M51 - equipment triumphs, processing fails

[John78]

So last night almost everything came together...

Cloud free windless night

Sharpcap polar alignment - easy, good and accurate

Guiding - seemed to be working well < 1" ??? EQ5 mount

Plate solving working.

Shame about the moon

So I took 2hrs of M51 - 5 minute subs 400 iso, EQ5, SW 80ED, Canon 1000D (astromodded)...  So time for the next hurdle - my inability to process the data...

Quite disappointed with my (lack of) achievement so far - the histogram for the data looks weird, but I'm baited for a moon free night now - as I'm sure that will bring drastic improvements to the raw data.  Tonight I'll try to find a target much further from the moon.

I really try and try with star tools but I just get hideously grainy mush - when I first open the FITS file it doesn't look too bad and I get quite hopeful, but by the time I've worked through the tutorials online I can find I get left with rubbish.  That image above is just gimped...

Fits file in case any wizards can make more of the data than me or critique why the histogram looks bizarre...

m51.FTS

Really disappointing is there are 2 or 3 massive blobs of debris somewhere in my image train that I need to get out.

Happy Sunday people!!

P.S - the FF/FR is on the shopping list for next month!!

[alacant]

1 hour ago, John78 said:

try with star tools

Hi. One of the best things I find with StarTools is its handling of noise. Could you post your workflow? But hey, to get any rgb with last night's moon is a credit to you. Cheers and clear (moonless!) skies...

Edit: had a go in StarTools. Had to crop a long way, but it looks OK. Take flat and bias frames too. They will improve the file a lot and get rid of the artefacts you mention. Autodev -crop - decon - contrast - wipe 92% dark anomalies 12 - Develop 80% - colour - mask fat stars - tracking - denoise 8 - repair stars ... That's it.

 

[John78]

I followed the tutorial linked by @bottletopburly

But basically followed it blind - what seem to kill my data everytime is that autodev button.

[alacant]

9 minutes ago, John78 said:

autodev

Try Develop until you get the feel for the autodev big hammer...

[Star101]

I had a quick go using PixInsight. As said above, Bias, Darks and Flats would help. But otherwise its not a bad shot

I cropped it a bit too. Here are both images. 

 

[John78]

There's 60 flats and 10 darks in the stack already, no bias as I forgot - the histogram looked really odd in DSS, I'll try restacking it later see if it improves with different settings.

Your processing skills are miles better than mine, thank you for having a go - I'm sure moonless nights will bring big gains now!!

 

[alacant]

47 minutes ago, John78 said:

60 flats and 10 darks

Hi. Lose the dark frames. With Canon, it's best to use light, flat and bias only. It looks as though there is a problem with the flat frames though. Maybe you rotated or removed the camera after the light frames were taken? HTH.

[John78]

For the flats I just pointed the telescope indoors at a fluorescent white light with a piece of white paper over the front without adjusting the focus or camera - there are a couple of "odd" looking frames among the 60 looking back at them now, probably walked in-front of the telescope when I was packing up

Tonight's plan is to try and get dithering working properly so I don't need to worry about the darks!  Just need to pick a target thats above the tree line.

[John78]

The histogram for the flats looks bizarre - that's whats screwing it up...

I'll try to use my laptop or iPad to shoot tonights flats.

 

Edit : DSS has odd looking histograms, even without stacking the flats - something odd going on ...

[bottletopburly]

4 hours ago, John78 said:

So last night almost everything came together...

Cloud free windless night

Sharpcap polar alignment - easy, good and accurate

Guiding - seemed to be working well < 1" ??? EQ5 mount

Plate solving working.

Shame about the moon

So I took 2hrs of M51 - 5 minute subs 400 iso, EQ5, SW 80ED, Canon 1000D (astromodded)...  So time for the next hurdle - my inability to process the data...

Quite disappointed with my (lack of) achievement so far - the histogram for the data looks weird, but I'm baited for a moon free night now - as I'm sure that will bring drastic improvements to the raw data.  Tonight I'll try to find a target much further from the moon.

I really try and try with star tools but I just get hideously grainy mush - when I first open the FITS file it doesn't look too bad and I get quite hopeful, but by the time I've worked through the tutorials online I can find I get left with rubbish.  That image above is just gimped...

Fits file in case any wizards can make more of the data than me or critique why the histogram looks bizarre...

m51.FTS

Really disappointing is there are 2 or 3 massive blobs of debris somewhere in my image train that I need to get out.

Happy Sunday people!!

P.S - the FF/FR is on the shopping list for next month!!

tried sharpcap last night dammned if i could get it working right wasted all night messing and computer was playing up ,eqmod crashed one of those nights ,what were your settings in sharpcap john ,exposure  brightness etc 

[John78]

6 hours ago, Star101 said:

I had a quick go using PixInsight. As said above, Bias, Darks and Flats would help. But otherwise its not a bad shot

I cropped it a bit too. Here are both images. 

 

Running that through HLVG makes a really quite good image ?

 

@bottletopburly sharpcap just default gains etc.. with 2 second exposures I pointed the rig straight up, Polaris roughly centred in the polar scope, did the first frame it solved straight away rotated and took the second then adjusted as directed and my polar alignment looked super - for me.  I just have an ASI120MM and a 9x50.  Quite surprised when I slewed to M51 I could see it in sharpcap on the guidscope with 2 second exposures, shame it doesn't have bigger pixels really could use it for imaging.

[bottletopburly]

cheers im using a qhy mini guidescope

[John78]

What is the objective lens diameter on that?  Qhy5 even though the same sensor as the ASI looks tricky to make work in all sorts of different programs - Mr Sharpcap @rwg is on the forum so maybe post up your issues in his thread or wait to see if he pops by ??

[assouptro]

Great triumph John! 

i have always struggled with m51 for reasons beyond my control (or my comprehension and or wallet) 

I'd be pretty encouraged by your results and would be praying to the moon gods to hurry up! 

Great image for a moonlit night!

well done 

clear (moonless) nights 

bryan

[ollypenrice]

According to my understanding of the dates, you were shooting this target close to full moon! Luminance may work reasonably well, given a lot of it, but this is not the time of the month for running OSC or RGB.

Shoot galaxies in the dark.

Olly

[John78]

21 minutes ago, ollypenrice said:

According to my understanding of the dates, you were shooting this target close to full moon! Luminance may work reasonably well, given a lot of it, but this is not the time of the month for running OSC or RGB.

Shoot galaxies in the dark.

Olly

Both close in distance target to moon and moon close to full ? I'm itching for an opportunity with no moon.

Unfortunately I'm pretty sure this is the first time I remember 3 days in a row of clear skies since birth and the first truly clear night in 2017 - beggars can't be choosers.

[Starwiz]

10 hours ago, John78 said:

 

 

 

I like your guiding data.  Mine looks all over the place compared to yours.  How are you guiding?  

John

[Oddsocks]

Deleted post, error in original evaluation. please ignore.

[vernmid]

21 hours ago, alacant said:

Lose the dark frames. With Canon, it's best to use light, flat and bias only. 

Hi alacant,

I too am new to this game and was wondering why you say not to use dark frames with Canon? Please could explain why?

Incidentally I also captured M51 for the first time on Saturday night. I only managed 30s exposures with no guiding but was really pleased to see anything :-) Having read this thread I understand now that the bright moon reduced my chance of any RGB!

Thanks

Vern

[John78]

13 hours ago, Starwiz said:

I like your guiding data.  Mine looks all over the place compared to yours.  How are you guiding?  

John

I spent the late afternoon setting up, levelling the tripod and balancing the rig in the garden - then I polar aligned using sharpcap and the guiding came out like that - by far my best results, just default settings really in PHD just calibrated and the started guiding.

 

Here's what HLVG did for the PI picture from above...

 

 

[alacant]

1 hour ago, vernmid said:

not to use dark frames with Canon?

Hi. The impossibility of temperature matching and something in the Canon circuitry dictates that dark frames will not work. I tried using a 15 px dither for the light frames. That, along with flat, bias and DSS's sigma clip algorithm does an excellent job from there on. YMMV but I believe that that's the general consensus amongst Canon users; the best way would be doing a with-and-without-dark-frame side by side comparison. HTH.

[vernmid]

Thanks alacant. I'll try a with and without comparison later :-)

[wimvb]

.. and here's my attempt with your image.

There is quite some noise in your image, most likely due to light pollution. The only way to get rid of this is to take more subs; MANY more subs. 2 hrs of 5 minute subs = 24 subs. I think that with strong light pollution you may have to triple that number. As Olly noted, doing that when the moon isn't showing itself, will give better results.

The background issue you have (shown clearly in @Star101's reply) is optics related. Most of it can be removed with DBE using division as a correction method. There was also some vignetting and a few dust bunnies, indicating that your flats aren't working as they should. The dust bunnies are still visible in the attached image.

PixInsight workflow:

Dynamic Crop

DBE: tolerance 1.7, sample size 21, 30 samples/row, correction: division + normalise

BN + CC, using the four corners as background reference

MLT on Lum, to get rid of some of the noise

Deconvolution with a star mask protecting the brightest stars

MLT on chrominance to reduce the colour noise in the background

Histogram Transformation

MLT on chrominance to reduce some of the colour noise in the galaxy

Curves Transformation to boost colour in the galaxy and stars

Morphological Transformation to reduce the stars

Dynamic Crop to clean up the edges

 

[John78]

Well that's a great result Wim 

Could the camera have rotated the flats, I notice that when they load in ImageEX they are all portrait and the lights and darks are landscape - although they seem to load landscape in DSS could they be upside down?

The LP here isn't too bad (Norfolk) as I'm at my parents, there is an LP filter in the image train also.

Time to go on a dust bunny execution mission....

[wimvb]

AfaIk, any dust bunnies that show in an image are close to the sensor, ie inside the camera. Dust bunnies on other optical parts, such as field flatteners, coma correctors and external filters, won't show up on flats, simply because their shadow is too large. For the same reason, dust on a Newtonian's mirror will never show up (thankfully). I would expect that camera orientation doesn't make any difference, since the dust bunnies rotate with the camera. Also stacking software doesn't care about camera orientation. While camera orientation is saved in EXIF information in the raw files, to my knowledge, DSS (and PixInsight) don't use this information.

Software like ImageEX may read and use the EXIF information when converting the raw image files into something that can be displayed on a monitor.

You could try loading some raw files in a program like Canon's own raw converter, and have a closer look there. If you took the flats at a later time/date than the light frames, dust may have entered the optical system. You may want to leave any optics such as coma corrector or field flattener permanently attached to the camera, if you don't use the camera for other photography.

Btw, the histogram of the original fits file looks ok to me, other than the channels not being aligned. The histogram is from the cropped, but otherwise unprocessed fits. The lower image in the screen dump is the DBE corrected image. (Same screen stretch applied)