Hi all,
I have been giving some thought to the procedure for next year's project to produce a movie of the Jovian atmospheric movement. Here is my suggested procedure. Please come back to me with any suggestions to improve this or if you have any concerns about any of the practicalities:
1) Time Period for the imaging
The opposition of Jupiter is May 9th in 2018. I suggest we aim to do the imaging for 1.5 weeks each side of the opposition. Lets say 23 days in fact, because we can include as many weekends as possible in the timeframe. I am assuming it is easier for people to image during the weekends so the greater the proportion of the time that is on the weekend the better.
Can I suggest therefore we start on Saturday 28th April and finish on Sunday 20th May 2018? This gives 4 weekends (8 days) and 15 week days.
2) The Imaging Process
Clearly you will all have your own techniques, so the following would not be considered mandatory, but probably the more standardised the approach the better.
Telescope: 200mm + objective diameter. The larger the better
Imaging camera: Monochrome or OSC but capable of high frame rates 100+ fps
RGB filters: Where monochrome any good RGB set up. Luminance generally not required nor preferred
Focal Length Multiplier (Barlow or Powermate): Optimal Multiplier = 5.15 x camera pixel size (microns) x objective diameter (mm)/ telescope focal length (mm)
This is to ensure adequate sampling
Eyepiece projection may also be used to give good results.
Seeing: 4 or 5 out of 5
Transparency: minimum 6 out of 10
Captures: I suggest a standard of 60 secs per colour channel for monochrome and 90-180 seconds for OSC. The videos thus produced do not need de-rotation prior to stacking. If you want to take longer videos, then you will have to de-rotate the videos before stacking using WinJUPOS.
I use Firecapture to manage the capture process and it is an excellent piece of software. The ability to crop to a small Region Of Interest (ROI) around the planet allows much higher frame rates to be achieved. The higher the frame rate the more data to improve the signal to noise ratio.
Imaging run duration: As long as you can manage. The longer you capture the more of the planet you are seeing as it rotates and the better the planetary map we will be able to generate, especially if you encounter good seeing.
I generally aim for as high a gain as possible, zero gamma and set the capture rate to give a histogram level of ~ 80%
3) Image Processing - Quality Assessment and Stacking
I think we all use Autostakkert to quality assess and stack the videos. RegiStax is also a possibility, however, I have found file size an issue and it is less stable than Autostakkert. I also understand that there is consensus that Autostakkert produces superior results.
You will no doubt use your own settings, but i would expect you would use Planet (COG) for image Stabilisation. For the Quality Estimator I would use Gradient, Noise Robust: 4 (if the seeing is good enough this should be fine) and Local (AP). After doing the Quality Assessment, I generally set the frame percentage to stack around the point where the Quality Graph drops below 50%. I have looked at a variety of frame percentages to stack which is a balance between more data and less noise against including poor data. I have generally found that the 50% Quality on the graph corresponds well with the optimal frame percentage to stack. However, feel free to bracket if you want or use your own criteria.
When it comes to stacking I generally select an alignment point (AP) size to get 15 - 25 APs.
4) Image Processing - Wavelet Processing
In order to sharpen the stacked image there can be many techniques. I use RegiStax, but you may have a different approach. The exact process to sharpen the image is highly subjective. What I do is to select Linked Wavelets. The first slider I push to ~ 40, the second to about 15 and the third to ~ 5. Depending on the seeing this sharpens but introduces noise. I then use the noise reduction on each slider, most on the first, until the noise has been reduced to an acceptable level. The zoom feature is useful to see the noise. I then play around with sharpening and de-noising until I achieve an acceptable result. It is useful to save the settings to apply to each image if RegiStax crashes. I then look at the histogram. I bring in the slider from the right until it reaches the histogram and press "stretch". If necessary I move the central slider to marry up with the peak of the histogram. This just seems to give a good exposure for me.
5) Image Processing - De-rotation
I think we all use WinJUPOS for this. I don't there is anything in particular to standardise for the process, except image size. For OSC outputs, when you do de-rotation of images can I ask that you select a Quadratic Image Size of 600 pixels. For monochrome with RGB, please also select this size when de-rotating images for each of the RGB channels and again when you are doing the De-rotation of R/G/B frames. This will ensure that we are dealing with a consistent image size and minimise the image measurement work when combining the images from multiple imagers to create a planetary map for each day. 600 pixels gives the planet occupying ~ 80% of the frame. Please also select North UP, so the images all have the same orientation.
I generally like to de-rotate and compile the monochrome outputs from 6 RGB runs. This gives me significantly better resolution than from a single RGB run, depending on the seeing. The resultant RGB combined image which has a time stamp at around the centre of the 6 runs, can be sharpened again with RegiStax. I generally start with half the slider levels of the first round. That is for first slider I push to ~ 20, the second to about 8%and the third to ~ 3. Again depending on the seeing this sharpens but introduces noise and I use the same process as previously successively de-noising and sharpening until the result looks satisfactory. I then RGB Balance in RegiStax which always improves the colour of the final image.
6) Image Processing - Finishing
If there have been any moon shadows creating a black smudge, you may choose to use something like Photoshop to remove this from your final image (s). Equally if you've ended up with some unwanted processing artefacts please feel free to eliminate these in whatever manner you see fit. However, if you don't want to do this, you can just submit the images with the blemishes and I can clean them up if required.
Where you have done multiple imaging runs and therefore have images that cover more than 180 degrees of the planet, then please generate more images from across the run to include more of the planet. So for example, supposing you have been imaging for 2 hours. During this time the planet will have rotated 2/ 10 ~ 70 degrees, so your images will cover 180 + 70 =250 degrees of the planet's surface. You could arrange the number of runs that you compile in WinJUPOS so that you cover the surface with three good colour images. For example one image from 0 to 180 degrees , one from 35 to 215 degrees and the next from 70 to 250 degrees.
This way when combined with the output from other imagers we will have the best chance of getting coverage of the whole planet each 24 hours.
7) Image Submission
So your aim will be to produce a number of high quality colour images of Jupiter each day you can. Don't worry if you can't due to personal commitments, weather or seeing. I would expect that you only succeed a proportion of the time.
I will give you a DropBox link and you will be able to upload your images to there. I think that will be easier than uploading them to AstroBin, although you can do that as well of course.
8 ) Planetary Maps
Hopefully each day I will be able to produce a planetary map of Jupiter from our combined images. Over the 23 days we may have some gaps and I am hoping to access some image morphing software to produce an "intermediate" map for those days, such that in the end we can make a movie showing the evolution of Jupiter's atmosphere over nearly 3 weeks.
Please come back to me on this.
I have been giving some thought to the procedure for next year's project to produce a movie of the Jovian atmospheric movement. Here is my suggested procedure. Please come back to me with any suggestions to improve this or if you have any concerns about any of the practicalities:
1) Time Period for the imaging
The opposition of Jupiter is May 9th in 2018. I suggest we aim to do the imaging for 1.5 weeks each side of the opposition. Lets say 23 days in fact, because we can include as many weekends as possible in the timeframe. I am assuming it is easier for people to image during the weekends so the greater the proportion of the time that is on the weekend the better.
Can I suggest therefore we start on Saturday 28th April and finish on Sunday 20th May 2018? This gives 4 weekends (8 days) and 15 week days.
2) The Imaging Process
Clearly you will all have your own techniques, so the following would not be considered mandatory, but probably the more standardised the approach the better.
Telescope: 200mm + objective diameter. The larger the better
Imaging camera: Monochrome or OSC but capable of high frame rates 100+ fps
RGB filters: Where monochrome any good RGB set up. Luminance generally not required nor preferred
Focal Length Multiplier (Barlow or Powermate): Optimal Multiplier = 5.15 x camera pixel size (microns) x objective diameter (mm)/ telescope focal length (mm)
This is to ensure adequate sampling
Eyepiece projection may also be used to give good results.
Seeing: 4 or 5 out of 5
Transparency: minimum 6 out of 10
Captures: I suggest a standard of 60 secs per colour channel for monochrome and 90-180 seconds for OSC. The videos thus produced do not need de-rotation prior to stacking. If you want to take longer videos, then you will have to de-rotate the videos before stacking using WinJUPOS.
I use Firecapture to manage the capture process and it is an excellent piece of software. The ability to crop to a small Region Of Interest (ROI) around the planet allows much higher frame rates to be achieved. The higher the frame rate the more data to improve the signal to noise ratio.
Imaging run duration: As long as you can manage. The longer you capture the more of the planet you are seeing as it rotates and the better the planetary map we will be able to generate, especially if you encounter good seeing.
I generally aim for as high a gain as possible, zero gamma and set the capture rate to give a histogram level of ~ 80%
3) Image Processing - Quality Assessment and Stacking
I think we all use Autostakkert to quality assess and stack the videos. RegiStax is also a possibility, however, I have found file size an issue and it is less stable than Autostakkert. I also understand that there is consensus that Autostakkert produces superior results.
You will no doubt use your own settings, but i would expect you would use Planet (COG) for image Stabilisation. For the Quality Estimator I would use Gradient, Noise Robust: 4 (if the seeing is good enough this should be fine) and Local (AP). After doing the Quality Assessment, I generally set the frame percentage to stack around the point where the Quality Graph drops below 50%. I have looked at a variety of frame percentages to stack which is a balance between more data and less noise against including poor data. I have generally found that the 50% Quality on the graph corresponds well with the optimal frame percentage to stack. However, feel free to bracket if you want or use your own criteria.
When it comes to stacking I generally select an alignment point (AP) size to get 15 - 25 APs.
4) Image Processing - Wavelet Processing
In order to sharpen the stacked image there can be many techniques. I use RegiStax, but you may have a different approach. The exact process to sharpen the image is highly subjective. What I do is to select Linked Wavelets. The first slider I push to ~ 40, the second to about 15 and the third to ~ 5. Depending on the seeing this sharpens but introduces noise. I then use the noise reduction on each slider, most on the first, until the noise has been reduced to an acceptable level. The zoom feature is useful to see the noise. I then play around with sharpening and de-noising until I achieve an acceptable result. It is useful to save the settings to apply to each image if RegiStax crashes. I then look at the histogram. I bring in the slider from the right until it reaches the histogram and press "stretch". If necessary I move the central slider to marry up with the peak of the histogram. This just seems to give a good exposure for me.
5) Image Processing - De-rotation
I think we all use WinJUPOS for this. I don't there is anything in particular to standardise for the process, except image size. For OSC outputs, when you do de-rotation of images can I ask that you select a Quadratic Image Size of 600 pixels. For monochrome with RGB, please also select this size when de-rotating images for each of the RGB channels and again when you are doing the De-rotation of R/G/B frames. This will ensure that we are dealing with a consistent image size and minimise the image measurement work when combining the images from multiple imagers to create a planetary map for each day. 600 pixels gives the planet occupying ~ 80% of the frame. Please also select North UP, so the images all have the same orientation.
I generally like to de-rotate and compile the monochrome outputs from 6 RGB runs. This gives me significantly better resolution than from a single RGB run, depending on the seeing. The resultant RGB combined image which has a time stamp at around the centre of the 6 runs, can be sharpened again with RegiStax. I generally start with half the slider levels of the first round. That is for first slider I push to ~ 20, the second to about 8%and the third to ~ 3. Again depending on the seeing this sharpens but introduces noise and I use the same process as previously successively de-noising and sharpening until the result looks satisfactory. I then RGB Balance in RegiStax which always improves the colour of the final image.
6) Image Processing - Finishing
If there have been any moon shadows creating a black smudge, you may choose to use something like Photoshop to remove this from your final image (s). Equally if you've ended up with some unwanted processing artefacts please feel free to eliminate these in whatever manner you see fit. However, if you don't want to do this, you can just submit the images with the blemishes and I can clean them up if required.
Where you have done multiple imaging runs and therefore have images that cover more than 180 degrees of the planet, then please generate more images from across the run to include more of the planet. So for example, supposing you have been imaging for 2 hours. During this time the planet will have rotated 2/ 10 ~ 70 degrees, so your images will cover 180 + 70 =250 degrees of the planet's surface. You could arrange the number of runs that you compile in WinJUPOS so that you cover the surface with three good colour images. For example one image from 0 to 180 degrees , one from 35 to 215 degrees and the next from 70 to 250 degrees.
This way when combined with the output from other imagers we will have the best chance of getting coverage of the whole planet each 24 hours.
7) Image Submission
So your aim will be to produce a number of high quality colour images of Jupiter each day you can. Don't worry if you can't due to personal commitments, weather or seeing. I would expect that you only succeed a proportion of the time.
I will give you a DropBox link and you will be able to upload your images to there. I think that will be easier than uploading them to AstroBin, although you can do that as well of course.
8 ) Planetary Maps
Hopefully each day I will be able to produce a planetary map of Jupiter from our combined images. Over the 23 days we may have some gaps and I am hoping to access some image morphing software to produce an "intermediate" map for those days, such that in the end we can make a movie showing the evolution of Jupiter's atmosphere over nearly 3 weeks.
Please come back to me on this.
