DSS alignment puzzle

[Tommohawk]

Fanfare bunting and much general celebration!! Well tentatively anyhow. I think I may have sorted it.. with a bit of help from my friends!

I reverted to Louise's idea of the number of stars. As posted before, I couldn't reduce the star count below about 1400 because if I upped the threshold a single percent further, it found no stars at all. BUT when I checked the "reduce noise by using median filter" (under advanced settings when registering) the star count dropped to 550 ish. So then after registering and stacking again, I have a result free from blur/melt whatever. I guess its asking for trouble imaging in Cygnus at F2.8.

SO, I will look again at some of my earlier stuff and see what the star counts were, and will post the end result when processed on a separate thread. 

Thanks again for the help and suggestions. Now... anyone know of a good counselling service?? 

[jazza]

Heyo Tommohawk,

Glad to hear you fixed it, just for completeness (please note I gave not an iota of consideration to background level or keeping colours nice etc, was just looking for the problem): 

First off, no probs with your individual frames (do gifs work here? about to find out :D)

I just had a look myself and can reproduce your fun times if I try hard enough... so at least you're not crazy?

I can also however get it to stack just fine:

Method that works fine:
Load lights only
Check all
Register checked pictures (untick stack after registering)
Compute offsets
Load dark
Load Flat
Load offset/bias
Stack Checked pictures.

As you've figured out I suspect you may end up with too many stars if you dark subtract before registering.

Sorry for the large post
 

[nucdoc]

 

I also had a look at your data, and agree with the others here that your subs are fine, and they integrate with no problem.  For this exercise I used PI just to make sure that the images would align and stack properly.

I see that you solved the problem, so congrats.

 

 

[Thalestris24]

14 hours ago, Tommohawk said:

Fanfare bunting and much general celebration!! Well tentatively anyhow. I think I may have sorted it.. with a bit of help from my friends!

I reverted to Louise's idea of the number of stars. As posted before, I couldn't reduce the star count below about 1400 because if I upped the threshold a single percent further, it found no stars at all. BUT when I checked the "reduce noise by using median filter" (under advanced settings when registering) the star count dropped to 550 ish. So then after registering and stacking again, I have a result free from blur/melt whatever. I guess its asking for trouble imaging in Cygnus at F2.8.

SO, I will look again at some of my earlier stuff and see what the star counts were, and will post the end result when processed on a separate thread. 

Thanks again for the help and suggestions. Now... anyone know of a good counselling service?? 

Glad you sorted it Tom! I owe you an apology... I've never used the threshold setting before myself and so didn't quite twig what you were doing. I've only ever adjusted the star count via the median filter. Had I twigged, I would have said more explicitly to do that before and might have saved you some grief! Still, it's all a learning experience

Louise

[Tommohawk]

19 hours ago, jazza said:

Heyo Tommohawk,

Glad to hear you fixed it, just for completeness (please note I gave not an iota of consideration to background level or keeping colours nice etc, was just looking for the problem): 

First off, no probs with your individual frames (do gifs work here? about to find out :D)

I just had a look myself and can reproduce your fun times if I try hard enough... so at least you're not crazy?

I can also however get it to stack just fine:

Method that works fine:
Load lights only
Check all
Register checked pictures (untick stack after registering)
Compute offsets
Load dark
Load Flat
Load offset/bias
Stack Checked pictures.

As you've figured out I suspect you may end up with too many stars if you dark subtract before registering.

Sorry for the large post
 

Thanks so much - appreciate folk taking the time to help out! I need to look in more detail at what actually goes on in this process - you seem to suggest that calibration can be done either before or after registration?

19 hours ago, nucdoc said:

 

I also had a look at your data, and agree with the others here that your subs are fine, and they integrate with no problem.  For this exercise I used PI just to make sure that the images would align and stack properly.

I see that you solved the problem, so congrats.

 

Kind of you to do that - and I think the PI version looks maybe better than the DSS version!

17 hours ago, Thalestris24 said:

Glad you sorted it Tom! I owe you an apology... I've never used the threshold setting before myself and so didn't quite twig what you were doing. I've only ever adjusted the star count via the median filter. Had I twigged, I would have said more explicitly to do that before and might have saved you some grief! Still, it's all a learning experience

Louise

Oooh now she tells me!!!

No worries, I wouldn't have sorted it at all if you hadnt got me pointed in the right direction.

Still sort of wonder why the star count is an issue. Out of intrest, I went through about 10 of my latest DSO projects, and although some have star counts in the hundreds, many have counts over 3 or 4 thousand with results that look fine. Its a mystery.

[Thalestris24]

35 minutes ago, Tommohawk said:

Thanks so much - appreciate folk taking the time to help out! I need to look in more detail at what actually goes on in this process - you seem to suggest that calibration can be done either before or after registration?

Kind of you to do that - and I think the PI version looks maybe better than the DSS version!

Oooh now she tells me!!!

No worries, I wouldn't have sorted it at all if you hadnt got me pointed in the right direction.

Still sort of wonder why the star count is an issue. Out of intrest, I went through about 10 of my latest DSO projects, and although some have star counts in the hundreds, many have counts over 3 or 4 thousand with results that look fine. Its a mystery.

Hi

I don't pretend to understand what was going on! It might be that the threshold setting affects the noise whereas the median filter can select the brightest stars. 

Louise

[Tommohawk]

Just adding this footnote to what is a very old thread, as it may help others - and very probably me at some point in the future when I've forgotten!

Leaving aside the star count issue, it seems DSS definitely doesn't like dealing with flats in TIFF format. I've just acquired a new scope, and did some new flat light and flat darks in TIFF format, where my previous ones were in FITS. Not sure why I used TIFF - I think I just find it easier to view them than FITS (lights and darks are all FITS)- and I get exactly the same problem as shown on my opening post on this thread - a big smeary area. Changing the star count makes no difference.

This resolves by using flats in FITS format, even with a star count over 4000. (10% threshold and median filter on).  So the take home message is... I think... use FITS, not TIFF, for all.

Funny thing is - I'm doing the exact same target, the Veil nebula. Spooky.