WinJUPOS TUTORIAL
Preliminaries…with
(semi!) updated editing…
This tutorial covers WinJUPOS using the “De-rotation of images…” and
“De-rotation of R/G/B frames…” – “De-rotation
of avi’s…” is not covered here. You can take 3 long mono avi’s (I’d
suggest a maximum 6 minutes per channel Saturn, 3 minutes per channel Jupiter)
& process the image from each channel and combine them using the same method - there
is nothing basically wrong with this but I discourage it because of the
possibility of "seeing" consistencies" - this was originally
written for combining images from 3 (or
more) mono sets of “classic” r-g-b runs (eg, 6 mins total r-g-b for Saturn, 3
mins for Jupiter.) You can also use
this perfectly well for colour camera images, or you can combine
each “classic” set of mono rgb’s and combine these coloured outcomes in WinJupos: just remember to
click on “Colour” in the Image measurement Window whenever you are measuring
colour images.
It also helps using larger images if the data can handle
this, or before any scale reduction if you have employed drizzle in AS!3 or Registax – making the “Alignment Frame” easier to set
precisely for better outcomes! Alternatively utilise the "zoom"
feature in the Measurement "Adjust" window.
For this tutorial I used 3
sets of r-g-b's taken in reasonable seeing on one night.....it’s preferable imo for them to be taken one after the other so that the
total timespan from the start of the first set to the end of the last set isn't
overly long (some small delays can easily
be accommodated for between sets but I
recommend for best results to keep your WJ combining to sets where the
time-span between the start of the first avi &
end of the last is within 30 minutes of each other - that's just a suggestion
for optimum outcomes...)
You'll need to process your
avi channels individually for mono imaging in AS!3, Registax (for wavelet application) and AstraImage or other
processing regimen you favour, and save the outcomes as .png
or .tif files.
It's worth doing most of
your initial sharpening to any image (mono's or colour) before measuring &
combining in WinJupos as the combined images will have there noise lowered by
the "stacking" WJ does.....this means you
can "push" the processing further than you might normally do with an
image - but be sensible about all of this..! :)
I'll start in this updated
re-editing there is no difference
between combining mono images (which is the main thrust of the following
explanations) or combining colour images - regardless of whether you
have used a colour camera or combined your red & green & blue channels
from mono captures before doing any WinJupos combining..... this
makes it a whole lot easier & quicker for mono camera users than
the following single channel explanations but it is EXACTLY IDENTICAL IN BASIC METHODOLOGY!
I now never work with the mono images after basic
processing of them, always combining the 3 channels whose total timespan for
those 3 stays within accepted limits for each planet without usually resorting
to WinJUPOS ie, (roughly 3 mins
for Jupiter, 6 mins for Mars or Saturn or 10-15 mins for Uranus & Neptune.
Some folks find 4 mins ok for a Jovian capture with AS!3... the image scale you
capture at affects these "rough" figures! :)
Great news for everyone because working with the mono
images is a rather tedious exercise - "normal" combining of the
channels into a composite colour image first using AstraImage,
P/shop etc requires much less.
First find out the
mid-point time of each single channel’s processed file - or that of a colour
camera's avi (FireCapture's text files make this easy!) - make sure
this value is Universal Time, & also check that the day/date you enter is
also UT: the date can change over an imaging session if you are up half the night.....
Say you have 3 sets of rgb's.....this means you'll end up
with 3 processed .png or tif
file outcomes in each channel (or only 3 files if we're talking colour images)...ie,
starting with nine in all. Both Registax & AstraImage save .png or .tif .
Loading processed files
& aligning/measuring
Start by opening WinJUPOS
and click on "Program" > "celestial body" and I’ll choose
"Saturn” for this exercise...but I’ll also include some Mars screenshots
to show the similarities/differences between the 2 planets.
Use the help info of
WinJUPOS to acquaint you with what all the keyboard functions do.....read and re-read it is my suggestion - it took me
several reads to get the gist, mainly about moving the image and the frame!
Then click on
"Recording" and select "Image measurement" and another
(large) window will open on top of the original (it looks like WinJUPOS has a window
within a window.)
Up on the top left hand side
the "Imag." tab should be highlighted with
an orange/brown bar above it.....just below here is
the "Open image (F7)" tab...open your selected image.
Your planet image will
appear along with an “Alignment frame” - although the two will most probably be
quite out of kilter with each other as per the next image… Once you've opened
one of your images you have to enter the time (seconds as decimal points of a
minute - each 6 secs = 0.1 mins and every 3 secs or more counts as an extra 0.1
mins, ie 15 seconds = 0.3 mins) as well
as the geographical location of where you took the images (lat. & long.).....you
can save co-ordinates for any imaging site by using the black arrow button just
below the latitude window... 
Note the filename and path
are at the very bottom of the windows.
Once you've entered all the
correct figures & rechecked them,
click on the "Adj." tab next to the "Imag."
tab - this is where you set up the alignment frame and planet image accurately
so that WinJUPOS can accurately measure the image for its calculations for de-rotating...
This is a red filter/channel image…..you select this in the dropdown box titled “”Channel (F9)”
This is where
you should really have read the “Help” info in WinJUPOS re all the various
controls you can see here…..although a couple of them
are incorrect therein..!
Altering the “Brightness” or “Contrast” may help you “see” the planet within the Alignment frame better…but I find the "zoom" tool the best helper.
Most important
are your “Page Up” and “Page Down” keys for altering the size of the Alignment
frame…..the ordinary up & down arrow keys of the
keyboard for moving the Alignment frame up and down and the “Left” &
“Right” arrow keys for shifting the frame left & right….. “N” & “P”
keys rotate the Alignment frame independently of the planet…..”L”
& “R” rotate both the planet & Alignment frame…often the Alignment
Frame will be the opposite orientation to the planet you've loaded...use the N
& P keys to get it the same as your planet & then the L & R keys to
get them both the appropriate orientation (usually "N" up.)
IMPORTANT -
you might find the adjustments just mentioned are too severe & wish to make
smaller rotations of the AF (Alignment Frame) with the "N" &
"P" keys...or the left & right, up & down or Page Up or Down
keys - holding down the "shift" key whilst operating any of these
keys makes the adjustments much smaller/finer!
Jupiter uses a
“simpler” alignment frame than Saturn which isn’t necessarily easier to set up
(I find having an outline for the rings as well as the disk is a bonus on
Saturn…..but the same keyboard controls work on both
planets.
For Jupiter
that “control H” will align the frame and planet horizontally on the screen
(when they’re synched) whilst “F11” is an automatic alignment operation that
works well…..you need to be aware of North & South
with Jove but you should find it very similar to Saturn in nearly all respects
otherwise. J
IMPORTANT!!! -
planets must NOT be flipped vertically or horizontally which can happen during
capture if you don't set the parameters correctly - it is relatively easy to
see if this has happened by
checking in the "Ephemerides" section of WinJupos to
compare your image/s with the graphic simulation there...the disk shadow on the
rings for Saturn or the moons position for Jupiter will help tell you if you
need to flip your image...upside down is easy to see..! :) Sort this out before loading into WJ.
Using the above keys (except the “L” & “R”) I have positioned the Alignment frame around the planet – you will note that I have made the frame a little larger (“Page Up”) than the planet to assist me in determining if the planet is centred within the frame…..don’t be fooled with Saturn by looking at the disk – the planet is the most oblate (flattened) of all planets it is far better to look at how centred the Alignment frame is around the ring system to get accurate alignment – doing this and entering the correct times/timestamps ensures optimum outcomes!
In the following image I have shrunk the Alignment frame to fit snugly around Saturn's rings and then also enabled “control H” to get the pair of them aligned horizontally on the screen! Using the “Outline frame” button on the LHS is meant to align both the frame and planet but this doesn’t seem to adjust the pair of them accurately in my experience, and I’ve found that manual positioning/alignment as above gave me the best results…..but your mileage may vary as they say! J
At this point
it pays to use the arrow keys (once you’ve “snugged” the frame around the
planet rings using “Page Down”) to double-check if they are indeed accurately
aligned…..by moving the frame one shift up and one
shift down as well as one shift left & right we can easily see if the frame
is indeed centred around the planet when the frame is snugly-fitting…..looking
to see what gaps there are with said movements…..
Then it’s back to the “Imag.” Tab and click “Save (F2)” Your image will be saved as an .ims file…..make sure you go to the correct folder to save it!!!
Also note that
the WinJUPOS .ims filename
has the date & timestamp in it…..this is important because when you make a
combined image from all 3 .ims files in each filter -
or when combining colour images from a colour camera or if you have already
combined your mono images - it will apply a date & timestamp to this
combination, and this is what you must use when you combine each of the
compound images for your final RGB output….!
Note also that I have added (highlighted) “1r” to the front of the filename…..this means this is the first of the 3 red channel measured images I am doing (remember, in the following explanations the 3 sets of rgb’s are WinJUPOS processed by measuring and obtaining an .ims file for each single red, green & blue and then combining the 3 reds, the 3 greens & the 3 blues…..and then combining those final 3 “compound” red, green & blue images for the output rgb image..!) J
Now when you open the second red image you’ll notice that it is much closer to being accurately aligned than the first one we opened, and much easier to adjust - in fact there is usually no adjusting to do whatsoever if you have "set" your first image accurately..!
Don’t forget
to check that the date is correct for when we put in the UT details for this 2nd
red image…..refer to the text file in your capture
files…sometimes (depending upon what time you were imaging) the UT day date may
change between images in any one session..!
Then go to the “Adj.” tab and make your fine adjustments…..save and put the prefix “2r” or somesuch to define this file…..and do the same again for the last red channel image .
You must
repeat all this with each set of 3 colour channel images…..make
sure you enter “Green” or “Blue” for the “Channel (F9)” dropdown box when using
those colours…..or if you are combining colour images
enter "colour."
In the next screenshot you can see I’ve saved all my red images (1r….. 2r….. & 3r…..if you use colour images then 1rgb, 2rgb, 3rgb etc is good) - I am now about to start saving my first green image (1g2011-04-22-1217.7)......
Here’s the last of the blue set…..all ready for the next process which combines each single red, green & blue.ims into single red, green & blue compound images for the final rgb combination: note the file-naming protocol I’ve adopted as previously mentioned…
We’ll briefly interrupt the WinJUPOS processing of Saturn to show Mars
being aligned within the frame just to
demonstrate the (slight) differences between planets – all the operations are
the same nonetheless & Jupiter is identical to Mars..!
Here is a screenshot of Mars being “fitted up” J with the Alignment frame to show the difference between it and Saturn
(remember that you must select whichever planet you intend processing right at
the start when you open WinJUPOS as per the first image in this tutorial.
In this Mars you can see by the examples that you must carefully consider
whether you have accurate alignment between the frame and the planet…..the
first Mars image appears to be correctly aligned within the frame…
But by pressing the “Page Up” key a few times and enlarging the Alignment frame we can see that Mars isn’t properly centred within this frame - you can see the "gap" on the LHS is greater than that on the RHS..!
Another
example of misalignment of the Alignment Frame not shown here atm is when the
AF needs to be rotated with respect to the planet - neither the planets or the
AF's are perfectly round & if this aspect is out of kilter you will notice
it by careful examination - correct it by rotating the AF for a nice uniform
fit around the planet & re-hit "Control H" to get both
planet & AF properly upright.
Here Mars is correctly aligned within the frame after using the arrow keys to shift the frame relative to Mars…and/or the N & P keys to re-orientate the AF...always hitting "control H" after any corrections to make sure both are in the correct orientation in the WJ window...
If we now use the “Page Up” key to enlarge the Alignment frame we can see that Mars is much better centred in it..!
After this little “Mars interlude” we
now return to the WinJUPOS processing of Saturn where
we had left off… J
We now click on “Tools > De-rotation of images” in the menu bar to bring up the next screen…
In the “De-rotation of images” window you won’t need to remove any files if this is the first time you’re using WinJUPOS - but after the 1st time you’ll need to click on the “Edit > Remove all” button to remove any files in the window…..a pop-up window will ask you if you want to remove the files and naturally you’ll click “Yes” if needed….!
Now click on “Edit > Add” and choose your first red channel image…..repeat for the other 2 red .ims images…..the following screenshot shows me about to load the 3rd red .ims file. (this is why you’ll appreciate it’s easier to call them “1r, 2r etc!)
There isn’t
any reason why you couldn’t have started with (say) 4 sets of rgb’s and be now combining 4 reds instead of 3 – WinJUPOS
allows for more than 3 files to be opened herein..!
Check that the
“Destin directory” is correct (ie, where you want your images to go – back with
the others!!!) and also note that when you hit the button “Compile image (F12)”
the resultant image will have its own timestamp and be a .png
file.
A new window comes up with the image you’ve just created (a combination of all your red images in the example shown) and it will be saved in the relevant folder/directory…..it will have its own timestamp that can identify it BUT you should really add a prefix to this filename to make it more readily identified – I chose “3rs” as per the screenshot below - if it were colour images I was working with it would be "3rgbs"…..I do this immediately after “Compile image F12” – just by bringing up the window and using “Date Modified” to get this last image up the top. You have to do it this way because WinJUPOS won’t allow you to alter the filename beforehand…
Close the “Image” window displaying the planet and repeat
these preceding steps for the 3 greens and the 3 blues…..here’s
the final compound blue image being saved with the prefix “3bs”
***NOTE - If you have been using
colour images you have now completed the job!!! Good luck!!! :)
Now close this last “Image” window of the blue compilation and the “De-rotation of images” window.
Okay – now we have to start all over again by measuring these compound .png files we’ve just created so that we can combine them as an RGB image…..meaning we turn these now into .ims files for WinJUPOS.
In the next screenshot you can see me opening the 3rs .png file and I have already entered the correct UT date & time taken from the filename I am opening (for the convenience of this tute!)
This timestamp
is what WinJUPOS has calculated as the midpoint of the 3 images we’ve just
combined for each channel.
You’ll notice that it’s “back to the start” with the Alignment frame & planet image also…..but you’ve already had a bit of practice….! J Just treat this compound image exactly in the same manner you set up each single image in the “Adj.” window – making sure that you have also set the correct “Channel (F9)” colour in there..!
We’ve aligned the planet & frame for WinJUPOS and gone back to the “Imag.” tab view to save this .ims file.
I have given the file its own distinctive prefix so that it is clear that this is the .ims compound red image…..there are similarities with some of the .ims & .png filenames and this makes it clear to me what this file is…..you may have a better system..! J
Repeat this for the green & blue sets also…..make sure that you
change the filter “Channel (F9)” colour in the “Adj.” tab view to the correct
colour each time as well as using the time given on the timestamp of the .png file!
Next, click on “Tools > De-rotation of R/G/B frames…”
This screen will appear and you open each of the compound .ims files you have just created…..check to make sure you’re directing the end-image to the appropriate “Destin directory” folder/directory..!
You can of course add another image (red for example) as a Luminance channel but I’m not overly impressed with WinJUPOS on that score – I prefer to do said in CS4 etc…..but you may like to experiment here to make your own judgement! J
Hokey-dokes..! You’ve just hit the “Compile image (F12)” button and Voila! – if you’ve set up each image in its frame and used all the correct times etc, you should have a good RGB image…..meaning there is virtually no colour misalignments other than the very slight limb colours here: if you’ve got distinctive “rgb rainbows” on the sides you haven’t properly followed these instructions….! J
The next step is really the “post-processing” but you will find with a good image that you can work it quite a bit: this image has 18000 frames - so personally I find that working it a bit harder than a “classic” 6 minute rgb set allows you to increase detail, smoothness and scale…..this image hasn’t turned out as good as my best Saturn WinJUPOS effort so far (the image at the start of this tutorial on 27th April 2011) but I am happy with it: seeing is always “King” and WinJUPOS won’t turn sows’ ears into silk purses, but I think that with reasonable/good data it is a very worthwhile processing tool to bring out the best in your captures..! J
I've done some small re-editing over time to this tute but there are always areas where improvements in it can be made: I had not covered adjusting the "LD" values in the final image compilation phase: “Tools > De-rotation of R/G/B frames…”
This is in the 2nd to last WinJupos screenshot above: you will note that on the RHS of each .ims file you have placed in each of the respective channels there is a gamma and LD adjustment box.....toggling the up or down arrows alters these values.
If your image (and I've noticed this particularly on Jupiter at times ) has some colouration along the limbs which can also appear as a "speckled" or semi-transparent broken line along parts of the limb (especially Saturn) then lowering the LD values will remove said.....you will have to experiment with the values you choose here this to fix this problem!
The reduction in LD values might need to be anywhere down to figures of 0.75 etc.....but in my experience the need to do this is always associated with incorrect/inaccurate Alignment Frame positioning & I'd recommend setting the AF such that these "limb artefacts" don't appear.....but having said this I note that many folks seem to not be so fussy & rely on altering the LD value to make their images look presentable.....although I personally feel the sharpness must suffer somewhat in taking this approach... :)