How I process Solar prominence frames in RegiStax 5.

   (by Eric Roel)


Since RegiStax appeared in 2001, we imagers started to use this incredible software. At that time, most of the now known as “High Resolution” imagers, started taking video streams with all sort of camcorders and webcams like the famous workhorse Philips ToUcam. Soon we all were using the first versions of Corīs RegiStax, exploiting their ever growing features. I must say that without the webcams and without programs to align and stack the frames, we would not be where we amateurs are standing today. At that time many amateurs were producing images that equaled and very soon surpassed all the images professional observatories had made since astrophotography started. Amateurs started to be leaders in those chores and even got close to the results obtained with the Hubble Space Telescope.

Circa 2001, I started to pester Cor Berreboets with questions, with suggestions, etc…, and soon I was invited by him to join a fraternity of just a few guys around the world as beta testers. Suddenly versions R2 and R3 became rapidly overwhelmed with famous version R4, multiple alignment points were born and a new revolution in imaging started. Amateurs were getting incredible high resolution images of the Moon, the Planets and the Sun never equaled before.

Then again after a long dormant period for us testers, Cor was again ready for a new version, a version that now has features that make it a more incredible tool.

I wanted to make this brief introduction because as one of the beta testers, I personally want to thank and congratulate Cor, for taking in account every suggestion or even every crazy idea from us, and then worked on them until fully developed and functional.

Finally, I also want to congratulate and thank all of my fellow beta testers for being such good friends, always trying to help and cooperate with ideas, even that we live in different continents, have different cultures, different languages, different time zones, to all I thanks for your efforts.

Some thoughts about the image picked to explain my workflow.

I did choose an avi stream of the Sun recorded the 8th of November 2008, because it was taken in medium to good seeing conditions. The instrument used was the old 6” f/12 Astro-Physics “Superplanetary” triplet refractor, working at (+) f/30 (5.1” ERF + 2XBarlow + Telecentric + Daystar ATM 0.45 A° H-alpha filter with a homemade tilter + DMK31). The camera resolution was set to 768x768 pixels and 2000 frames were taken at 15fps, starting at 10:04:25 am. CST and finishing 2ī13” later.

Solar prominence imaging is tricky due to several factors like, daylight low and high atmospheric turbulence, the subtle tendrils that conform the protuberance, the continuous change of the developing prominence, the very uneven dynamic background created by the Hydrogen monochromatic light, dispersion from the etalon itself as also from high atmospheric humidity, the mechanical tilt of the etalon that distorts the image, the high  gain settings needed for the camera to record dim prominences against an uneven background , all  that results in a difficult and noisy image.

I have tried to make my workflow as easy as possible and I must say, that it does not work always, you have to get the feeling for every avi stream taken, each is different and therefore it must be dealt individually, so for this example, I feel that the workflow shown is the one I like better for this particular avi stream.


From the 2000 frame 1.09 GB avi stream, I picked the best or let’s say, one of the best frames by moving the lower slider. In this case frame number 745 was chosen.


Figure 1.



I took a few moments to analyze the image I had in front before clicking on the align tab, by moving the lower slider slowly from frame 1 to frame 2000 several times, and watching for drift, for changes in the structure of the prominence, for the turbulence pattern, etc… In this case, I felt that the big prominence could be aligned and stacked with just 2 alignment points, by using an alignment box size 256.

In prominence imaging, I have found that it is better to use just as few alignment points as possible, so not to produce the dreaded seams and changes in intensity from adjoining alignment points that show when you have less or more frames stacked on different alignment sections.

This problem most of the time appears when using many aligning points on an object like a prominence, that has very low contrast and that is dancing all the time in front of a very light and uneven background. (On the moon, the planets or in solar disc surface feature imaging this is not a big problem, but it sure is in solar prominence imaging).

From the Alignment Options tab, I did set:

a) Alignment method to Multi,

b) Alignment box to 256.

c) Quality settings to Gradient2.

d) Lowest quality to 90.

e) Click on pre-blurring.

f) On tracking settings clicked to track object, processing area clicked to 1024 and checked all auto processing buttons, the features not mentioned were left as default.


Figure 2.

Placed the first alignment box on the lower part of the prominence and after that, the second box on top of the first one, trying to cover all the prominence and to anchor them on some feature that shows some contrast, if happy with it, clicked on the Align tab, if not, cleared the alignment points by clicking clear on the multialign panel at right, and did it again until satisfied. (Figure-2)


Figure 3

3- REFERENCE frame.

After all the frames were aligned (figure 3), clicked on LIMIT tab to start making the default 50 frame Reference Frame, that will become the sample for all the frames to be stacked further on the process.


Figure 4

Before the reference is ready to serve as the sample for all the frames that will be stacked and optimized (figure 6), it has to be sharpened a bit with the Wavelet tool, so generally I use it in Gaussian and linear mode and just use layer 1 to get enough resolution from the sample, in the case shown it is 65%.

figure6Figure 6


Figure 7.

After the Reference Frame process is finished, we then click on Continue to start the Optimizing process.


After you click on Continue, one has two choices, one is to click on the Optimize tab, and the other on the Optimize & Stack tab. Choose the Optimize tab, so to set these optimizing parameters:

a)     Set the optimize setting to fast.

b)     Opt for v5 mode by clicking on the button.

c)      Opt for absolute.

d)     Leaving the rest of the parameters as default.

Figure 9.

When optimizing is finished, you are ready to begin stacking the frames you decided to use for your final image.


After opening the Stacking tab we have to set the parameters that are needed for stacking our frames, in this case:

a)     Normalize intensity over frames.

b)     Stackgraph for further tweaking.

When looking at the Stackgraph Panel, you will find 2 sliders, the one on the top graph can be moved horizontally and the one on the lower graph vertically.


Figure 10

I did not move the sliders, because the graphs show a rather flat and not overly spiky result. Though I suggest that it is worth reading thoroughly the latest manual, to understand how these tools work.


Figure 11


Once the frames are stacked to the parameters set from the beginning, then you are ready to enhance the resultant image, so click on the WAVELET tab.



As soon as you press this tab, you will find the final stacked image enhanced to the value you choose for the Reference Frame. I normally use on this type of objects just Layer 1, so to get the finest details possible on the final stacked image.


Figure 12.

Analyzing the final stacked image, you find that it can handle a bit more sharpening without getting too noisy and also that it needs a stretch to normalize the image, so for that we have to use the Histogram tool to see how far we need to bring up the black point. When happy with the result you can just save your image in the format you wish, or tweak the final image a bit more by clicking on the FINAL tab.


Figure 13


Usually I donīt leave as the final image the one shown, because after recording the prominence avi stream, I also take an avi stream of just the solar surface (without showing the prominence), so later I can make a composite with the 2 resulting images, giving the overall visual appearance of Prominence plus the solar disc detail. This procedure uses other well known processing programs, to further tweak the resultant image by increasing sharpness, increasing contrast, reducing noise, removing gradients, etc…, programs that are not part of RegiStax5.

Nevertheless, I used the FINAL tab to give the RegiStax final image an antique tint, by using the saturation and hue sliders, so to keep out the harsh look of the bright solar disc, the dimmer prominence and the uneven background.


Figure 14


    Figure 15.