M106 and NGC4217 and others

[Tommohawk]

Evening all. I had a bash at M106 last year and was quite pleased with the results, but not realising that NGC 4217 was a bit of a looker, I inadvertently chopped it in half. So I had another go the other night, and quite pleased with the outcome although not sure M106 has quite as much detail as before - maybe the sky not quite so good. I think this framing works as a pair just about.

Packed the refractor away, and back to the 200P for this. First time using the Newt with the new Astronomik deepsky/ L3 filter set - seems to perform similar to the ZWO filters perhaps not surprisingly. Worth noting that these filters don't seem parfocal even with the Newt, though this could be down to the coma corrector optics.

Used NINA which has become my default and generally works well, except that my filter sequence went a bit weird and I ended up with a double ration of green. Guiding was good generally - it always seem to be better with the Newt which is bonkers given its 5 times the weight of the frac. Maybe something to do with the area of sky I'm looking at. 

The strangest thing was - not a drop of dew anywhere to be seen. By 1am it was all bone dry so I let it run overnight and it was still dry at 5am. Guiding log shows something ghastly happened at about 4am - probably the wind. 

I think I need to rethink my gain settings. I've always used 139 for LRGB but the LUM seems noisy even with 200+ subs. Also I processed in APP and used auto quality for integration - this is great but I can never be sure what its actually done, ie how many subs its used/ignored. Not sure if there's a way of knowing?

The other thing I did different was to not use any flats. I combined RGB in APP rather than PS, and then used the excellent LP tool to correct the vignetting.  This does a great job of keeping the backgound neutralised and avoids any gradient/cast.... I think!! My optics are pretty clean and I cant see any bunnies etc. Maybe not looking hard enough! 

So 200P with SW 0.9CC, ASI1600 cool, NINA, APP, PS.   217 x 60s LUM, 79 x 60s RED, 217(!) x 60s GREEN, 72 x 60s BLUE (RGB binned 2 x 2) All at -20 deg C.

Happy to hear any  feedback/comments, thanks for looking.

[Tommohawk]

Hi folks and thanks for all the likes!

Just FYI, I had another look at this the following day and found a way to reduce the noise by tweaking the stretch settings in APP manually, rather than just using the drop down options. The difference isnt massive so I wont repost, but the take home for me is that doing as much processing in APP as possible makes sense not least because I think at this point its still 32 bit data. 

[tooth_dr]

55 minutes ago, Tommohawk said:

Hi folks and thanks for all the likes!

Just FYI, I had another look at this the following day and found a way to reduce the noise by tweaking the stretch settings in APP manually, rather than just using the drop down options. The difference isnt massive so I wont repost, but the take home for me is that doing as much processing in APP as possible makes sense not least because I think at this point its still 32 bit data. 

That’s good to hear, I do like APP. Did you run star colour calibration?  I only ask because the image has a pink tinge to it 

[Tommohawk]

1 hour ago, tooth_dr said:

 I only ask because the image has a pink tinge to it 

Aaah thats interesting. One of the issues I have when processing is that my red/green colour vision is a bit pants. I typically run the PS eyedropper thingy over the background to check it, and I found that APP did the background neutralization stage nicely, according to this check method. So I'm surprised you say there is a pink tinge - you mean the background, or the stars? I did try the star colour correction but it seemed to do some odd things to the star colours - I tihnk it seemed to desaturate the yellows from memory. But maybe thats how it should be - I'll look again.

[tooth_dr]

6 minutes ago, Tommohawk said:

Aaah thats interesting. One of the issues I have when processing is that my red/green colour vision is a bit pants. I typically run the PS eyedropper thingy over the background to check it, and I found that APP did the background neutralization stage nicely, according to this check method. So I'm surprised you say there is a pink tinge - you mean the background, or the stars? I did try the star colour correction but it seemed to do some odd things to the star colours - I tihnk it seemed to desaturate the yellows from memory. But maybe thats how it should be - I'll look again.

It looks pink overall, particularly noticeable in the galaxies (looks like the yellows are pinkish).  I’ve attached my image as a comparison, but I think mines just slightly on the green side 

 

[Tommohawk]

ok great thanks for that - got my hands full ATM but will play around with it later.

 

[Tommohawk]

Right I had another look and certainly if I crank up the saturation I can see the halo of M106 is very pink, as are some of the other pale galaxy areas. But if I do star colour calibration it really does something horrible - all the blue goes, and the whole thing has an obvious yellow/green haze. I tried randomly fiddling with the slopes but couldn't really get the colour back to anything like normal.

One thing about APP which drives me bonkers is that the sliders, slider labels, and slider controls aren't segregated,  so its hard to see which slider to actually move! I might post something in the APP suggestions area. But maybe its just me!

Going forward I'll have to figure a way of tweaking the colours even if its just a subjective local adjustment in PS. I don't like doing local adjustments though because that always seems wrong ie rigging the data. 

Edited April 13, 2021 by Tommohawk
typo

[tomato]

If you are combining the channels in APP, you can adjust the individual channel sliders in the combine RGB tool on the left hand side until you get the desired colour balance then save this file.

I would agree with @tooth_dr, your original looks to have a pink hue overall. Having said that you have some great detail, and the outer arms of M106 are very well defined, something I struggle to achieve.

Great framing also.👍

[Tommohawk]

3 minutes ago, tomato said:

If you are combining the channels in APP, you can adjust the individual channel sliders in the combine RGB tool on the left hand side until you get the desired colour balance then save this file.

I would agree with @tooth_dr, your original looks to have a pink hue overall. Having said that you have some great detail, and the outer arms of M106 are very well defined, something I struggle to achieve.

Great framing also.👍

Thanks for that, very constructive! I realised I could change the multiplier value when combining but was reluctant to fiddle with it. In theory given I had a triple ration of green I was expecting to have to dial that down. But as you say that's probably the best stage to tweak it - I'll have another go. Right now I'm playing with some limited data from Markarians Chain - it certainly suffers the same pink hue, so looks like its just the way this filter set works. Once I've got a formula that works hopefully it will be good all - I'll have another twiddle!

 

[Clarkey]

If you balance the multipliers in the RGB combining as suggested then you may still find if you calibrate the star colours the whole hue of  the photo can change. Personally I find APP too blue once the star colour calibration is carried out. However, you can use the sliders in the star calibration tool to adjust the ratios of the different colours moving away slightly from the 'standard model' used in APP. Generally I now leave the RGB ratios proportional to the number of subs and then 'tweak' the colours in the star colour module. I find this gives then best 'natural' look to the stars and the galaxy.

Great image though.👍

[Tommohawk]

2 hours ago, Clarkey said:

If you balance the multipliers in the RGB combining as suggested then you may still find if you calibrate the star colours the whole hue of  the photo can change. Personally I find APP too blue once the star colour calibration is carried out. However, you can use the sliders in the star calibration tool to adjust the ratios of the different colours moving away slightly from the 'standard model' used in APP. Generally I now leave the RGB ratios proportional to the number of subs and then 'tweak' the colours in the star colour module. I find this gives then best 'natural' look to the stars and the galaxy.

Great image though.👍

Great thanks for that. I experimented with these APP setting yesterday doing Markarian's Chain, but TBH the data was pretty ropey and not really a fair test. I'll revisit M106 and see what I can achieve. Ideally I would find a formula that works for my filter set that I can apply to all images. Due to a capture glitch I had way more green subs, but this didnt make for an overly green image. So not sure that using the number of subs as a guide to RGB ratios would be effective with my setup.

What doesn't help is that I'm experimenting with using LP correction to resolve all vignetting issues, and not using any flats. This works remarkably well, but the RGC combine is done prior to LP correction so when doing RGB combine there are huge gradients such that judging colour balance is tricky. Maybe I'll have to get this done best possible, and then do the final tweak in star colour correction as you say.

[Clarkey]

7 hours ago, Tommohawk said:

 there are huge gradients such that judging colour balance is tricky.

You can remove light pollution and gradients on the individual stacked images before combining. This may make the colour balance easier. I think you can redo the gradients on the stacked image too (but I'm not 100% sure on that).

 

7 hours ago, Tommohawk said:

So not sure that using the number of subs as a guide to RGB ratios would be effective with my setup.

Maybe the file data on the stacked images contains the total duration? Not sure. I generally have similar amounts of RGB so the colour balance is normally neutral. However, on some targets I have used 4 or 5 to one in the colour balance, so again I don't think there are hard and fast rules.

[Tommohawk]

1 hour ago, Clarkey said:

You can remove light pollution and gradients on the individual stacked images before combining. This may make the colour balance easier. I think you can redo the gradients on the stacked image too (but I'm not 100% sure on that).

Yes, and that's what I used to do - but a bit laborious having to do multiple boxes in the LP for each channel. Way easier to combine first then do LP routine all together. However, what I am looking at now is doing as you say and doing LP on each channel first and then combining just in order to get a formula which works. Then in future use that formula  in the RGB combine, and then do LP all channels together - if that makes sense!

 

1 hour ago, Clarkey said:

...so again I don't think there are hard and fast rules.

Agreed - a bit of experimentation required I guess

Edited April 15, 2021 by Tommohawk