Jupiter 30/12/13 and reprocess from 17/12/13.

[stev74]

Hi all,

There were a few of us out on the 30th of December but unfortunately the seeing was well below average and for me high cloud kept coming over in the middle of some of the AVI's until it eventually clouded over after a hour or so. I think I must of got lucky with my first run as it was miles better than anything else I got - I've included an animation as well, more to show what Io and Europa were doing over an hour and to show how the seeing conditions deteriorated than anything else.

Jupiter002 13-12-30 23-50-00 with Io and Europa by stev74, on Flickr

https://www.youtube.com/watch?v=-jDSMpJ4b6w

After taking some advice from Neil I've had a go at reprocessing my images from the 17th December which were originally lets say on the bold side of processing!! The 17th were the best seeing conditions I've managed to image in so far this apparition so it was good to have another go at processing the data.

These are all 10000 frame avi's with the best 5000 stacked in AS!2 including a 1.5x Drizzle, then wavelets in R6 one with Dyadic (wavelet 2 only) and one using Linear (wavelets 1,2 and 3) with both full sized and also then reduced to 75%. I've also included a Winjupos version which is 3 separate images (Jupiter 0005/6/7 in the run sequence) combined using 'derotation of images'. All finished with a final tweak in Image Analyzer. 

Any feedback on these would be great to see if I'm getting wavelets etc about right?

Again a link to an animation best viewed by clicking on the 'large player' icon on Youtube I found. 

First up the Winjupos version but all the info is in each title...

2013-12-17-0133 Jupiter0007 reprocess Winjupos 75% downsize.. by stev74, on Flickr

Jupiter0007 13-12-17 01-36-23 reprocess 75% downsize, dyadic wavelets.. by stev74, on Flickr

Jupiter0007 13-12-17 01-36-23 75% downsize, linear wavelets by stev74, on Flickr

Jupiter0007 13-12-17 01-36-23 reprocess 1.5x Drizzle, dyadic wavelets.. by stev74, on Flickr

Jupiter0007 13-12-17 01-36-23 reprocess, 1.5 drizzle.. by stev74, on Flickr

Finally the animation..

https://www.youtube.com/watch?v=myoWvUjJAd8

[Uplooker]

Hi Steve,

I am not best placed to comment on the wavelet specifically but I think you have done a good job on the processing in general. You have gotten some very good images, well done. I would be very happy with those.

I used to think people had a real hankering to moan about the conditions we get but I am slowly becoming one of them. The conditions have been dire this autumn / winter.

Ian

[neil phillips]

Hi Steve one thing that is difficult to get across. is how I do my full processing. So of course using dyadic before winjupos combine is only part of the sharpening I would do. so comparing a dyadic attempt. against a linear one. which clearly has more strength can be misleading. There is also the problem of focal length. My processes are geard for quite high focal length. anything from 5 to 11 meters. As such using Dyadic will give variable results depending on focal length. Seeing conditions ect. I was thinking about this recently. As my current blurry images were having rather different wavlets applied. much more dyadic 3 wavlet than I suggested to you. Mainly to pull out what little detail there was in my current rather poor captures. I think the best one can say is to experiment and find ones own way to getting the image to look how you want it. On occasion I have mentioned I have used a tad linear wave 1. mostly don't say it. because mostly I don't use it. But if I had certain types of images and the images were appearing as soft as what the dyadic version of your shot is showing. I would likely try some linear wavelet 1 in combination, as at your focal length it appears to me dyadic wave 2, is really not fine enough. wave 1 will produce a lot more noise though. which is one reason I avoid it but I have used it, it can be done especially with good noise removal. In your case it appears like linear is working better.  Having said that the extra wavlets are producing a finer stronger appearance which I prefer here. though it does bring out the artifacts more. The dyadic looks too soft. Because I do a lot more sharpening other than dyadic. its hard to judge a comparison saying linear is better. Here at your focal length with your sharpening routines after wavlets applied. linear is better. But not all circumstances will react the same. It seems me and Freddie are the only ones who like dyadic. At high focal length, drizzle 1.5  you would be surprised how fine dyadic can actually appear.

[Space Cowboy]

Nice repros Steve. The linear image gets my vote despite the limb artifacts which makes me ask were you using gamma at 100 during capture as I've had those artifacts when dropping gamma below default 100?

[neil phillips]

As a example of how every little thing can affect processing. The 100% dyadic faired a tad better. Probably because you was too soft on the dyadic version at 1.5x. But at 100% the dyadic wavelet would be stronger. which it looks like it needed in comparison to the strength you used on linear.

 Anyway I took your 100% Dyadic resized it. Did some further sharpening noise reduction. And manual artefact removal. downsized to roughly your linear shot. It no longer looks night and day in favour of linear. You cant compare Dyadic very easily to linear. In a sense of using the same wavelet strength. The comparison wont work. As dyadic is stronger. The most you can do is say. One looks better over another. And I have given the multiple of reasons why that would be so.

[neil phillips]

Btw I think your image looks better. But I still have the feeling that more fineness could be extracted. At the moment I am not sure how. Without actually spending ages working on the raws from scratch. But I think things are improving. That artefact is very sever though hence my crude attempt to remove it all

[neil phillips]

The more I look the better it is looking. Other than that artefact its going in the right direction isn't it Steve. I get those artefacts when I try to eek out too much contrast. But this looks different I am not sure ?

[U Cyg]

Nice repros and another vote for the linear processing! 

Angie

[stev74]

Nice repros Steve. The linear image gets my vote despite the limb artifacts which makes me ask were you using gamma at 100 during capture as I've had those artifacts when dropping gamma below default 100?

Thanks Stuart. The limb artifact is something that seems to appear to some degree in most if not all my images with the worse artifacts appearing with worse seeing and the strange thing is I never touch gamma it has always been set to 100 as I've seen plenty of threads on here about changing it. Maybe it's something to do with the way I use alignment boxes on AS!2 but I always use manual alignment boxes, size 100 and keep them away from the edges then use 50 on the inner area. All the captures are generally 60fps, 1/60 exposure with gain set using the histogram so it's 65 % to 85% full, on average probably 75% on most occasions. Maybe it's something I'm adding at some stage during capture or early processing but I've not got to the bottom of it yet but I'll keep trying. Any other ideas of the cause from yourself or anyone else would be helpful?

The more I look the better it is looking. Other than that artefact its going in the right direction isn't it Steve. I get those artefacts when I try to eek out too much contrast. But this looks different I am not sure ?

Neil thanks for all your input to this thread, the further info on dyadic wavelets and the reprocess - I should say though I wasn't trying to get two matching images or to do a direct comparison between linear and dyadic wavelets just trying out different processing routes. 

Your also right about different focal lengths, seeing conditions and I suppose the main difference between our set ups would be I'm using a colour DFK618 whereas you are using a mono camera so I wasn't expecting the wavelets to work exactly the same. For these images I was using my 4x Imagemate with the camera pulled an extra centimetre along the length of the camera 1.25 adapter so I'm imaging at just over 6 metres focal length (scope is 1500mm) which is at the lower end of what you are doing. 

Basically I was finding if I pushed dyadic  wavelet 2 (with default filter) any further the image was starting to bloat a bit but I think it was still worth posting up the results and experimenting with them in general as it's something I'd never really tried before. I must say though if you sit slightly further away from the screen (2/3 feet) the surface detail seems to pop out more on the dyadic images (maybe it's just me or my screen). 

Anyway it's always worth experimenting as much as possible with our images I'd say (as I know you do) infact most of the good capture practise and processing routines/techniques I've picked up are from yourself, Stuart, JamesF to name just a few of the great planetary imagers on here!! I'll attach the original stacked image as a PNG so if yourself or anyone else would like to have a play with the data then feel free. Oh and any info about how you took out some of the limb artifact in processing would be great. 

Nice repros and another vote for the linear processing! 

Angie

Thanks angie. 

Here's the original stacked image if anyone would like to have a play with it feel free (1.5 drizzle, 5000 stacked from 10000 frame run in AS!2)...

[neil phillips]

Nice repros and another vote for the linear processing! 

Angie

Hi Angie

Cleary Steve has under done the dyadic sharpening to even make a quick comparison like that meaningful in any way

. Hes under done the 150% A lot more than hes under done the 100%. As clearly the 100% is sharper than the 150% version.

But both 100% dyadic. and 150% dyadic are under sharpened. compared to the linear versions. I certainly don't understand why that has happened. As clearly a touch more dyadic wave 2 on the 100% would have made them look very comparable ?

 The 150% dyadic is so off. in a strength comparison. To be virtually meaningless. To try and compare.

 Either more dyadic. Or less linear was neeed to even be able to tell any meanigful differences. But this was not why I suggested Steve try. Every image is different. And sometimes some images at certain focal lengths will not always respond as we think. having said that

I can pretty much make dyadic look almost identical to linear.On most  images ( perhaps not all ) Think I showed this well to Stuart many moons ago if I could find the thread ?   The reason I use it, is,  I find it easier to control what I am doing. There isn't a a major advantage visually for me to be doing so. But if I am controlling my processing better doing it that way. Then I suppose there is a advantage in that

[neil phillips]

Thanks Stuart. The limb artifact is something that seems to appear to some degree in most if not all my images with the worse artifacts appearing with worse seeing and the strange thing is I never touch gamma it has always been set to 100 as I've seen plenty of threads on here about changing it. Maybe it's something to do with the way I use alignment boxes on AS!2 but I always use manual alignment boxes, size 100 and keep them away from the edges then use 50 on the inner area. All the captures are generally 60fps, 1/60 exposure with gain set using the histogram so it's 65 % to 85% full, on average probably 75% on most occasions. Maybe it's something I'm adding at some stage during capture or early processing but I've not got to the bottom of it yet but I'll keep trying. Any other ideas of the cause from yourself or anyone else would be helpful?

Neil thanks for all your input to this thread, the further info on dyadic wavelets and the reprocess - I should say though I wasn't trying to get two matching images or to do a direct comparison between linear and dyadic wavelets just trying out different processing routes. 

Your also right about different focal lengths, seeing conditions and I suppose the main difference between our set ups would be I'm using a colour DFK618 whereas you are using a mono camera so I wasn't expecting the wavelets to work exactly the same. For these images I was using my 4x Imagemate with the camera pulled an extra centimetre along the length of the camera 1.25 adapter so I'm imaging at just over 6 metres focal length (scope is 1500mm) which is at the lower end of what you are doing. 

Basically I was finding if I pushed dyadic (with default filter) wavelets any further the images was starting to bloat a bit but I think it was still worth posting up the results and experimenting with them in general as it's something I'd never really tried before. I must say though if you sit slightly further away from the screen (2/3 feet) the surface detail seems to pop out more on the dyadic images (maybe it's just me or my screen). 

Anyway it's always worth experimenting as much as possible with our images I'd say (as I know you do) infact most of the good capture practise and processing routines/techniques I've picked up are from yourself, Stuart, JamesF to name just a few of the great planetary imagers on here!! I'll attach the original stacked image as a PNG so if yourself or anyone else would like to have a play with the data then feel free. Oh and any info about how you took out some of the limb artifact in processing would be great. 

Thanks angie. 

Here's the original stacked image if anyone would like to have a play with it feel free...

Drizzle15_Jupiter0007 13-12-17 01-36-23_pipp_g3_b3_ap31.png

 Mono Images may well respond differently. I suspected the images may bloat a bit at lower focal length. As to the reason you under sharpened both compared to the linear makes a bit more sense now.

that normally is a focal length thing I have found. As higher focal lengths bloat less quickly. I will have a look at your image to see what I think.

[neil phillips]

Hi Again Steve. Cant sleep here,  so ive been playing with your png. Your right it isn't easy to get dyadic on this image to work well. I did use a combination of wave 1 and 2. Your image is actually very noise free. So it allowed wave 1 dyadic to be used with no Gaussian applied.

I tried to make the image look both similar to yours. ( to back up my claim dyadic can look very similar to linear ) But I also tried to process it in a way, where the artefacts were  more controlled.

I used a combination of sharpening that didn't highlight the artefact. ( certain things did I found ) what I couldnt control during sharpening I blured away. Some may not like this. But seeing as the artefact is false. I don't believe it looks any more false with less of a obvious artefact to me. ( see if you agree )

I also wanted to see if if could get more fine detail out. with less smear. I found it hard to be honest. But if you look closely especially on the north polar region. You can clearly see there is less smear. with more fine details in the pole. Again see if you agree. I tried a few approaches with your image. Each one crashed and burned lol. Untill I hit on this. I think it looks ok. I may be able to do better with further experimenting. I am not sure.

But your processing is quite good Steve. It wasn't easy to make the image look any finer than yours without bad noise. So your doing well, as I tried a few times. My image is a tad more noisy ( you don't get something for nothing in this game )

But the smear is much reduced. Especially on the poles as mentioned.

So although dyadic is renowned for being easy to overbloat. ( which it is )  it can work. Hope ive shown this. I tend to use it because its a lot easier. And I find I can control the noise quicker and easier with my images. Though I am always Experimenting. And think I should certainly start experimenting with linear again. To be certain. ( you've got me wondering now )  I am not losing anything with my images. I don't think I am. But experimenting. only costs time. And I have quite a bit. Just feel sometimes its more constructive to get away from the voting system. We all do when someone suggests something new. or goes against the grain. Its easy to jump to conclusions. About a process. When often theres merit in both sides. I hope you agree. And have found this interesting

[sologuitarist61]

What great images - well done!

[Space Cowboy]

Neil has worked his magic on those!

Do you get those artefacts at 30fps Steve? I was never happy with my images at 60fps with the DFK.

[neil phillips]

Neil has worked his magic on those!

Do you get those artefacts at 30fps Steve? I was never happy with my images at 60fps with the DFK.

You have jogged my memory I do remember artifacts at 60 fps. infact it was a known issue if I remember correctly. Stuart is right. 30 fps should cure the problem.

Good thinking Stuart. I forgot about that. its been a while, I know a lot of imaging source users were annoyed by this. This was with the old chip mind. I didn't try with a 618 ccd colour ?

Does that behave the same way ?

[theo]

My old DFK618 was ok at 60fps, never had any artifacts that I can remember. I think this issue was confined to the older chipped version as you say Neil.

Pete

[stev74]

Thanks Neil for taking the time to work on my image, it has been a great learning experience for me and has given me a baseline to work with processing wise. With your image as a template to follow I had another go at processing using both Dyadic and Linear wavelets and while you have nailed the processing with both types of wavelets my Dyadic version still has a way to go compared to yours but I think my Linear version is now nearly as good. Mind you that is to be expected I suppose as you've been using Dyadic wavelets for much longer so it's something I will definitely continue to experiment with especially as I've managed to pick up a secondhand 5x Powermate so next time I'll be able to image at a higher focal length. 

Thanks also for your input into the artifact issue, I've made use of the blur tool in gimp in this final image which is something I hadn't considered before. Also thanks to Stuart and Pete for your input as well, I'll have a go at 30fps and see if that helps and also have a go at processing in R5.1 instead of AS!2 to see if it's the camera or my processing and I can take it from there. 

Anyways here's my final process for now.....

[neil phillips]

Really like that  Linear version Steve. Much better than your original posts. Working on your image I can tell you for what ever reason. Being a single shot colour camera. Focal length ( I didn't realize you was at 6 meters ) So focal length seems unlikely. Dyadic really needs extra sharpening meaning after wavelets. I also had to use wavelet 1 and 2 because of the softness that you can clearly see in your process. I must admit. seeing how different images behave with these settings. Opened my eyes too. And certainly has made me want to experiment with linear in my own images a lot more. Though clearly I have made it generally work for my own images. that's not to say I may not prefer another process, that I had given up on as more difficult. If I was you Steve. seeing how your images behaved. ( your not the only one Stuart doesn't like it either ) I would stick to linear. When I first mentioned try it. I think I half expect it to work similar to my own captures. and working on your image. I can tell you clearly it did not. So that was something that took me by surprise a bit. Your images are very noise free. where mine can on occasion be quite noisy. dyadic being softer can suit that better. working with the image rather than against it. I think your images work against the technique . Even though in my example I could make it work. Being a learning thing, for you and me, was really the reason I didn't see a voting avalanche in favour of linear, being constructive in anyway. It may have given a false impression that the technique itself was unworkable. which your example seemed to suggest. Which I hope you can see why I would take issue with that type of voting response. Yes your linear was better. But I was more interested in why. Rather than some members possibly coming away with the belief the technique is a waste of time. Its difficult, if it isn't part of a whole host of other sharpening routines to get right. and may work better with particular captures. But hope you can see, that's far from being unusable. I was worried the comparisons being shown seemed to suggest that strongly. With the early voting starting to re inforce that opinion. This was the reason I showed dyadic could work even on your image, even if I had to change the procedure quite drastically to achieve it. Sometimes all is not what it seems Steve. BTW the linear dyadic post I made. The linear was your version. I like the blur tool in Gimp. some do not. you can do similar using clone stamp ( ive heard ) photoshop registax has de ringing. But I never got that to work. Layer masks are also another way. Though I haven't tried that. at some point I would iike too.

Nice process Steve. You had a good night. But I like to be positive and think you will do better. Always look over the hill Is what I say