This guide will take you through the process of transforming a video capture of Jupiter into an image you can show your friends. We'll be using two pieces of popular software, AutoStakkert!2 and Registax 6 to do so.
1. Stacking in AutoStakkert! 2 (AS!2)
Download and install AutoStakkert! 2 for free from here.
Running AS!2 will open the following windows:
The window on the left is the main program window where we will load the video to be stacked and perform most of the processing. The preview window on the right will display the video and allow you to preview each frame; it will also be used to set our alignment points later on in the process.
Click on the button labelled "1) Open", navigate to the capture video and open it.
The number of frames in the video and the video filename are now displayed in the bottom left of the main windows and our video is also displayed in the right hand window. Let's have a quick rundown on the options we need to go through before we can click '2) Analyse'.
This is where we select what type of object we are stacking.
Surface - Used for the Moon and Sun, where we are imaging craters and sunspots.
- Expand - This makes the image as large as possible with the side-effect that the edges may be distorted due to lack of data.
- Crop - This crops the image to an area around your selected alignment points which will result in sharper edges.
Planet - used when we stack images of planets like Jupiter that are a discreet object in the view, as in the screenshot above. It uses a method called Centre of Gravity (COG).
- Dynamic Background - This should always be checked as it affects the COG calculations
This sets how AS!2 calculates the quality of each frame in the video.
Gradient - This is used for large planets, or surface pictures.
Edge - Used for smaller planets such as Mars, Venus or Mercury.
Activating 'Edge' will then display a circle with tick boxes around it. When processing video of the inner planets during their phases, you should untick the boxes that correspond to the unlit edges as they appear in the window on the right.
- 'Noise Robust' - This should be set according to how noisy your video footage is. If you can see fine details you should set this to a lower number, if the footage is very noisy set this higher. It's recommended that you leave this on 3 until you gain more experience with stacking.
- 'Force Global Quality' - You should only need to use this if visible seams appear on your stacked images. This can happen with low quality video and multi-point alignment.
As we are going to process some footage of Jupiter you should choose Planet and tick Dynamic Background in the 'Image Stabilization' section. In the 'Quality Estimator' panel, choose Gradient, leave Noise Robust as 3 and ensure Force Global Quality is not ticked.
Analyse the video
Clicking the "2) Analyse" button will instruct AS!2 to examine each frame of the video and create a small stack of images that it will use as a reference frame to align all the other frames to.
The progress of this process will be displayed in the central portion of the main window and, once complete, it will display a graph that represents the quality of each frame as it was in the original video (grey line) and quality of each frame sorted by high to low quality (green line).
There are two options for analysing the data which can be found in the Reference Frame panel.
- 'Last Stack is Reference' - This will use the last stack of images you created and use that as the reference.
- 'Auto Size (Quality Based)' - This is the default setting and uses the best set of frames in the currently loaded video.
Setting Alignment Points
We now need to use the preview window, that is displaying our video footage, to place alignment points (APs). These points will be used by AS!2 to ensure that each frame of our video is aligned with every other frame, as mis-aligned frames will result in distortions in our final, stacked image.
Again, lets run through the options.
- 'Frames' slider - This allows you to scroll through all the available frames of the video, and will be sorted by descending quality. You can click on the 'Frames' text to switch it back to the frame order of the video. Another click will switch it back to quality order.
- 'Set Size' - This can be used to crop the size of the stacked images. If you find that the stacking process is taking a long time to complete, you can use this to reduce to effectively speed up the process. It can also be set before analysing the video in the previous step.
- 'Details' - This toggles the display of text describing the colour format of the frame, the frame number, the quality score of the frame and the brightness of the frame.
- 'Draw APs' - This toggles the display of alignment points on the image frame, which can be useful for viewing the detail in the frame when obscured by a large number of alignment points.
- 'Single' - This is the least accurate method of aligning the frames of your image. It's unlikely that you will ever need to use this.
- 'Multiple (MAP)' - This is the most accurate aligning method and will use either automatically or manually placed alignment points. Selecting this method will display a series of options in the area below it.
The area immediately below the alignment method selector displays the number of APs that are currently placed. You can also clear the currently placed APs by clicking the 'Clear' button.
- 'Manual Draw' - Ticking this allows you to manually draw the APs on the preview image. Left click to set the first corner of the AP box, move the mouse to increase the size of the box and then left click to fix the size. APs can be removed by right clicking them.
- 'AP Size' - This setting allows you to set the size of AP boxes when automatically or manually placing points by single left-clicking on the preview image. You can either use the four preset sizes (25, 50, 100, 200) or set a custom size by using the up/down arrows or your mouse wheel.
- 'Min Bright' - This allows you to set the minimum brightness of features that AS!2 will automatically place an alignment point at.
- 'Place APs in Grid' - This tells AS!2 to automatically place APs down on the preview image. It is generally not recommended to use this feature with planets, as it tends to place the APs too close to the planet's edge. However, I often use it to quickly place APs and then manually delete the ones I don't want.
For our Jupiter footage we need to do the following:
- Select 'Multiple(MAP)'.
- Ensure that 'Manual Draw' is unticked.
- Choose an alignment size of 40.
- Click 'Place APs in Grid'.
The preview screen should now look like this:
5. We should now remove the eight APs that are sitting on the edge of the planets by right clicking within the blue bounding box of each AP. Once you've done this the preview screen should look like this:
We could add some additional APs to cover the top left and top right areas of the Jupiter but this should be sufficient to start with.
Our alignment points are set and we just have one last thing to do ... generate the final stack.
We now move back to the main window and work through the options on the right hand side.
- Image Form - The first choice we need to make is whether to our final stack is going to be in TIF or PNG format. This will always be TIF as this is the format we need to load into Registax.
- 'Number of frames to stack' or 'Frame percentage' - This is where you can choose the number of frames that will be used to create the stack. You can either choose a specific number of frames or a percentage of frames. You should see a green tick mark that indicates which one is currently selected.
- 'Normalize stack' - This equalises the brightness of the frames that are being stacked.
- 'Sharpened images' - This will generate an additional image file that has been sharpened, it can be useful to generate this in order to check that there aren't any obvious stacking errors in the stack. You can also select a percentage of the raw stack to be blended into the sharpened image by using the up/down arrows.
- 'Save in folders' - Leaving this ticked will save the stacked image into a folder that will be located in the same folder that the video is in.
- 'Prefix' - This allows you to specify a prefix that will be added to the stacked image filename.
- 'HQ Refine' - For the best results this should always be left on. Only turn it off if your computer encounters performance issues.
- 'Drizzle' - This can be used to improve the output from under sampled images as well as increasing the size of your image. This isn't really needed as unless under sampled, this will degrade the quality of the image and isn't worth the extra processing time that is required to use it.
Let's set the final options:
1. Select TIF.
2. Click on the percentage box and enter a value of 10%. I have selected this value as it is roughly the percentage of frames that have a quality of over 75%. You can rough guess this or click on the graph at the point where the green line crosses the blue horizontal 75% mark. This will also select that frame in the preview window allowing us to see the frame number and % in the top left of the preview.
3. Select 'Normalize Stack' and choose 75%.
4. Leave 'Sharpened Images' un-ticked.
5. Leave 'Save in Folders' ticked.
6. Leave 'HQ Refine' ticked and make sure that 'Drizzle' is set to 'Off'.
7. Click '3) Stack'.
8. After a wait, the stacked image will be saved in a folder called something like: AS_p10_Multi (AutoStakkert_percentage10_Multipoint).
Time for Registax Wavelets.
2. Using Registax Wavelets
Download and install Registax for free from here.
Load the Stacked Image
When you open Registax you will be confronted by a multitude of tabs and options ... two thirds of which we can ignore as we did our aligning and stacking in AS!2.
All we are interested in is using Registax to process our stacked image. So click on the 'Select' button in the top left and browse to the folder where our stacked image has been saved. In the 'Open file(s)' window you will need to change the 'Files of Type' dropdown to 'Tiff frame(s)(*.tif)' as Registax expects us to be loading a video file by default.
Once you have loaded the stacked image, you will be prompted as to whether you want to stretch the intensity levels answer yes. Registax should then automatically take you to the 'Wavelet' tab and display a bright and blurry image of Jupiter.
Set the wavelets
There are a lot of options here but we are only interested in the 6 sliders on the left hand side of the window. These sliders govern how Registax processes the image; essentially the sliders will accentuate varying sizes of detail from large (1) to small (6). There are no hard and fast rules on what these sliders should be set to, and you will use different settings according to the object being processed and the conditions that the footage was taken in.
Each slider has 'Denoise' and 'Sharpen' settings, these modify the default behaviour of each slider, or how aggressive you want the modifications to be.
For this footage (taken with a TIS DMK21) you can try the following settings:
- Denoise 0.20, Sharpen 0.120, Slider set to 92.0
- Denoise 0.05, Sharpen 0.120, Slider set to 60.0
- Do not change
- Slider set to 8.5
- Slider set to 15
- Slider set to 50
This will result in this:
A lot of personal preference comes into this process, so don't worry about playing around with the sliders until you're pleased with the results.
One tip I'd give, however, is try and favour underdoing your wavelet tweaks rather than overdoing them. Aggressive use of the sliders can result in rather artificial looking detail. I'd suggest doing a few tweaks, going away for a few minutes, and coming back to it with a fresh pair of eyes to see if the detail looks natural before proceeding.
Once you're happy with the results, click on 'Do All' followed by 'Save Image' and you're done!
One option that's useful when processing colour images is the 'RGB Align' button found on the right hand panel. This option will adjust the red, green and blue channels of the image, correcting any imbalances caused by the sensitivity of the webcam to particular light frequencies. Clicking the button will open an additional window and display a green square over the image. Resize the square so that your object sits inside it and then click 'Estimate'. You can then save the image as normal.