I have tried processing in Registax 6, but, while I think the stacked result is quite good, I can't seem to get a feel for using wavelets to actually improve the image.
Any suggested would be most appreciated.
I have posted a stacked, and two overprocessed images.
I have read some tutorials, such as those below, but, they seem to say 'do it' without quite saying what to look for, to improve the image.
Not bad for starting out. My suggestion is that the VX-3000 will only produce these kinds of images. When you haven't much data to work with in the first place it makes it hard to process for an improvement. I had more success with a dslr than cheap web cams.
or
A DMK or equivalent with a 2x powermate will produce what you are after.
Hi Philip, I think your problem is more with capturing some good data not with processing it. Too much magnification and bad seeing make it near impossible to get good video data. A cheap web cam can get good results on a steady night but looking at how wobbly the edges are on your images I think that the seeing was not very favorable for a good capture that night.
Here's a pic taken with an old Phillips toucam web cam.
Can you tell us a bit more about your telescope setup please? What were the atmospheric conditions like when you were taking the video files?
a few things i noticed that may help.
1/ Definitely overkill on the wavelets....i try to use as little as possible. On an image of this size, you will be limited as to the amount of processing you can do to improve a very small image to begin.
2/ The Red Green Blue RGB Alignment could do with some adjustment. You can see a red and blue halo around the edge of the image. On your tool panel on the right side of screen when you are adjusting wavelets, you will see an RGB adjust button. It allows you to move the red and blue elements of the image to re center them over the green. It may help to tidy up the image a bit.
3/ Center the object and rotate your camera so that the cloud bands are horizontal ....it always looks more pro that way.
4/Try using a 2x barlow. Bigger image on the ccd will help.
5/ One misty night on a mountain in France the great Thierry Legaut walked out of the fog and joined me for a hot chocolate....I had similar images to show him. His underlined point was, collimation, collimation, collimation.....and more hot chocolate.
6/ Try a better camera and find good control software. A good camera is actually very cheap these days.
7/ Keep trying....only time and patience pays off with this sort of thing.
With the video you've captured I think you can get plenty more just with processing. Here's what I suggest:
download PIPP and use that to split out the video into still BMP files centering the target in the frames. You can also try outputting say first 1,000 ordered by quality. This I call pre-aligning, it will align jupiter to account for drift and some seeing wobble.
Now throw these into registax and run with No Alignment and little or no quality rejection if you did that with PIPP already. once stacked on I think the last screen near top left is a button something like "Realign with processed stack". Click on this and it should take you back to the start and has marked alignment points, now run through with this until you have your stack again. I find even with horrid seeing wobbles Registax produces a very well aligned result.
Next the wavelets. These are confusing at first and still are to me, but you can think of them mathematically doubling each time from lowest (top of the list in registax) to highest layer value. First layer acts on the scales of 1 pixel which is both detail and noise typically in astrophotography. Next layer is double so acts on scales of 2 pixels, again detail and some noise. Next layer sizes of 4 pixels, then 8 then 16 etc. So the highest layer typically won't improve anything but will create those ringing artifacts you've got in your images. I think of wavelets as a contrast/sharpening operation so as I adjust a slider or value I am looking for contrast changes, but too much and you get processing artifacts like in a heavily over sharpened image. Each layer multiplies up through the ones above it. So starting with layer 1 first, If you know what a bell curve is, I adjust a little here and bump up noise reduction for it a tiny bit too (experiment and you'll see a lot change for small movements). Usually I don't expect to see much or any change in the image with this step, I'm hoping small features are being enhanced without enhancing the noise here. Next I adjust layer 2 a bit further and you should start seeing contrast changes keep a tiny noise reduction on here too. Layer 3 is about where I am doing the most adjusting, no noise reduction, Layer back down a bit with no noise reduction. So like a bell curve the peak is around layer 3 with values either side dropping off. As you adjust you are looking for two things, one is contrast changes which gives you details and ringing artifacts, these are in your processed images where a light and dark edge forms on contrast boundaries and these you do NOT want. So its a balancing act. Do lots of experimenting with the sliders and values boxes and push each value to see where the ringing artifacts appear badly and keep your values way below this. As I said the layers multiply upwards so if you dont back off enough the artifacts are there mathematically and will effect how far you can push each layer up. Also if you adjust sliders randomly or start with the highest layer first you'll be forever changing sliders and its a frustrating exercise. don't worry if you see little or no changes when adjust the first two wavelets layers (in any program) but DO be aware that noise mostly lives in layers 1 and 2 and detail in 2 and 3 so being aggressive can give you noise as well as signal.
None of these mathematical descriptions hold true i dont think, but in practice its an easy way to understand what I'm working on and what to look for and expect.
In PixInsight I still get about the same increase in detail in my images with wavelets as registax gives and what AutoStakert does too, so it seems like its a good approach and my data is limiting me now.
Once you get to layer 3 Jupiter and saturn will be sharpening up and detail coming through and maybe close moons are starting to brighten out of the surrounding blackness. Good luck and be patient with it. I never got seeing as great as in your video and am pretty sure there will be much you can reveal following the above steps.
Thanks for the replies. I will work through the suggestions and see how it goes. They sound damn good.
I suspect that patience will be required!
I tried this webcam mainly to see what it can do, and I am impressed with the result even as is.
I do have a DSLR and have taken some footage, but, not had the chance to process it. Due to my ignorance, I think the webcam footage is better than the DSLR footage that I have. I don't doubt that the DSLR will be better in the long run
I would love to get a 'proper' camera, but, that is going to have to wait a while.
Again, thank you for the replies, and all the extra work I now have to do!
Philip, some of the replies are ok but none i thought addressed your actual question for now. Buying more gear is poor advice for improving when you can improve without cost, makes my wonder how much they are missing out on with their gear.
After rechecking your pics and video they all exhibit the ringing i mistook for purely wavelet processing based. Not sure what else causes the same effect. If you have trouble focusing a bahtinov mask will greatly help, are you sure you didnt record using a compressed format?
Hopefully someone else can provide a useful comment on what causes the same artifact instead of assuming its your gear which i dont believe for a second.
