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

First DSO Attempt M45 - Help with weirdness!

  • Please log in to reply
39 replies to this topic

#1 Monkeybird747

Monkeybird747

    Vostok 1

  • *****
  • topic starter
  • Posts: 145
  • Joined: 31 Aug 2017
  • Loc: Louisville, KY

Posted 08 January 2018 - 09:15 PM

I could really use some help on this from the CN brain trust! The short, rambling story (apologies for the stream of consciousness approach):

 

During the last year I've been assembling my AP setup. During this time I've been practicing processing other people's data in PI. I've become comfortable interfacing with the processes and scripts, but don't claim total mastery. I've made a few equipment test runs, but this was my first full DSO imaging session. Full image data set is here (some screen shots below):

 

https://www.dropbox....CnrFp2Lpxa?dl=0

 

I thought things went pretty well, despite some questionable guiding. During preprocessing I noticed a couple of things going on. There were some significant black lines beneath bright stars. Also, the gradient was odd and multi-colored, but I guess it sort of fits with where the brighter lights of the city were oriented (lower right). Hard to describe. See image below. I tried a few test runs of ABE and DBE, but all they did was change the color and location of the gradients.

 

I read some forums on Canon 1100D sensor problems/limitations that produces these black lines below bright stars, but before I start blaming the equipment, I want to make sure there is nothing obvious to more experienced imagers that I may be missing. The only fix I've read is in Photoshop, and I'm wondering what tool in PI might fix these lines?

 

Aside from that the stars are not 100% round. There is a small chance I was briefly guiding through some tree branches... Not sure about that though.

 

 

Basically, I'm wondering if my data is garbage, and if it is did I make it garbage, or was there an equipment fault? Should I bother taking this data further into preprocessing, or will it prove to be even more frustrating that it already is?

 

Let me know what other data I could provide to help diagnose.

 

 

iEQ45 Pro, Canon 1100D, SVR90T, Polemaster, PHD2, ASI120 guide camera, 60mm guide scope, BYEOS, Pixinsight

20 lights 800ISO 300s

30 flats 800ISO 1/13s (flats look kinda weird to me. tshirt fabric on sewing loop with ipad white screen)

10 darks same ISO/Temp as lights

150 bias

 

It was wicked cold (14 degrees F), if that matters. 

 

Here is the calibrated, corrected, and debayered integration attempt;

 

Screen Shot 2018 01 08 At 20.04.29
 
Here is a preview during the CosmeticCorrection process. Is the weird checkered pattern normal during live view mode?
 
Screen Shot 2018 01 08 At 19.30.53

 

 

Here is master flat. It looks very different depending on the program opening it. I debayered it for the screen shot.

 

Screen Shot 2018 01 08 At 20.10.09

 

A single flat frame in Photoshop:

 

Screen Shot 2018 01 08 At 20.18.56
 
Super bias and Master bias with auto stretch:
 
Screen Shot 2018 01 08 At 20.04.56
Screen Shot 2018 01 08 At 20.06.34

 

Histogram in BYEOS of single 300s light:

 

IMG 9701

 

IMG 9702

 

Some PHD graphs. Things got worse as night progressed.

 

IMG 9683
IMG 9685
IMG 9705

 

Snapshot of seeing conditions. FYI, ADM plates are physically painful to the touch when cold-soaked at 14 degrees!

IMG 9689

Edited by Monkeybird747, 08 January 2018 - 09:19 PM.

  • artem2 and Augustus like this

#2 Jim Waters

Jim Waters

    Apollo

  • *****
  • Posts: 1056
  • Joined: 21 Oct 2007
  • Loc: Phoenix, AZ USA

Posted 09 January 2018 - 12:10 AM

From doing a quick PI stacking and initial processing it looks like the 'Lights' are over saturated.  The BYEOS histogram tends to support this.  Try to keep the histogram between 25-30%.

 

FWIW - I have never had good results with a Super BIAS.


Edited by Jim Waters, 09 January 2018 - 12:12 AM.


#3 Monkeybird747

Monkeybird747

    Vostok 1

  • *****
  • topic starter
  • Posts: 145
  • Joined: 31 Aug 2017
  • Loc: Louisville, KY

Posted 09 January 2018 - 01:37 PM

Jim, thanks for taking the time to do that. The super bias looked a little off compared to previous ones, and I ran it with the normal master with the same/similar result. What do you think about the flats? They were right on the 2/3 mark on BYEOS histogram. Is 1/13s exposure too long? I would have to make a thinner T-shirt fabric hoop, or get a brighter light source to take shorter exposure flats.

I took a lot of test exposures and examined in PI statistics. I was given advice to shoot for an exposure that produced a mean ADU/DN of 800-1000. I thought I had it dialed in, but it looks like I didn’t set the right exposure, as these are showing up around 1200 ADU/DN, so i definitely over shot. I’ll shoot for the histogram target you suggested on the next run.

So will the oversaturation cause the black lines beneath bright stars? Does that also produce issues with background extraction/gradient removal?

Do you think this data is salvageable, or would you scrap it and start over? If I can’t get past the background extraction I don’t think there is much point in continuing with it.

I also read some threads on issues with black lines on this canon model at 800 iso and lower. I might try 1600.

#4 the Elf

the Elf

    Ranger 4

  • -----
  • Posts: 351
  • Joined: 06 Sep 2017
  • Loc: Germany

Posted 09 January 2018 - 02:57 PM

Neither saying it is the camera nor it is not. I image with an EOS600D and never had any trouble so far. As used ones are available cheap you might try one. When I got it new 5 years ago I asked for the 1100D and the shop attendad said "I don't want to sell that to you." I do not know any particular reason. I recommend the 600D as it has still some sensitivity at H-alpha even unmodded. I checked it in the optics lab at work. Perhaps you try and then sell the one you like less to a daytime imager. What I can say for sure: too much exposure cannot be fixed. Trust Jim. He knows what he is talking about!


  • Jim Waters likes this

#5 Monkeybird747

Monkeybird747

    Vostok 1

  • *****
  • topic starter
  • Posts: 145
  • Joined: 31 Aug 2017
  • Loc: Louisville, KY

Posted 09 January 2018 - 03:16 PM

Thanks Elf, I was considering a different camera at some point, but that opens up a whole other can of worms. An newer canon is not the worst idea I’ve heard, as it would plug and play with my current setup.

Jim has given good guidance before, and I was glad to see he chimed in on this topic. I’ll just start a new batch of data with the proper histogram at 800iso, and maybe test 1600 and go from there. If the lines are still there then I might start shopping around.
  • Jim Waters likes this

#6 Jim Waters

Jim Waters

    Apollo

  • *****
  • Posts: 1056
  • Joined: 21 Oct 2007
  • Loc: Phoenix, AZ USA

Posted 09 January 2018 - 04:19 PM

Some random inputs - my opinions here...

 

Your SV90T optics need to be cleaned especially the sensor of your 1100D.  But its likely your sensor.

 

You are getting a huge amount of vignetting.  When I was using my Ha Mod T3i (APS-C) I got much less and I was using my SV90T-CF scope.  Not sure why you are getting so much.  Is your focuser 2.5" or 2"?  Are you using the SV screw-in FF?

 

Your Flats look over exposed.  I keep my histogram around 30-40%. 

 

I have seen some posts that say bright or over exposed stars produce streaks on the 1100D sensor.  Try renting an APS-C camera such as a 7DMkII but keep the histogram between 25-30%.  No higher.  Use PI Canon Banding Reduction Script to fix any banding.  Do not try using a FF camera with the SV90T.  You will be disappointed because of vignetting.

 

I imported the Raw 'Light' CR2 files and lowered the histogram to 25% using DPP and reprocessed in PI with and without BIAS and Flats.  I get basically the same results.  You can find YouTube videos that advocate 'Expose To The Right' for Astrophoto's but I have never seen good results.

 

If it were me I would rent another camera and give it a try.


Edited by Jim Waters, 09 January 2018 - 04:21 PM.

  • sharkmelley likes this

#7 Monkeybird747

Monkeybird747

    Vostok 1

  • *****
  • topic starter
  • Posts: 145
  • Joined: 31 Aug 2017
  • Loc: Louisville, KY

Posted 09 January 2018 - 07:14 PM

Some random inputs - my opinions here...

Your SV90T optics need to be cleaned especially the sensor of your 1100D. But its likely your sensor.

You are getting a huge amount of vignetting. When I was using my Ha Mod T3i (APS-C) I got much less and I was using my SV90T-CF scope. Not sure why you are getting so much. Is your focuser 2.5" or 2"? Are you using the SV screw-in FF?

Your Flats look over exposed. I keep my histogram around 30-40%.

I have seen some posts that say bright or over exposed stars produce streaks on the 1100D sensor. Try renting an APS-C camera such as a 7DMkII but keep the histogram between 25-30%. No higher. Use PI Canon Banding Reduction Script to fix any banding. Do not try using a FF camera with the SV90T. You will be disappointed because of vignetting.

I imported the Raw 'Light' CR2 files and lowered the histogram to 25% using DPP and reprocessed in PI with and without BIAS and Flats. I get basically the same results. You can find YouTube videos that advocate 'Expose To The Right' for Astrophoto's but I have never seen good results.

If it were me I would rent another camera and give it a try.


Good stuff Jim. I have access to another canon I can borrow. I forget the model but it is APS-C and has the digic 5 processor. It sounds like you took calibration frames out of the equation, so it must be something with the lights or the camera. I might back off my flats exposure too. If I get some clear skies soon I think I’ll just grab some star field shots at different histogram points (including flats) and see if I can figure this out.

I have the 2.5” stellarvue focused (Svr90t-25sv), and stellarvue threaded .8x field flattener/reducer, and t-ring adapter. I’m not sure about the excessive vinetting. The flattener is supposedly built for the specific scope in terms of spacing, with some tiny shims added to get it just right. Maybe the spacing is a little off?

Multiple PI processing runs take a lot of time. Folks like you that take the time to dig that deep into an issue are what really make this forum great. Thanks again!

#8 Jim Waters

Jim Waters

    Apollo

  • *****
  • Posts: 1056
  • Joined: 21 Oct 2007
  • Loc: Phoenix, AZ USA

Posted 09 January 2018 - 07:43 PM

You are using the SV 0.8x FR/FF.  That's likely causing some of the vignetting.  I just use their threaded FF.   I tried the TeleVue 0.8x FR/FF and the images look awful.  Flats don't correct the problem.  The light falloff is too great.  I am considering getting their SVQ100-3FT Astrograph to use with my 6D.  If I do I will end up selling my 90T.

 

Try imaging with and without the SV FR/FF.

 

Keep us posted.



#9 Monkeybird747

Monkeybird747

    Vostok 1

  • *****
  • topic starter
  • Posts: 145
  • Joined: 31 Aug 2017
  • Loc: Louisville, KY

Posted 09 January 2018 - 07:46 PM

You are using the SV 0.8x FR/FF. That's likely causing some of the vignetting. I just use their threaded FF. I tried the TeleVue 0.8x FR/FF and the images look awful. Flats don't correct the problem. The light falloff is too great. I am considering getting their SVQ100-3FT Astrograph to use with my 6D. If I do I will end up selling my 90T.

Try imaging with and without the SV FR/FF.

Keep us posted.


That’s interesting. I hadn’t considered imaging without it. I’ll give it a go.

Edited by Monkeybird747, 09 January 2018 - 09:25 PM.


#10 the Elf

the Elf

    Ranger 4

  • -----
  • Posts: 351
  • Joined: 06 Sep 2017
  • Loc: Germany

Posted 10 January 2018 - 06:29 AM

"expose to the right" is not for astro and there are many occasion where it is second best for daytime as well. It is one of the statements that must be right because some expert says it. It is like "the atom cannot be split". The true question is how many over exposured pixels can you tolerate and where. One has to understand that there is over exposured and over exposured in a DSLR / any OSC. All three channesl blown out white my be easy but only one or two saturated while mostly the red one is still ok needs treatment. When I check histograms on my camera display I can't get rid of the feeling they take samples from the image only and never count all 18 million pixels. I guess BYEOS is doing it well. My best advice: keep the histogram peek low and look for the smal stuff right of the main peak. This might be data you need in your image.

 

I used the regular master bias and the super bias and I can't see much of a difference (created from 100 frames). It may depend on the camera. If you visually compare, make sure you use the same screen stretch!


  • Jim Waters likes this

#11 Monkeybird747

Monkeybird747

    Vostok 1

  • *****
  • topic starter
  • Posts: 145
  • Joined: 31 Aug 2017
  • Loc: Louisville, KY

Posted 10 January 2018 - 11:55 PM

Ok Gents, here is something to look at. I found these two images in the preview section from that same imaging session. They weren't guided. One is me trying to find the Rosette nebula and is just a star field at 800 ISO 60s. BYEOS saves the preview as a jpeg, but the histogram is in the 25-30% zone. The other image is me failing to find, and subsequently giving up on, the Rosette nebula and shooting for the obvious bright target above me (still fiddling with Astrotortilla). The  M45 frame is 800 ISO 120s, and comes in at the 50% histogram mark. I still see the streaks on the M45 120s frame, but they are less pronounced than the 300s subs I posted earlier. The would probably get more pronounced after stacking.

 

I guess my questions now is, will 60-90s subs be the best I'll get from my light polluted back yard, and is it worth it? I mean, I would need several hundred subs to get the hours of exposure I was hoping for. I would think I would be bumping up against the law of diminishing returns at that point. I'm on the edge of a red zone/orange zone at something like Bortle 6-7. 

 

Also, is it normal for the red channel to be further right on the histogram than the B and G? Is this equivalent somehow to skyglow?

 

Do these jpegs provide any additional clues?

 

DELETE PREVIEW 20180106 19h28m36s059ms
Screen Shot 2018 01 10 At 23.57.04

 

MAYBE PREVIEW 20180106 20h35m45s845ms
Screen Shot 2018 01 10 At 23.57.59

 

 


Edited by Monkeybird747, 11 January 2018 - 12:01 AM.


#12 sharkmelley

sharkmelley

    Surveyor 1

  • -----
  • Posts: 1528
  • Joined: 19 Feb 2013

Posted 11 January 2018 - 08:32 AM

 

I have seen some posts that say bright or over exposed stars produce streaks on the 1100D sensor. 

The "black lines below stars" does affect some (but not all) Canon 1100Ds.

 

Examples:

https://stargazerslo...es-below-stars/

http://www.iceinspac...ad.php?t=123486

Both those threads contain links to other threads discussing the same issue.

 

It's a sensor fault and there's nothing you can do except to change the camera.

 

Mark


Edited by sharkmelley, 11 January 2018 - 08:33 AM.

  • 17.5Dob likes this

#13 entilza

entilza

    Gemini

  • *****
  • Posts: 3207
  • Joined: 06 Oct 2014
  • Loc: Oakville, ON, Canada

Posted 11 January 2018 - 10:55 AM

Hey back to your original question, no the data is not garbage, anything that is focused well regardless of LP is usually usable data.  I would consider frames unfocused or with clouds garage data that's about it.  The rest is you just battling what we all do, light pollution and learning the hobby.

 

Take some of the tips here regarding exposure length and keep at it you're doing really well!  Having less exposures is not going to be a big deal just a bit more integration time, but you'll end up with a better calibrated image and easier to post process.

 

Once you get to some dark skies you'll use all your experience to really make use of that time under the stars.



#14 Monkeybird747

Monkeybird747

    Vostok 1

  • *****
  • topic starter
  • Posts: 145
  • Joined: 31 Aug 2017
  • Loc: Louisville, KY

Posted 11 January 2018 - 11:53 AM



I have seen some posts that say bright or over exposed stars produce streaks on the 1100D sensor.

The "black lines below stars" does affect some (but not all) Canon 1100Ds.

Examples:
https://stargazerslo...es-below-stars/
http://www.iceinspac...ad.php?t=123486
Both those threads contain links to other threads discussing the same issue.

It's a sensor fault and there's nothing you can do except to change the camera.

Mark

Thanks Mark, I was afraid of that. I read that 1600 ISO might work better for this model, with respect to the black lines, so I’ll test that next to some 800 ISO shots at the exposure times recommended above. If lines are still there I will have start shopping around. The 1100d was free, so not a big loss. Stand by for a “New Canon DSLR vs Cooled Astro Camera” thread. I’m sure there are quite a few out there.

#15 Monkeybird747

Monkeybird747

    Vostok 1

  • *****
  • topic starter
  • Posts: 145
  • Joined: 31 Aug 2017
  • Loc: Louisville, KY

Posted 11 January 2018 - 11:55 AM

Hey back to your original question, no the data is not garbage, anything that is focused well regardless of LP is usually usable data. I would consider frames unfocused or with clouds garage data that's about it. The rest is you just battling what we all do, light pollution and learning the hobby.

Take some of the tips here regarding exposure length and keep at it you're doing really well! Having less exposures is not going to be a big deal just a bit more integration time, but you'll end up with a better calibrated image and easier to post process.

Once you get to some dark skies you'll use all your experience to really make use of that time under the stars.


Thanks for the encouragement! You’re right though. Every failure is a learning opportunity. If I can process this LP stuff, then some good frames under dark skies will be a cinch!

#16 Monkeybird747

Monkeybird747

    Vostok 1

  • *****
  • topic starter
  • Posts: 145
  • Joined: 31 Aug 2017
  • Loc: Louisville, KY

Posted 13 January 2018 - 01:55 AM

So this was about the best I could come up with. Had to sacrifice some detail in processing. Its a pretty disappointing image. Gave me more pixinsight practice though.

 

M45 Max Cold Integration

 



#17 sharkmelley

sharkmelley

    Surveyor 1

  • -----
  • Posts: 1528
  • Joined: 19 Feb 2013

Posted 13 January 2018 - 07:24 AM

So this was about the best I could come up with. Had to sacrifice some detail in processing. Its a pretty disappointing image. Gave me more pixinsight practice though.

 

Unfortunately those black streaks will always be a severe limitation :(

 

Mark



#18 the Elf

the Elf

    Ranger 4

  • -----
  • Posts: 351
  • Joined: 06 Sep 2017
  • Loc: Germany

Posted 13 January 2018 - 07:42 AM

Hey, this is not a bad image! Stars are small and circular, you have good guiding. Don't be disappointed. Regard it as one step - one important step - on your way to the final image. If you can say what in detail you don't like in the image you can make a plan what you change next time. I made a lot of progress by asking myself "what exactly don't I like in the image". Not critizising, just as an example you might say you dislike the halos on the bright stars or the lack of background which is pitch black or the lack of reflection nebula or the missing colors of all the small stars (there are very few images that show some red stars). By you own knowledge or by a discussion here you can figure out what to do to change this. Some suggestions that might be further discussed

 

- From the discussion it seems like a different camera it the right step and if you go for the 600D it is not expensive and you can have one within a few days. Or you make the big step to an astro cam.

- More darks might help. This is work you can do at daytime.

- Some shorter exposures with less blown out bright stars could be integrated by the HDR process or photoshoped in later.

- Keep the data and reprocess in one year when you have more experience with PI.

 

Depending on the weather here and my housekeeping motivation I might download your original data and do some PI practice. For me it is a good test as I killed Windoze and installed Fedora to my Laptop. This would be my first work with the new OS. Either I get something decent and can tell you a trick or I create a far worse version and you will know that your processing is not that bad. Either way we win. I must admit I have processed only nebulae and galaxies so far and no open clusters. Thanks for putting your data online.

 

the Elf



#19 Monkeybird747

Monkeybird747

    Vostok 1

  • *****
  • topic starter
  • Posts: 145
  • Joined: 31 Aug 2017
  • Loc: Louisville, KY

Posted 13 January 2018 - 12:17 PM

If you get time that would be awesome the Elf, thanks. Your examples of observations are all things I agree with, and I'm going to take another few passes starting fresh with the original integration. I basically start with the maximum and minimum applications of a tool to see what effect they have and go from there. Using Inside Pixinsight and the PI forum as a guide along the way. I can't find a tool in PI that really lets me isolate the black line problem to apply a correction. I'm still working on it, but for now I have been trying to deal with it in PS using the healing brush, but thats a lot of healing. I did see one place in PI where you could specify x and y coordinates of a row or column of pixels, and then set a limit on the height or width and apply a correction, but I still need to play with it. I may inquire on the PI forum.

 

In the posted image, I tried to raise the black point until the lines were mostly gone, but as you noticed it creates an artificial looking black background. I've been experimenting with different range and star masks along with HT, and have seen some minor progress. Using masks it is tricky to not make the nebulosity look as if it is detached from the rest of the image.

 

I just missed a t3i full spectrum modified by spencer's camera by a day or so. They pop up for $300-$350 now and again. I would like to experiment with a full spectrum modified and a Skytech CLS CCD clip in filter with built in UV/IR blocking, and add a Ha filter to combine with the rgb data. I've read arguments both for and against light pollution filters, but I'd like to try it for myself and see. My gut feeling is that I have a better chance at correcting the color from a LP filter than I do trying to process out the gradient. My gut could be out of calibration though. I might take this route before exploring a dedicated astro cam. If I do that, I would want to go the mono/filter wheel route. I really wanted to get a few successful captures under my belt before sinking another 2k into the hobby.

 

Thanks for the replies. I'll post some more attempts here over the next couple of days. If I find something that works I'll try and make a walkthrough with some screenshots. 



#20 the Elf

the Elf

    Ranger 4

  • -----
  • Posts: 351
  • Joined: 06 Sep 2017
  • Loc: Germany

Posted 13 January 2018 - 01:21 PM

How do you like this?

After processing I tell you one thing: get rid of that camera!

Attached Thumbnails

  • the_Elfs_Processing.jpg


#21 the Elf

the Elf

    Ranger 4

  • -----
  • Posts: 351
  • Joined: 06 Sep 2017
  • Loc: Germany

Posted 13 January 2018 - 01:47 PM

Some more feedback:

 

there is a lot of good in your data:

- you got excellent tracking, stars are circular and there is virtually no movement between the subs

- your temperature management is exellent, one sub 0°C, all others 1°C

- your darks perfectly match in temperature one -2°C, all other 0°C ( I use exiftool -CameraTemperature *.CR2|grep Temp in the Linux console, available for Windoze as well)

- you flats are quite nice, nothing to complain

- a lot of small stars to be seen

 

What I did and what I found:

- going thru the subs with blink the background is getting brighter and brighter and brighter. Maybe clouds or dust? Finnally I used only the first 10 lights to avoid shot noise from the bright background.

- You have a satelite trail in sub #4 at 20h57m50s, could be removed during image integration using Winsorized Sigma sigma high = 4.3

- I applied ABE twice, first irst subtraction, second division.

- went thru the standard process of Warren Kellers book Inside Pixinsight: Color Calibration, SCNR, Lin Denoise

- then used a range mask to get the black lines in the background and removed them by histogram transformation

- main stretch was done by masked stretch with standard settings

- fought the lines again protecting stars with a different range mask and build a custom kernel for morphology: a vertikal line 11x3 pixels and applied a median

- nonlinear denois, not much

- fiddeled quite a while on local histogram equalization to enhance some structure that distracts the eye from the bloody black lines

- color saturation of the blue/teal reflection stuff

- color saturation of golden stars to get some other color than blue in the image

- some saturation boost with stars under the mask

- another range mask for the very bright stars that were pink: simply desaturated with curves tool

- for the thumbnail above I applied a gaussian convolution prior to downsample

 

I must admit, I was unable to get the dark lines out of the blue nebula. Without the ever raising background all subs could have been used and a lot more faint reflection would have been there.

 

An astro cam with a filter wheel is by no means a guarantee for perfect images. Depending on the cromatic aberrations of your scope and how parfocal the filters are, they might introduce color fringe. I am still waiting for the repair of a mono mod DSLR and I will try to combine the unmodded OSC and mono L and/or Halpha. Whatch my posts and sooner or later you will find that (unless the camera is broken again).

Use a different DSLR and life will be much easier! Don't frustrate yourself with a shisho process and start with good data. A regular 600D is still better, you do not neccessarily need a full spectrum mod.


  • Monkeybird747 likes this

#22 Monkeybird747

Monkeybird747

    Vostok 1

  • *****
  • topic starter
  • Posts: 145
  • Joined: 31 Aug 2017
  • Loc: Louisville, KY

Posted 13 January 2018 - 01:49 PM

That's looking better!  Yeah, its a bear of an image to work on. I'm in camera shopping mode. Its a flown arrow:)

 

Had I used the exposing techniques suggested by Jim above, and maybe gone 1600ISO, this might have been workable. The bottom line is there is much better equipment available for not that much more $$, and considering I got this one for free, its not really heartbreaking. My wife saw me working on the image and when she finally got my attention she asked "what are all those black lines?". I told her there might be an issue with the camera, and it was taking me a lot of time to fix the image. She said "you need a new camera". Strike while the iron is hot.

 

Soooooo...ASI1600mm cooled and filter wheel, the color version cooled, or a dslr that will plug-n-play with my current setup? I have some thinking to do. There is a 1600mm cooled with electronic filter wheel (but no filters) in the classifieds now for $900. Sounds tempting.

 

MB



#23 Monkeybird747

Monkeybird747

    Vostok 1

  • *****
  • topic starter
  • Posts: 145
  • Joined: 31 Aug 2017
  • Loc: Louisville, KY

Posted 13 January 2018 - 01:55 PM

Some more feedback:

 

there is a lot of good in your data:

- you got excellent tracking, stars are circular and there is virtually no movement between the subs

- your temperature management is exellent, one sub 0°C, all others 1°C

- your darks perfectly match in temperature one -2°C, all other 0°C ( I use exiftool -CameraTemperature *.CR2|grep Temp in the Linux console, available for Windoze as well)

- you flats are quite nice, nothing to complain

- a lot of small stars to be seen

 

What I did and what I found:

- going thru the subs with blink the background is getting brighter and brighter and brighter. Maybe clouds or dust? Finnally I used only the first 10 lights to avoid shot noise from the bright background.

- You have a satelite trail in sub #4 at 20h57m50s, could be removed during image integration using Winsorized Sigma sigma high = 4.3

- I applied ABE twice, first irst subtraction, second division.

- went thru the standard process of Warren Kellers book Inside Pixinsight: Color Calibration, SCNR, Lin Denoise

- then used a range mask to get the black lines in the background and removed them by histogram transformation

- main stretch was done by masked stretch with standard settings

- fought the lines again protecting stars with a different range mask and build a custom kernel for morphology: a vertikal line 11x3 pixels and applied a median

- nonlinear denois, not much

- fiddeled quite a while on local histogram equalization to enhance some structure that distracts the eye from the bloody black lines

- color saturation of the blue/teal reflection stuff

- color saturation of golden stars to get some other color than blue in the image

- some saturation boost with stars under the mask

- another range mask for the very bright stars that were pink: simply desaturated with curves tool

- for the thumbnail above I applied a gaussian convolution prior to downsample

 

I must admit, I was unable to get the dark lines out of the blue nebula. Without the ever raising background all subs could have been used and a lot more faint reflection would have been there.

 

An astro cam with a filter wheel is by no means a guarantee for perfect images. Depending on the cromatic aberrations of your scope and how parfocal the filters are, they might introduce color fringe. I am still waiting for the repair of a mono mod DSLR and I will try to combine the unmodded OSC and mono L and/or Halpha. Whatch my posts and sooner or later you will find that (unless the camera is broken again).

Use a different DSLR and life will be much easier! Don't frustrate yourself with a shisho process and start with good data. A regular 600D is still better, you do not neccessarily need a full spectrum mod.

That's great feedback. I followed a pretty similar workflow, but you did some stuff I want to try. I noticed the brightness in Blink and came to the determination that M45 was tracking from the rural side of me, toward the city side as the night progressed. The gradient fits with the location of city lights.

 

I will probably go the dslr route to get up and imaging sooner, but its fun to think about fancy filter wheels and autofocusers wizzing away through the night. I've seen a lot of nice images with the 600d.



#24 entilza

entilza

    Gemini

  • *****
  • Posts: 3207
  • Joined: 06 Oct 2014
  • Loc: Oakville, ON, Canada

Posted 13 January 2018 - 01:56 PM

Could you please upload an integrated stacked image? Thanks!

#25 the Elf

the Elf

    Ranger 4

  • -----
  • Posts: 351
  • Joined: 06 Sep 2017
  • Loc: Germany

Posted 13 January 2018 - 07:20 PM

Here is a share of the masters, integrated lights and my post processing steps:

 

https://my.cloudme.c...keybird747s_M45

 

Monkeybird, some general facts to think about when choosing a camera:

- a DSLR works without a PC and without extra battery and it has got a display

- astrocams need a PC

- cooled astro cams consume a lot of power so you need a big battery or mains power

- an OSC either astro or DSLR takes all colors at a time. When clouds roll in you have what you have

- a mono with filter wheel takes one color at a time. When clouds roll in you might be missing one or two channels

- an LRGB (+Halpha)- workflow is time consuming. You need flats for each filter, need to integrate each filter seperately, combine RGB, process L and combined color seperately and combine both

- good filters are not cheap. Bad filters are frustrating.

- with a mono + filter wheel you can do narrowband, ideal in light polluted cities, but the artificial look might not be to your taste

- there are beautiful mono images. You can use a mono camera with the L-filter only if you like b/w.

- even uncooled astro cams may be better than a DSLR

- an astro shutter is very slow like 1s shortest exposure. This makes flats a bit more complicated: you have to reduce intensity by ND foil or some sheets of paper

- the sensor must not be larger that your scopes image circle (I could not find any information how large it is for your scope, but as we can see it covers APS-C)

- if the image circle is large enough for full frame you can work with a reducer and make it faster like f/5 instead of f/7

- the sensor may be (much) smaller than the image circle but all the light is lost

- noise comes with small pixels. You have about 1500 e- per square micron. SNR is square root of electron number. Large pixels have a better SNR. (dark current and read noise adds to it)

- any decent astro cam can be binned to create larger pixels. But beware! Make sure the ADC can deal with it. Some cannot use full well when binned.

- if you have a bright sky background most noise might be shot noise and cooling is not offering a benefit. Cooling helps very much at a dark site.

- don't buy a too old CCD. There is a thing called anti blooming gate. If you don't have that you have white lines instead of the black lines.

- if you buy a mono modded DSLR (bayer pattern removed) the micro lenses are missing and it is by far not as sensitive as expected.

- there are Halpha filteters for DSLRs as well but the low quantum efficiency, the red filter and the fact that only one out of four pixels is red makes this a questionable encounter

- if you shoot Halpha or O-III you need long exposure times (20min+). This requires a good mount and good guiding and a lot of time.

- there is no one astro camera perfect for DSO and planets. If you do both and have high expectations you need two cameras.

 

Strike while the iron is hot, but do it wisely! In your case of light pollution I recommend you discuss with the experts how much noise is shot noise from the sky and how much is dark current, reduced by cooling. You can calculate shot noise as stated above. SNR is different for each pixel! The darker ones have a lower SNR. Dark current should be documented in the camera data sheet. It roughly doubles with each 6°C raise in temperature or is reduced to 1/2 with each drop of 6°C. We don't want your wife ask you "where does all the noise come from though you have a cooled cam?"

 

the Elf




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