Jump to content

  •  

CNers have asked about a donation box for Cloudy Nights over the years, so here you go. Donation is not required by any means, so please enjoy your stay.

Photo

IC 1805 - bortle 5 hazy skies - over 7 hours integration ...

  • Please log in to reply
44 replies to this topic

#26 kevinbreen

kevinbreen

    Gemini

  • -----
  • Posts: 3,118
  • Joined: 01 Mar 2017
  • Loc: Wexford, Ireland

Posted 21 September 2021 - 03:31 PM

Hi everyone

2 nights not favorable, hazy , second one moon in the sky , lit over 80 percent.


https://www.cloudyni...11_13344101.png

Nice detail though ,especialy the sparkling birth chamber of stars..at least that is how i see it.

Background stars bit less good...ah well..

HOO approach, HA in the red channel , OIII and Hb in the green and blue.. L-enhance 76 mm scope at F/4.5,unmodded DSLR


Kinda like it though...


Excellent image Freddie. I'm very interested in the L-Enhance filter. Did it block the 80% moon for you? I ask because I read somewhere that it doesn't but I found that hard to believe. I use a modded Canon 600d and I'd like to be able to image on moonlit nights. Any advice you can give I'd be very grateful.

#27 F.Meiresonne

F.Meiresonne

    Cosmos

  • *****
  • topic starter
  • Posts: 7,971
  • Joined: 22 Dec 2003
  • Loc: Eeklo,Belgium

Posted 21 September 2021 - 03:33 PM

Excellent image Freddie. I'm very interested in the L-Enhance filter. Did it block the 80% moon for you? I ask because I read somewhere that it doesn't but I found that hard to believe. I use a modded Canon 600d and I'd like to be able to image on moonlit nights. Any advice you can give I'd be very grateful.

Imho it helps alot. I used it on the Rosetta too with the moon half lit and about 30° away and it still produce alot of nebulosity.

I quite like it.



#28 mackiedlm

mackiedlm

    Viking 1

  • -----
  • Posts: 521
  • Joined: 16 Jul 2019
  • Loc: Galway, Ireland

Posted 21 September 2021 - 04:41 PM

can you elaborate?

I'm still seeking a standard HOO process in PI. Thank you!
 

Hi Elf,

 

I guess I should have said "simple" rather than "Standard" I am very far from being experienced or knowledgeable enough on PI to say what is "standard". But I guess what I did is currently my "standard" for this type of data. I expect you'll be able to point out plenty things I'm doing wrong or could do better. And I would highly value such comments.

 

Workflow was;

Dynamic crop,

Extract channels, Discard B (sometimes, if it looks interesting, I may keep B it and blend it in later)

R=Ha, G-Oiii

Deconvolution on Ha

DBE on both

HT to non-linear -  a moderate stretch and eyeballing background to have them similar (there must must be a better way to have them equal? )

Smoothed range mask on Oiii then CT to boost the brightness of the Oiii signal

Combine with SHO-AIP script, (I really dont know what I'm doing here so its trial and error. On this one I got what I wanted with.. 

fredheart.JPG

 

Note the Ha used as L also

 

masked with luminance mask combined with starmask

LHE

CT for Saturation and brightness

SCNR (strength 0.5)

TGVD

 

I then took it to Photoshop for final tweaks - I expect these can all be done in PI but I dont know how yet blush.gif

 

this was - 

 

Camera raw filter to selectively adjust the colours to exactly what I wanted

Selective Highpass routine to make the inner core details pop

Some healing brush to get rid of some of the big halo's (star cores masked)

Final levels for background and curves for brightness.

Saved to TINY CN file size jpg 

 

If this had been my own data and I was looking to get it as good as I could possibly make it, I may have done some more noise reduction, some star reduction and maybe spent more time on the colour (as Freddy says - maybe a bit soso wink.gif )

 

I hope this is of some value to you.

 

David.


  • the Elf likes this

#29 mackiedlm

mackiedlm

    Viking 1

  • -----
  • Posts: 521
  • Joined: 16 Jul 2019
  • Loc: Galway, Ireland

Posted 21 September 2021 - 04:45 PM

How did you get so much nebulosity out of it? Did you stretch it further,? Also the overal look is quite more refined...

I think the big thing on bringing out the nebulosity was using your Ha data as L also. And the boost of the Oiii probably helped too. See my flow above 



#30 mackiedlm

mackiedlm

    Viking 1

  • -----
  • Posts: 521
  • Joined: 16 Jul 2019
  • Loc: Galway, Ireland

Posted 21 September 2021 - 04:53 PM

Excellent image Freddie. I'm very interested in the L-Enhance filter. Did it block the 80% moon for you? I ask because I read somewhere that it doesn't but I found that hard to believe. I use a modded Canon 600d and I'd like to be able to image on moonlit nights. Any advice you can give I'd be very grateful.

In my limited experience the l-enhance does a great job - it makes the difference between being able to image or not with a bright moon.

 

But I do believe that when you have a bright moon and not the clearest of sky - high haze - you are going to get moonlight leaking in on the Oiii channel - it does not destroy the Oiii but it does make it more difficult to bring out. A good strong oiii target will still show but a very faint oiii may be washed out almost entirely. But it is still an absolute saviour - I just take care with choice of target in such conditions.



#31 vidrazor

vidrazor

    Surveyor 1

  • *****
  • Posts: 1,656
  • Joined: 31 Oct 2017
  • Loc: North Bergen, NJ

Posted 21 September 2021 - 09:51 PM

Nice capture!

 

Honestly, I don't know how to deal with Ha and/or Oi data, so I just went mostly for luminosity here. Siril, Starnet++, and Photoshop. I experimented a bit with a channel mixer, but used the process centrally only, it looked weird around the perimeter. As I was opening up data, I noticed what looks like walking noise, and I thought of processing it, but I just left well enough alone. I started hitting too much noise, so I cut it there. Color-wise, I looked around online for references, and as usual with such targets, the sky's the limit on how you want it to look, so I made it like this. I'm sure there's much better ways to process this data, but as I said I'm not familiar with processing Ha and/or Oi data. Hope you'll like it! smile.gif

Attached Thumbnails

  • IC1805-C.jpg

Edited by vidrazor, 21 September 2021 - 10:10 PM.

  • F.Meiresonne likes this

#32 Ivo Jager

Ivo Jager

    Vendor ( Star Tools )

  • *****
  • Vendors
  • Posts: 552
  • Joined: 19 Mar 2011
  • Loc: Melbourne, Australia

Posted 21 September 2021 - 10:41 PM

Hi Freddy,

 

The background appears very, very uneven with lots of dark streaks (running at ~10/190 degrees) that do not constitute real detail (you should be able to see this immediately during the first/diagnostics AutoDev). Did something go wrong during calibration? It makes it really hard to push your data without starting to show detail that doesn't actually exist (as can be seen in some renditions here) :(

 

Did dithering work as expected those two nights? Though walking noise is not too prevalent, correlated noise is, however. This doesn't look like the data you typically produce... Did something change in your setup?


  • F.Meiresonne and 17.5Dob like this

#33 F.Meiresonne

F.Meiresonne

    Cosmos

  • *****
  • topic starter
  • Posts: 7,971
  • Joined: 22 Dec 2003
  • Loc: Eeklo,Belgium

Posted 22 September 2021 - 01:03 AM

Hi Ivo,

 

I noticed some of that but i that the filter had something to do with.

I did not change anything in my setup , i never do. I coud not see any issues whilest calibrating or dithering.Maybe i did not dither severe enough,i believe i lowered a bit the parameters.I will check this tonight.

I can't see any reason, then my dithering setting was too low then....

 

What is correlated noise?

 

Thanks for your reply and good you see and mention this.. I was unaware  , i did not use even then walking noise denoiseing feature

 

bit of a bummer , cause this call for a complete redo..;ah well, it's fungrin.gif


Edited by F.Meiresonne, 22 September 2021 - 01:15 AM.


#34 F.Meiresonne

F.Meiresonne

    Cosmos

  • *****
  • topic starter
  • Posts: 7,971
  • Joined: 22 Dec 2003
  • Loc: Eeklo,Belgium

Posted 22 September 2021 - 01:14 AM

Nice capture!

 

Honestly, I don't know how to deal with Ha and/or Oi data, so I just went mostly for luminosity here. Siril, Starnet++, and Photoshop. I experimented a bit with a channel mixer, but used the process centrally only, it looked weird around the perimeter. As I was opening up data, I noticed what looks like walking noise, and I thought of processing it, but I just left well enough alone. I started hitting too much noise, so I cut it there. Color-wise, I looked around online for references, and as usual with such targets, the sky's the limit on how you want it to look, so I made it like this. I'm sure there's much better ways to process this data, but as I said I'm not familiar with processing Ha and/or Oi data. Hope you'll like it! smile.gif

Very nice processing , i must say...



#35 the Elf

the Elf

    Soyuz

  • *****
  • Posts: 3,949
  • Joined: 06 Sep 2017
  • Loc: Germany

Posted 22 September 2021 - 05:14 AM

Excellent image Freddie. I'm very interested in the L-Enhance filter. Did it block the 80% moon for you? I ask because I read somewhere that it doesn't but I found that hard to believe. I use a modded Canon 600d and I'd like to be able to image on moonlit nights. Any advice you can give I'd be very grateful.

The moon is neutral in color. It reflects the sunlight without changing the spectrum. The sun's spectrum is the yellow curve (5777K) in this diagram

640px-BlackbodySpectrum_loglog_de.svg.pn

https://de.wikipedia...m_loglog_de.svg

 

Your camera (and you as well) only sees the visible range, marked by the rainbow. As it is a color camera it basically sorts left, middle and right part of the rainbow to the three color channels. The actual sensitivity for an unmodded T3i is this (actual measured data):

rel_Sensitivity_Canon_600D.gif

 

Your modded camera will probably look like this (my estimation, line drawn by hand from knowledge of other cameras, rough estimation)

modded.png

 

The emission objects are at the well known Ha and O-III lines, as well as some S-II and Hb.

Important thing to understand: the emission nebula signal is the same if you use the L-enhance or not! The filter can only block the light you don't want. The l-enhance transmission curve is available in the internet:

l-enhance_filter_curve-600x292.png

https://images.squar...rve-600x292.png

 

You have to multiply every value form the sun spectum (5777K curve) with each colored line in the relative sensitivity of the camera and calculate the area under each curve.

Next you cut away all the parts from the sun spectrum that the filter blocks (that is multiply the sun curve by the filter curve). Sum up the area for each channel again. Even without actually doing the math it is quite obvious that a huge amount of the sun or moon signal is blocked.

 

You must be aware of the fact that all star color is killed. Best thing you can achieve is white stars with slight variations between green and red. If you want a natural image you have to shoot an additional data set without filter for the stars and mix the true color stars in in post processing.


Edited by the Elf, 22 September 2021 - 05:17 AM.


#36 kevinbreen

kevinbreen

    Gemini

  • -----
  • Posts: 3,118
  • Joined: 01 Mar 2017
  • Loc: Wexford, Ireland

Posted 22 September 2021 - 05:18 AM

The moon is neutral in color. It reflects the sunlight without changing the spectrum. The sun's spectrum is the yellow curve (5777K) in this diagram
640px-BlackbodySpectrum_loglog_de.svg.pn
https://de.wikipedia...m_loglog_de.svg

Your camera (and you as well) only sees the visible range, marked by the rainbow. As it is a color camera it basically sorts left, middle and right part of the rainbow to the three color channels. The actual sensitivity for an unmodded T3i is this (actual measured data):
rel_Sensitivity_Canon_600D.gif

Your modded camera will probably look like this (my estimation, line drawn by hand from knowledge of other cameras, rough estimation)
modded.png

The emission objects are at the well known Ha and O-III lines, as well as some S-II and Hb.
Important thing to understand: the emission nebula signal is the same if you use the L-enhance or not! The filter can only block the light you don't want. The l-enhance transmission curve is available in the internet:
l-enhance_filter_curve-600x292.png
https://images.squar...rve-600x292.png

You have to multiply every value form the sun spectum (5777K curve) with each colored line in the relative sensitivity of the camera and calculate the area under each curve.
Next you cut away all the parts from the sun spectrum that the filter blocks (that is multiply the sun curve by the filter curve). Sum up the area for each channel again. Even without actually doing the math it is quite obvious that a huge amount of the sun or moon signal is blocked.

You must be aware of the fact that all star color is killed. Best thing you can achieve is white stars with slight variations between green and red. If you want a natural image you have to shoot an additional data set without filter for the stars and mix the true color stars in in post processing.


Thanks very much Elf

#37 F.Meiresonne

F.Meiresonne

    Cosmos

  • *****
  • topic starter
  • Posts: 7,971
  • Joined: 22 Dec 2003
  • Loc: Eeklo,Belgium

Posted 22 September 2021 - 12:50 PM

Hi Freddy,

 

The background appears very, very uneven with lots of dark streaks (running at ~10/190 degrees) that do not constitute real detail (you should be able to see this immediately during the first/diagnostics AutoDev). Did something go wrong during calibration? It makes it really hard to push your data without starting to show detail that doesn't actually exist (as can be seen in some renditions here) frown.gif

 

Did dithering work as expected those two nights? Though walking noise is not too prevalent, correlated noise is, however. This doesn't look like the data you typically produce... Did something change in your setup?

Ivo,

 

Is this what you mean, this is the initial stretch...

 

 

I checked my dithering distance. Guess it was wrong cause it was still on the setting for my 102 mm scope 3 x  3 scale PHD2 *.95 image scale is roughly 9 pixels, normaly i try to dither 18-20 pixels...

I just  bluntly forgot to alter the setting....in APT..

Can't think of any other reason.

Attached Thumbnails

  • Strikes.JPG

Edited by F.Meiresonne, 22 September 2021 - 01:36 PM.


#38 vidrazor

vidrazor

    Surveyor 1

  • *****
  • Posts: 1,656
  • Joined: 31 Oct 2017
  • Loc: North Bergen, NJ

Posted 22 September 2021 - 08:36 PM

Ivo,

Is this what you mean, this is the initial stretch...

I checked my dithering distance. Guess it was wrong cause it was still on the setting for my 102 mm scope 3 x  3 scale PHD2 *.95 image scale is roughly 9 pixels, normaly i try to dither 18-20 pixels...

I just  bluntly forgot to alter the setting....in APT..

Can't think of any other reason.

Yeah, that's the noise I came up against. At first I wasn't sure what it was. It's all over your image. Dunno if you stack more data on top if you can neutralize it, or just start over.


  • F.Meiresonne likes this

#39 F.Meiresonne

F.Meiresonne

    Cosmos

  • *****
  • topic starter
  • Posts: 7,971
  • Joined: 22 Dec 2003
  • Loc: Eeklo,Belgium

Posted 23 September 2021 - 01:02 AM

It calls for a redo, i did not quite see the noise as such at start. But processing it in Startools , well it was difficult to get something descent , i struggled with it and that should have rang a bell. If you cannot get a good first picture from Startools in say about 10 minuts, you can bet something is wrong with the stack. That is another reason these threads are so usefull, 10 pair of eyes see more then 1 pair...

 

I have difficulties to evaluate still many things, and how to tackle it then...


  • Ivo Jager likes this

#40 the Elf

the Elf

    Soyuz

  • *****
  • Posts: 3,949
  • Joined: 06 Sep 2017
  • Loc: Germany

Posted 23 September 2021 - 01:24 AM

Mono DSLRs need strong dither. I use a 30x30 pixel region. As I'm using an OAG the arcsec per pixel in the guide camera is about the same as in the imaging camera. (The guide camera has got smaller pixels but also has got it's own reducer in front so it's about equal).

With the 7nm Ha filter I need 30+ subs for a clean image. Imho 20 subs is the minimum for dither to do something at all. That is for the T3i. Other cameras are different for sure.


  • F.Meiresonne likes this

#41 Ivo Jager

Ivo Jager

    Vendor ( Star Tools )

  • *****
  • Vendors
  • Posts: 552
  • Joined: 19 Mar 2011
  • Loc: Melbourne, Australia

Posted 23 September 2021 - 01:51 AM

Is this what you mean, this is the initial stretch...

Yep!

 

"Correlated" noise is any sort of noise that is not just 1-pixel speckles, but looks like blobs of pixels, streaks of pixels, etc. E.g. it is bigger than just one pixel in size. Walking noise typically looks like single pixel-wide streaks.

 

If you see "broader" streaks going into one direction (the direction the sky drifted in), it can also be an issue with you flats (usually flats not matching the actual uneven lighting in your image).


  • F.Meiresonne likes this

#42 F.Meiresonne

F.Meiresonne

    Cosmos

  • *****
  • topic starter
  • Posts: 7,971
  • Joined: 22 Dec 2003
  • Loc: Eeklo,Belgium

Posted 23 September 2021 - 04:10 AM

Yep!

 

"Correlated" noise is any sort of noise that is not just 1-pixel speckles, but looks like blobs of pixels, streaks of pixels, etc. E.g. it is bigger than just one pixel in size. Walking noise typically looks like single pixel-wide streaks.

 

If you see "broader" streaks going into one direction (the direction the sky drifted in), it can also be an issue with you flats (usually flats not matching the actual uneven lighting in your image).

Hmmm, so the dithering may not be the only issue here...



#43 F.Meiresonne

F.Meiresonne

    Cosmos

  • *****
  • topic starter
  • Posts: 7,971
  • Joined: 22 Dec 2003
  • Loc: Eeklo,Belgium

Posted 23 September 2021 - 04:11 AM

Mono DSLRs need strong dither. I use a 30x30 pixel region. As I'm using an OAG the arcsec per pixel in the guide camera is about the same as in the imaging camera. (The guide camera has got smaller pixels but also has got it's own reducer in front so it's about equal).

With the 7nm Ha filter I need 30+ subs for a clean image. Imho 20 subs is the minimum for dither to do something at all. That is for the T3i. Other cameras are different for sure.

30 pixels !?  That will affect the settle time....



#44 the Elf

the Elf

    Soyuz

  • *****
  • Posts: 3,949
  • Joined: 06 Sep 2017
  • Loc: Germany

Posted 24 September 2021 - 02:59 PM

0.733 arcsec per pixel x an average of 15 pixels movement = about 10 arsec to move per dither. Guiding runs on 0.5x sidereal tracking, that is 7.5 arcsec /second. A bit more than one second. For the quad at 2 arcsec/pixel it is about 3 seconds, sometimes 5, not more. Compared to 15 minute subs not an issue. When it comes to the 16mm Rokinon it is a different story. LOL.



#45 F.Meiresonne

F.Meiresonne

    Cosmos

  • *****
  • topic starter
  • Posts: 7,971
  • Joined: 22 Dec 2003
  • Loc: Eeklo,Belgium

Posted 25 September 2021 - 01:15 AM

My settle times are about 20 sec , or around that...it seems not important, more important is settting the distance...cause in this thread i forgot it over 2 separate sessions. Still can't believe i forgot it so bluntly cause i calculated them for each of my scopes ...stupid.




CNers have asked about a donation box for Cloudy Nights over the years, so here you go. Donation is not required by any means, so please enjoy your stay.


Recent Topics






Cloudy Nights LLC
Cloudy Nights Sponsor: Astronomics