StarTools

Here are some notes relating to using the Stereo 3D module.
They relate to StarTools version 1.6.392beta and later
Please let me know if anyone sees any errors or has any additional advice they think helpful.
I will update this 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.

Stereo 3D Module

Purpose:
To synthesise depth information based on astronomical image feature characteristics to allow displaying 3D-like images in various ways.
This can be an important tool in interpreting the subject - making us understand its structure - and guiding us to process the data to bring out that structure.

Description
This module synthesises plausible depth information to allow images to appear three-dimensional.
Images can be created for viewing with: Red/Cyan glasses, VR Headset, Facebook 3D photo feature, or in standard web browser using the 2.5D option.
This depth information is not in the data but is suggested based on common rules and assumptions about gas and radiation behaviour.

Useful Sources
Stereo 3D: Plausible depth information synthesis for 3D-capable media and Virtual Reality
There are a number of useful links in StarTools Links and Tutorials.

The notes below relate to StarTools version 1.5, 1.6, and 1.7

When to use:

• When Tracking is Off.
• Standalone - on any image processed previously. No need to reprocess.
• At end of workflow.

Example Workflow (v1.6):
AutoDev-{Band/Lens}-Bin-Crop-Wipe-AutoDev (or Develop)-{Contrast/HDR/Sharp/Decon/Flux/Life}-Color-{Entropy/Filter}-Denoise (or Denoise 2)-{Layer/Shrink/Heal/Repair/Synth/Stereo 3D}
Key: {...} optional modules

Example Workflow (v1.7):
{Compose}-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/Stereo 3D}
Key: {...} optional modules

Method:
This is a way of using the module which should give good results in most cases:

1. Choose the Mode that is most comfortable for you to perceive depth.
2. For Side-By-Side 'Cross' or 'Parallel' Modes - Zoom out so you see both L and R images.
3. You can process using the easiest mode to view (e.g. Anaglyph with glasses) and then convert later.
4. Choose the way(s) of getting depth information best for this subject:
   • Dark Detail in front of light background - e.g. dust clouds - Luma to Volume - Shadow Dominant
   • Light Detail in front of dark background - e.g. M20 - Luma to Volume - Highlight Dominant
   • If there are still bright elements you want to bring forward - Simple L to Depth - Increase from 0%
5. Control the embedding of elements:
   • Large bright areas - e.g. emission or reflection nebulosity - Highlight Embedding - use higher values to embed
   • Bright objects within reflection nebulosity - e.g. bright stars within reflection nebula - Highlight Embedding - use higher values to embed
   • Small structures in front of large - Structure Embedding - use lower values.
6. Use 'Before'/'After' and 'PreTweak'/'PostTweak' button to see the effect of changes.
7. Set the Mode to 'Depth Map' to see the areas that have been brought forward (white).
8. Choose the final Mode required.
9. Export to: Web 2.5, Virtual Reality or Facebook 3D Photo as needed.
10. Press 'Keep' when done.

What result to look for:

• Image should give a sense of three-dimensional solidity.

Ways of getting better results:

• Often less is more. Overdoing it may spoil the illusion.
• For cross-eye or parallel viewing, portrait images are recommended.
• For Virtual Reality and anaglyph viewing, landscape images are recommended.

After Use:

• Enjoy.

Special Techniques:
<Contributions welcome!>

Description of Controls:

Exporting Presets:
In addition to the the type of image being produced which is set by the Mode control.
The image can be output to support different 3D representations using these buttons:

• Web 2.5D - Exports as an HTML file for viewing using a browser.
• WebVR - Exports as an HTML file for viewing using a Virtual Reality headset.
• Facebook - Exports as an JPEG file for viewing using the Facebook 3D Photo feature.

Simple L to Depth:
Does simplistic depth mapping - assumes lighter stuff is nearer than darker.

• Larger values mean brightness will be taken more as an indication of depth.
• 0% means that brightness is not taken as an indicator of depth.
• It is usually sufficient to leave this at 0% and rely on the Luma to Volume control.
• Default 0%. Range from 0% to 100%.

Mode:
This sets the output mode. Select based on the type of output needed for the way it is to be viewed.

• Side-by-Side Right/Left (Cross) - For Cross-eyed depth perception without aids. The left hand image is for the left eye, the right hand image is for the right eye.
• Side-by-Side Right/Left (Parallel) - For Parallel depth perception without aids. The right hand image is for the left eye, the left hand image is for the right eye.
• Anaglyph Mono (screen) - For use with red/blue filter glasses - Mono image that minimises ghosting on sRGB calibrated screens.
• Anaglyph Mono (print) - For use with red/blue filter glasses - Mono image suitable for printing.
• Anaglyph Color - For use with red/blue filter glasses - Keeps some color tones intact.
• Depth Map - Outputs the depth map. White is the near plane, black is the far plane.
• Default is 'Side-by-Side Right/Left (Cross)'.

Depth:
The 3D effect is obtained by modelling two planes. A near plane closer to the viewer and a far plane further from the viewer.
The Depth parameter defines the distance between the planes.

• The smaller the value the less pronounced will be the 3D effect.
• The larger the value the more pronounced will be the 3D effect.
• Default is 21 pixels. Range from 0 to 42 pixels.

Protrude:
The Protrude parameter defines how much of the scene appears to be in front of the image and how much behind.

• At 0% the entire scene will appear to be inside the screen or print.
• At 100% the entire scene will appear to be in front of the screen or print.
• Default 0%. Range from 0% to 100%.

Highlight Embedding:
This controls whether highlights appear embedded within the larger structures around them.

• Bright objects often cause emissions around them. It makes sense to embed these.
• The higher the value the more embedded the highlights are.
• Default 25%. Range from 0% to 100%.

Min Structure Size:
Controls the size of the smallest details that are acted on.

• Smaller values are best for wide fields with lots of small detail.
• Larger values are best for narrow fileds with many larger scale structures.
• Larger values may take longer.
• Default 1.6 pixels. Range from 1.0 to 5.0 pixels.

Structure Embedding:
This controls whether small-scale structures should appear embedded or intersect larger scale structures.

• Lower values mean more structures float in front of larger structures.
• Higher values mean more smaller structures intersect larger structures.
• Default 25%. Range from 0% to 100%.

Depth Non-Linearity:
This controls how matter is distributed across the depth.

• A value of 1.0 distributes matter evenly.
• Values greater than 1.0 progressively skew distribution towards the near plane.
• Values less than 1.0 progressively skew distribution towards the far plane.
• Default 1.00 Range from 0.00 to 5.00.

Luma to Volume:
This controls how brightness acts as a predictor for volumetric detail.

• Larger 'shadow dominant' values (slider to left) give larger dark structures more volume. Objects against a bright background.
• Larger 'highlight dominant' values (slider to right) give larger bright structures more volume. Objects against a dark background.
• Default 50% highlight dominant. Range from 100% shadow dominant to 100% highlight dominant.

Intricacy:
This controls what size detail is distributed in depth.

• Lower values show more depth changes for larger scale structures.
• Higher values show more depth changes for the finer detail.
• Default 50%. Range from 0% to 100%.

Background Notes:

The benefits of 3D representations.
One of the benefits of this module is to encourage new ways of looking at a subject. It is possible to make assumptions about the target which may be wrong - such as:

• Dark objects are in the background
• Light objects are in the foreground

This are assumptions we make from our experience of our own world - but this is based on reflection of a bright light source.
However with DSO's the following may be true:

• Sources of reflected light are less commonly visible than emission sources.
• Dark areas may be in the foreground - blocking the light coming from elements behind.
• Bright DSO cores are often embedded in, surrounded by, and shrouded by, bright gases.

Understanding how the object looks in 3D may help you make better decisions when processing:

• Highlighting local detail at the expense of the shroud of gas surrounding it - giving a false impression.
• The way you process dark areas may make a DSO with a dust cloud in front of it look like a DSO with a hole in it.

3D viewing without aids - Free Viewing
There is a useful web site with information about 3D images here: Basic 3D Viewing Terms
You need to train your brain to see using Freer viewing methods - see here for a general discussion: The logical approach to seeing 3D pictures
The main ways of seeing 3D images without aids are:

• Cross-viewing or The Cross-eyed Method - This link describes how to see images using Cross-viewing: The Single Finger Method
• Parallel-viewing or The Parallel Method - This link describes how to see images using Parallel-viewing: Parallel-viewing Stereo Photographs

Ways of identifying depth
The Stereo 3D module gives us a number of ways of identifying depth based on common rules and assumptions about gas and radiation behaviour:

• Simple L to Depth - Assumes the lower the luminance the further away it is.
• Luma to Volume - Shadow dominant e.g. Pillars of Creation , Highligh dominant e.g. M20.
• Highlight embedding - bright stars with emissions gas around them - embed.
• Structure embedding - small structures embedded in larger ones.

Types of 3D images

• Anaglyph - Creates a 3D image by Layering of different colours to create a stereoscopic image. Requires color filter glasses to view.
• Lenticular - Places thin lenses over the image to restrict the view of each eye to a particular part of the image.

Depth Map
The Depth Map tells how far away from the viewer each pixel is.

• Can be used to convert a 2D image to a 3D image.
• Facebook and Photoshop can use a depth maps to create a 3D image.