3-dimensional effect.
  1. Select the preset depending on the type and amount of effect is needed.
  2. Set Strength - to adjust the impact of the Super Structure module. Normally somewhere between 50-100% but can be less - use your judgement.
  3. Set Saturation - Change this to adjust the amount of 'glow'.
  4. When you are finished 'Keep' the result.
What result to look for:
  • Isolate Preset - objects should be lifted from the noise and background stars. Watch for halos around bright stars.
  • Other presets - Images should look less flat and detail looks more 3-dimensional.
After Use:
  • Often followed by turning Tracking Off & Denoise module.
Special Techniques:

Push Back noise using the Isolate preset with a mask.
This is described in the video 'StarTools: Lifting structures from heavy noise with the Life module' from the start to 5m27s. Note that using a manually created mask can affect the 'documentary' value of the image.
  1. Open the Super Structure module.
  2. Create a mask which is initially clear but which you put green in the areas of detail you want to exclude.
    • for long thin wisps (e.g. veil) - create a trail of individual green pixels (Using Single Pixel Toggle Brush Mode) then use Grow until they join up and cover the whole area of detail.
  3. Press 'Isolate' preset.
  4. Increase 'Mask Fuzz' to 3-4.
  5. Press 'Do'.
  6. Press 'Keep' when happy with the result.
Description of Controls:

For general instructions on using mask see Mask.
  • A mask can be used to selectively 'add life'.
  • Use of the mask will compromise the 'documentary' value of the final image. If this is of concern only use a mask derived purely from the image (e.g. Auto Star Mask with optional Invert)
  • The mask can be used in conjunction with the 'Isolate' preset to select the object(s) to be lifted.
  • DimSmall - Pushes back anything that isn't superstructure. Overall image brightness stays the same.
  • Brighten - Brightens the superstructure.
  • Isolate - Lifts objects from noisy data. Similar to DimSmall but overall image brightness is not compensated for.
  • AiryOnly - Sets the parameters so the output is the result of the Airy model only.
  • Saturate - Saturates the colours in the superstructure.
Sets an initial gamma correction before compositing.
  • Default is 0.5 (DimSmall,Brighten, Isolate, Saturate presets), 1.00 (AiryOnly). Range is 0.00 to 10.00.
Compositing Algorithm
This setting defines how the calculated diffraction glow is combined with the original image:
  • Screen - works like projecting two images on the same screen.
  • Power of Inverse - Combination based on Power of Inverted Pixels (PIP) function.
  • Multiply, Gamma Correct - multiplies foreground and background and then takes the square root.
  • Multiply, 2x Gamma Correct - as above but doubles the Gamma Correction.
  • Default is 'Multiply, Gamma Correct' (DimSmall, Isolate),'Power of Inverse' (Brighten), Screen (Saturate), None (AiryOnly).
Brightness, Color
Sets whether brightness, colour or both are processed.
  • Process Both
  • Only Brightness
  • Only Color
  • Default is Process Both (DimSmall, Brighten, AiryOnly presets), Only Brightness (Isolate), Only Color (Saturate).
Controls the overall strength of the effect.
  • Range is 0% to 500%.
  • Default is 100% (for DimSmall, Isolate, AiryOnly, Saturate presets), 75% (Brighten).
Sets the colour saturation level of the extracted diffracted light (glow) before it is combined with the original.
  • Range is 0% to 500%.
  • Default is 50% (DimSmall), 100% (Brighten,Isolate,AiryOnly) 200% (Saturate) .
  • Change this to adjust the amount of 'glow'.
Brightness Retention
Sets how the brightness of the output is controlled and adjusted relative to the input.
  • Off - No adjustment
  • Local Median - corrects the local brightness by adjusting so the median value is preserved.
  • Global Mode Align, Darken Only - Does a gamma correction which makes the histogram peak of the output the same as the input. It then selects only the darker of the input and result pixels for output.
  • Default is Local Median (DimSmall), Off (all other presets)
Detail Preservation
Sets the way that the detail in the parts of the image to be brightened is preserved.
  • Off - does not preserve any detail. You can use this temporarily with the Compositing Algorithm set to 'Multiply, Gamma Correct' to see the areas targeted by the module.
  • Min Distance to 1/2 Unity - uses the pixel that is closest to mid grey (half unity).
  • Linear Brightness Mask - uses a brightness mask that progressively masks out brighter values and uses the original values instead.
  • Linear Brightness Mask Darken - uses a brightness mask that progressively masks out brighter values. Only pixels that are darker than the original image are kept.
  • Default is Linear Brightness Mask (for DimSmall preset) 'Min Distance to 1/2 Unity' (Brighten), Linear Brightness Mask (Isolate,Saturate), Off (AiryOnly).
Detail Preservation Radius
Sets a filter radius that is used for smoothly blending processed and non-processed pixels, according to the algorithm specified.
  • Only applies when Detail Preservation is set to 'Min Distance to 1/2 Unity'.
  • Default is 20.0 pixels (Brighten preset). Range is 1.0 pixel to 50.0 pixels.
Airy Disk Radius
Sets the radius of the Airy disk point spread function (PSF) that is used to diffract the light.
  • Match the radius close to the radius of the stellar profiles. Larger values if stars are large.
  • With large fields a smaller value is likely to work best - say 2-5%. A wrong value will make the image look fuzzier.
  • With small fields a larger value is likely to work best. A wrong value will make the structure less well defined.
  • There are subtle rings around the stars after processing. The way this looks can be adjusted using this setting.
  • The effect of this parameter is quite subtle.
  • It allows the module to account for how light was diffracted in the image so as to reduce unwanted effects in the small scale detail whilst controlling the larger structures.
  • This value is used to replace a number of different values which combine to define the radius of the Airy Disk. This simplifies the control of the diffraction effects and allows you to easily change the effect of diffraction on the results. It is not a percentage of any one thing specifically.
  • Default is 50%. Range is 0% to 100%.
Mask Fuzz
When using a mask, this sets the smoothness of the transition between masked and non-masked elements.
  • Default is 1.0 pixel. Range is 1.0 to 40.0 pixels.
  • Increase to make the transition smoother - values of 10 pixels are common.
Background Notes:

Adding Life - Modelling Light Diffraction
The Super Structure module tries to compensate for the fact that some sorts of processing can cause bright objects to lose their 'glow'.
It assumes the light went through a perfect lens with a circular aperture and produced an Airy Disk.
It tries to model the original Airy Disk by taking the energy represented by the sample pixels and redistributing that energy into an Airy disk.
This is done using a point spread function that produces an Airy Disk diffraction model. The calculated Airy Disk diffraction model is then added back into the original image.

Airy Disk
The Airy Disk diffraction model is defined by the Airy Disk Radius. This sets the radius of the point spread function and so dictates how the resultant Airy Disk diffraction model looks.

Adding the Airy Disk model back into the image.
The calculated Airy Disk diffraction model is added back into the image in two steps.
  • Creating a composite of the model and the original image
  • Selectively merging the composite with the original image.
The process is controlled by the following settings:
  • Compositing Algorithm - this defines how the contribution from a pixel in the Airy Disk model is combined with the contribution of the corresponding pixel in the original image (e.g. Screen, Power of Inverse, Multiply, Gamma Correct) to create a composite.
  • Detail Preservation - Defines how to decide between the detail in the image and the diffraction model composite, or the proportions of both. This is done by selecting the algorithm to use in merging the composite of the model and the original image back into the image.
    • Off - None of the detail is preserved - the area to be redistributed is completely replaced by the new Airy Disk.
    • Linear Brightness Mask - uses a brightness mask with a linear relationship between input and masking - and with the composite of Airy Disk model and original as input - so, as the composite gets dimmer it masks out more of the composite and uses more of the original instead.
    • Max Contrast - chooses between the composite and image detail values based on which adds most contrast with its neighbouring pixels. Takes the Detail Preservation Radius as a parameter.
    • Min Distance to 1/2 Unity - chooses between the composite and the image detail values based on which value is closest to 1/2 unity (gray). Pixels closest to gray tend to show the most human detectable detail. Takes the Detail Preservation Radius as a parameter.
  • Detail Preservation Radius - used with binary either/or Detail Preservation algorithms to smooth the transition. As you move away from where a processed (composite) pixel is selected, a greater proportion of the relevant non-processed (original image) pixel is combined until there is no processed pixel contribution. This parameter defines the radius of that transition. Used in the Detail Preservation algorithms 'Min Distance to 1/2 Unity' and 'Max Contrast'.
Post Reply