single image v 7 de-rotated images

[Pete Presland]

I thought it might interesting to see the 2 images i recently produced side by side for a comparison between 3 minutes of data v 21 minutes.

This is purely a comparison of what my relatively modest  processing skills are capable of, given the two sets of data. All 7 SER files were treated the same, AS3 for stacking, REG6 for colour balance and wavelets. 

The first image is the single image, the second is the 7 x 3 minute images de-rotated in WinJupos. Even in my hands the difference is well worth the effort.

I did have a few issues with WinJupos, it left an annoying artefact just off the right hand limb, there is also a slight "edge rind" on the left (which i can live with 🙂 )

I removed the more annoying artefact by adjusting the levels at the black end of the histogram which seems to have hidden it successfully.  I think i might have lost a little bit of the limb all round doing this though. Probably another re-process is on the cards, as the data is the best i have for a while so worth more experimentation.

I tried a few different alignment measurement processes in WinJupos, but the artefact seem to be there on all of them to some degree.  I made the outline on the measurement routine as good as possible.

I would  be very  interested to hear what others think to the de-rotated image in comparison to the more subtle view in the first image.

[si@nite]

 A big improvement Pete, well worth the effort a smoother more detailed image you have atained, i would tend to use w j to make good data better myself, imo if the data is poor the results can come out poor too with artifacts & anomalies!

A comparative i did some years ago now shows w j can bring out some extra detail with average to good data but we must not get carried away with what it can do ( Hope you dont mind me posting on your topic) if i can recall think it was 6 mins 3x rgb tiffs derotated!

 

 

 

Edited July 30, 2021 by si@nite

[neil phillips]

3 hours ago, Pete Presland said:

I thought it might interesting to see the 2 images i recently produced side by side for a comparison between 3 minutes of data v 21 minutes.

This is purely a comparison of what my relatively modest  processing skills are capable of, given the two sets of data. All 7 SER files were treated the same, AS3 for stacking, REG6 for colour balance and wavelets. 

The first image is the single image, the second is the 7 x 3 minute images de-rotated in WinJupos. Even in my hands the difference is well worth the effort.

I did have a few issues with WinJupos, it left an annoying artefact just off the right hand limb, there is also a slight "edge rind" on the left (which i can live with 🙂 )

I removed the more annoying artefact by adjusting the levels at the black end of the histogram which seems to have hidden it successfully.  I think i might have lost a little bit of the limb all round doing this though. Probably another re-process is on the cards, as the data is the best i have for a while so worth more experimentation.

I tried a few different alignment measurement processes in WinJupos, but the artefact seem to be there on all of them to some degree.  I made the outline on the measurement routine as good as possible.

I would  be very  interested to hear what others think to the de-rotated image in comparison to the more subtle view in the first image.

Yes WJ is very important. Trying video de rotation a lot myself lately. Try LD on winjupos reduced Pete see if it kills the edge. Not sure if its my monitor but lots of areas on your image look whited out. white clipped if you will. I dont see that in Simons images below. Btw the detail is superb 

Edited July 30, 2021 by neil phillips

[orion25]

Brilliant work, Pete. Definite enhancement of detail and a more natural look in the de-rotated image.

Regards, Reggie🚀

[CraigT82]

Nice work Pete, obvious improvements in the derotated image. As Neil says play around with the LD value to control the bright limb if it appears in the final 

[Pete Presland]

11 hours ago, neil phillips said:

Yes WJ is very important. Trying video de rotation a lot myself lately. Try LD on winjupos reduced Pete see if it kills the edge. Not sure if its my monitor but lots of areas on your image look whited out. white clipped if you will. I dont see that in Simons images below. Btw the detail is superb 

does look a bit clipped either side of the orange belt area in the middle. Simons images are excellent, i will definitely have another go with this data.

[Pete Presland]

11 hours ago, si@nite said:

 A big improvement Pete, well worth the effort a smoother more detailed image you have atained, i would tend to use w j to make good data better myself, imo if the data is poor the results can come out poor too with artifacts & anomalies!

A comparative i did some years ago now shows w j can bring out some extra detail with average to good data but we must not get carried away with what it can do ( Hope you dont mind me posting on your topic) if i can recall think it was 6 mins 3x rgb tiffs derotated!

 

 

 

Very happy for you to post in this thread, thanks for sharing your images, i will definitely try again and keep  a better balance to the image.

 

[Pete Presland]

Any suggestions for the IMS alignment procedure?

[neil phillips]

18 minutes ago, Pete Presland said:

Any suggestions for the IMS alignment procedure?

I  am probably not a lot of use. it was over 5 years ago i was doing this. But i do think the LD value being reduced to 50 if i remember sorted out those weird edges. Even if the the number is wrong that is what you want to play with for a clean edge.

Automatic outline seems do a reasonable job on jupiter. Though of course you can align it manually if you dont like the look of the auto outline.  What times are you putting in. there was talk many years ago of inputting the mid point time.  However recently i measured jupiter with two moons in the image. it definately lined up better inputting the start capture time.  With the mid point time. the moons were more off ? not sure of others experiance here.  But trying  to brush up myself on all this too. Your images combined here are worth the hassle of watching clipping Pete. As stated earlier its very good.  Excellent captures 

Edited July 31, 2021 by neil phillips

[CraigT82]

8 hours ago, neil phillips said:

I  am probably not a lot of use. it was over 5 years ago i was doing this. But i do think the LD value being reduced to 50 if i remember sorted out those weird edges. Even if the the number is wrong that is what you want to play with for a clean edge.

Automatic outline seems do a reasonable job on jupiter. Though of course you can align it manually if you dont like the look of the auto outline.  What times are you putting in. there was talk many years ago of inputting the mid point time.  However recently i measured jupiter with two moons in the image. it definately lined up better inputting the start capture time.  With the mid point time. the moons were more off ? not sure of others experiance here.  But trying  to brush up myself on all this too. Your images combined here are worth the hassle of watching clipping Pete. As stated earlier its very good.  Excellent captures 

I wonder if your compute clock is slightly out? Have you synced it recently? 

if not then I guess an alignment issue may arise in Autostakkert if the timestamp of the output tiff doesn't match the time of the reference frame it used for the stack? I.e. if AS decides to use a frame near the beginning as the reference frame and the outputted tiff has the midpoint of capture time that could make the moons not line up properly in WJ?

Might go and  do some testing on existing Jupiter data to see what happens.

[neil phillips]

1 hour ago, CraigT82 said:

I wonder if your compute clock is slightly out? Have you synced it recently? 

if not then I guess an alignment issue may arise in Autostakkert if the timestamp of the output tiff doesn't match the time of the reference frame it used for the stack? I.e. if AS decides to use a frame near the beginning as the reference frame and the outputted tiff has the midpoint of capture time that could make the moons not line up properly in WJ?

Might go and  do some testing on existing Jupiter data to see what happens.

I just checked the laptop time, no issue. If your right Craig. It all sounds a bit hit and miss if that is indeed the cause ?

[CraigT82]

Just thinking about it some more... winjupos shouldn't really care what the time of the capture is, what matters is the time lapsed from one capture to the next, as this will determine how far it needs to derorate the frame to match the reference frame (the middle one). 

As long as the measurement frame is accurately placed around the globe, (which let's WJ align the images to be stacked) and the timestamps of the individual tiffs loaded in are representative of the actual time lapsed between capturing them, then the time and date could be wrong and it should still work??

[neil phillips]

6 minutes ago, CraigT82 said:

Just thinking about it some more... winjupos shouldn't really care what the time of the capture is, what matters is the time lapsed from one capture to the next, as this will determine how far it needs to derorate the frame to match the reference frame (the middle one). 

As long as the measurement frame is accurately placed around the globe, (which let's WJ align the images to be stacked) and the timestamps of the individual tiffs loaded in are representative of the actual time lapsed between capturing them, then the time and date could be wrong and it should still work??

All i know is i did the midpoint time. and two moons were slightly off. So i thought weird ok try capture time. they were nearly. (And underline) nearly aligned at that point. I wouldnt have even noticed if the moons hadnt been in my image often i crop them if they are far out 

Edited July 31, 2021 by neil phillips

[morimarty]

I think the WJ image is a big enhacement Pete it seems to have more substance to it. You can also clearly see it in Simons brilliant image. I 'm also currentlly having a play around with WJ but it does look a bit of a learning curve

[U Cyg]

Quite an improvement there Pete, well worth the effort.  I have never used WJ and doubt that I could learn a new software now even if I wanted to!

Angie