The "No EQ" DSO Challenge!

[rotatux]

Hi again, here's my progress on M45.

First image is for reference a session from this year in august, 50km from Paris, but with my EQ mount (sorry, just to show difference with others).
Info: Capture = 19 lights x 30s x 2500iso, 8 darks, Olympus E-PM1 with Skywatcher 130PDS (no CC) on Omegon EQ-300 tracking RA only, no filters; Processing = Regim, Fotoxx, Gimp

Second image is Alt-Az from late november 2016 from Paris suburbs.  Having read the story about dynamic range and sensorgen.info I wanted to experiment with lower ISO.
Not conclusive since 1/ my cam only has 12-bits depth, 2/ a simple contrast filter is not enough to dig faint signal like this from huge light pollution and 3/ the mount wasn't reliable at 30s so not many subs.
Info: Capture = 10 good of 44 lights x 30s x 800iso, 17 NG darks, Olympus E-PM1 with Skywatcher 130PDS and CC on Celestron SLT mount, TS contrast filter; Processing = Regim, Fotoxx

Third and my best image so far is Alt-Az from last 2016 evening at my deep country dark spot.  Still not much structure, but anyway amazed that I can get far better images than with an EQ though high number of subs helps a lot with both noise reduction and signal depth.
Info: Capture = 72 good of 98 lights x 20s x 2500iso, 38 NG darks, Olympus E-PM1 with Skywatcher 130PDS and CC on Celestron SLT mount, TS contrast filter; Processing = Regim, Fotoxx

PS: Ken I know m45 goes high in the night but this was taken early (local 18h30-19h00) about 50-60° alt because sky was already very dark and also somewhat dry and cold (had to keep outside by -5°C).

BTW I found that sensorgen.info must be taken with care: the measures are for a single average shot so especially noise and digital range may not be directly applicable to astro-imaging (because of subs stacking).  The noise level should be a guide on how many subs you need to sort it out; IMO Real good info are the saturation level, as guide to sub length depending on subject, and quantum efficiency (I now lurk towards E-PL5 or E-PM7 :-P).

[Filroden]

36 minutes ago, rotatux said:

Hi again, here's my progress on M45.

That's nice progress. Comparing the EQ with the latest AZ version it's interesting to see you have no star colour in the EQ image but the stars in the AZ image are very nice, showing a good range of white through to orange. I think the EQ image suffers most from limited total exposure time. Under 10 minutes of total time is not a lot, even for M45. You have twice as long on the AZ shot and I think the shorter exposures actually help retain the dynamic range.

I think there is still either a background gradient from light pollution or from vignetting. You don't say if you have used flats. I don't know processing software for Linux but I know both Photoshop (GradientXTerminator) and Pixinsight (DynamicBackgroundExrtactor) are both very good at removing background gradients.

Altitudes between about 30 and 60 are usually good - high enough to be above the worst of the light pollution but low enough not to suffer excessive field rotation. My concern with M45 at the moment is that it is also south. The combination of a southerly azimuth and high altitude means rotation is higher and you'll have to crop more. I think at my latitude (about 5 degrees north of you) I can still get 30s exposures in the area of sky.

I think to be fair to EQ mounts, a true side-by-side comparison between a good polar aligned, unguided EQ mount and a good aligned AZ mount would favour the EQ mount. Even unguided I would expect the EQ mount to be able to track for 60-120s without the issue of rotation and cropping. If you had 20x120s images from the EQ mount and compared them with 80x30s images from the AZ mount, you might notice less noise in the AZ image (from the extra stacked images) but you'd have better signal in the EQ image (from the longer exposures). It would be close though on a target like M45 with only 40 minutes of data. Now on a target like NGC1333 or M78, the EQ mount will be far better! The if you guided the EQ mount and were taking 2x1200s images...

I'd never claim we can beat EQ imaging, just that we don't have to be so far behind

[The Admiral]

3 hours ago, rotatux said:

Hi again, here's my progress on M45.

First image is for reference a session from this year in august, 50km from Paris, but with my EQ mount (sorry, just to show difference with others).
Info: Capture = 19 lights x 30s x 2500iso, 8 darks, Olympus E-PM1 with Skywatcher 130PDS (no CC) on Omegon EQ-300 tracking RA only, no filters; Processing = Regim, Fotoxx, Gimp

Second image is Alt-Az from late november 2016 from Paris suburbs.  Having read the story about dynamic range and sensorgen.info I wanted to experiment with lower ISO.
Not conclusive since 1/ my cam only has 12-bits depth, 2/ a simple contrast filter is not enough to dig faint signal like this from huge light pollution and 3/ the mount wasn't reliable at 30s so not many subs.
Info: Capture = 10 good of 44 lights x 30s x 800iso, 17 NG darks, Olympus E-PM1 with Skywatcher 130PDS and CC on Celestron SLT mount, TS contrast filter; Processing = Regim, Fotoxx

Third and my best image so far is Alt-Az from last 2016 evening at my deep country dark spot.  Still not much structure, but anyway amazed that I can get far better images than with an EQ though high number of subs helps a lot with both noise reduction and signal depth.
Info: Capture = 72 good of 98 lights x 20s x 2500iso, 38 NG darks, Olympus E-PM1 with Skywatcher 130PDS and CC on Celestron SLT mount, TS contrast filter; Processing = Regim, Fotoxx

PS: Ken I know m45 goes high in the night but this was taken early (local 18h30-19h00) about 50-60° alt because sky was already very dark and also somewhat dry and cold (had to keep outside by -5°C).

BTW I found that sensorgen.info must be taken with care: the measures are for a single average shot so especially noise and digital range may not be directly applicable to astro-imaging (because of subs stacking).  The noise level should be a guide on how many subs you need to sort it out; IMO Real good info are the saturation level, as guide to sub length depending on subject, and quantum efficiency (I now lurk towards E-PL5 or E-PM7 :-P).

That's showing significant improvement, and with only 24 minutes total exposure. I managed to stack just over an hour's worth on my M45, but I had discarded about a half-hour's worth to get that. Call me picky! There's a bit of structure there though. My tired old eyes aren't as all-seeing as Ken's, and I didn't spot any gradients, but it is worth noting that the nebulosity extends out quite a way and I'd be uncertain whether any gradient is down to that rather than processing. See Rogelio Bernal Andrea's superb image, which can be found here.

You might find the following thread informative re. sub length and sub number. https://stargazerslounge.com/topic/245183-to-stack-or-not-to-stack-30-x-1s-1-x-30s/#comment-2664662 .

As for field rotation, here is a chart plotting the sub-duration for a 0.1° rotation with direction, if you are not already aware of it.

Cheers, Ian

[Filroden]

7 minutes ago, The Admiral said:

My tired old eyes aren't as all-seeing as Ken's, and I didn't spot any gradients

I'm not convinced there is a gradient but there seems to be a central brightness that seems unrelated to the reflection nebulosity, something that might be there if there were not flats applied and it was then stretched. But as you say, there is a lot of dust in the region and I've struggled to see what is dust and what is just gradient before (often incorrectly erring on the side of dust).

[rotatux]

4 hours ago, Filroden said:

I think the EQ image suffers most from limited total exposure time. Under 10 minutes of total time is not a lot,

Agreed. Even the second one at 5mn total time is better.

The lack of color may also be caused be me (I don't recall for sure now) using gamma-based stretching on that one rather than my now usual brightness stretching. That former tends to burn colors and make all grey/white or the dominant color, so I rarely use it now.

4 hours ago, Filroden said:

I think there is still either a background gradient from light pollution or from vignetting. You don't say if you have used flats

Ok I must remember about explicitly adding "0 flats" in my info then next time

But I see no gradients; I see a decreasing noise zone around the general shape of the cluster, but I believe it to be dust as Ian pointed out. Anyway I didn't stretch it too much / nor shaved the background to keep the noise at a pleasing level (I like this grainy look of amateur astro images).

In fact each time I have tried flats they have ruined the result, introducing awful gradients rather than removing them. But the flats I made on that occasions allowed me to check I have less than 0.5% variation in corners relative to center; So I don't bother with them anymore, I rather use darks which are necessary and efficient for my setup. Remember I have a 4/3 sensor so 80% smaller than Canon APS-C, that may be why.

4 hours ago, Filroden said:

I'd never claim we can beat EQ imaging, just that we don't have to be so far behind

Right. However better do alt-az imaging when like me you're unable to PA an EQ correctly and get more than 30s subs from it, at least alt-az gets the subject centered...

58 minutes ago, The Admiral said:

You might find the following thread informative re. sub length and sub number

Thanks Ian, I have already read it as an sidestep from this thread when I saw your previous mention of it (also mentionned in the 130pds thread).

1 hour ago, The Admiral said:

As for field rotation, here is a chart plotting the sub-duration

Already see this, however I prefer a PDF cited somewhere in this this thread -- search for "Field Rotation V3.pdf", because it gives graphs per observing latitude, is based on absolute pixel displacement limit (though for APS-C), and gives the formulas to compute everything if not satisfied.

[Nigel G]

Its raining here and been cloudy for a few nights so I had another bash at processing my M 42 with modded camera.

I have deliberately started to over saturate the colours to see what would happen, and reduced highlights.

I'm quite amazed what colours have come out, I've seen so many different M42's I really don't know what the natural colours are but green has started to be quite noticeable in the middle.

Nige.

[happy-kat]

@rotatux startools works on Linux

[Filroden]

1 hour ago, Nigel G said:

I'm quite amazed what colours have come out

Very ethereal! I wouldn't be surprised to see that win the Turner prize  I'd call it "Butterfly: are you pulling my wing?".

[Stormbringer]

very nice and very different 

[The Admiral]

2 hours ago, Nigel G said:

Its raining here and been cloudy for a few nights so I had another bash at processing my M 42 with modded camera.

I have deliberately started to over saturate the colours to see what would happen, and reduced highlights.

I'm quite amazed what colours have come out, I've seen so many different M42's I really don't know what the natural colours are but green has started to be quite noticeable in the middle.

Nige.

I agree Nige, it is quite a challenge to decipher what the 'correct' colours should be, but according to Ivo the core is meant to be green, one of the few occasions such a colour appears in the universe. Strangely, even my subs of 2s don't reveal any green, even though they are not overexposed. Go figure!

Anyway, a nice rendition Nige, even if the colour scheme isn't quite my taste . I'm leaning towards calling the ST colour module the 'gaudiness control', although 'control' might be a misnomer!

Ian

[rotatux]

From what I read here: https://en.wikipedia.org/wiki/Orion_Nebula#Coloration

The center should be blue-violet because of the hot trapezium, and only some highly ionized O++ rich dust around ("high vacuum" regions) should show green. I also believe the green should come combined with other colors and come as yellow (with H) or cyan (with S), as no region should be pure oxygen !

Nice image Nige, very artistic and enhances many features !

[rotatux]

15 hours ago, happy-kat said:

@rotatux startools works on Linux

I know I know but it's only for image processing right ? But then since I'm quite satisfied with Fotoxx for now and it's free, I'm inclined to keep it like this.

Will reconsider the day I will manipulate more-than-16-bits images of course. Or maybe Gimp 2.9+ will do the trick (I may build it to check -- needs free time).

[Nigel G]

Clear but freezing night last night, I got some imaging before the moon came up.

2 targets, Jellyfish nebula and a double cluster I'm looking for the name atm.

Having problems with the Jellyfish, very bright green and blue pixels cover the image. Must be some bad sub's or a DSS issue.

The double cluster, 30 minutes of 30s and 20s. flats & bias. 150p, modified 1200d with CLS filter. DSS & ST.

EDIT. M35 & NGC 2158

Nige.

 

Edited January 15, 2017 by Nigel G

[Filroden]

I also managed a few minutes out last night (literally a few minutes). I got the SCT set up and focused to within a few miles of infinity (maybe a few thousand). So I had another pop at M1:

This is only:

B: 45 seconds (3x15s) (Although I had many more they just would not align with any other sub)

G: 105 seconds (7x15s + 1x5s)

R: 75 seconds (4x15s + 3x5s)

L: 415 seconds (26x15s + 5x5s)

So I'm pretty amazed there is even an image, let alone colour in the image  And please don't comment on the focus/vibration/tracking trails! If I'd discarded any more subs from the night I wouldn't have any.

[Filroden]

1 hour ago, Nigel G said:

The double cluster, 30 minutes of 30s and 20s. flats & bias. 150p, modified 1200d with CLS filter. DSS & ST.

I do love clusters. I've been so focused on nebula recently that I've not really looked at clusters but they do make nice images. 

[Nigel G]

38 minutes ago, Filroden said:

I also managed a few minutes out last night (literally a few minutes). I got the SCT set up and focused to within a few miles of infinity (maybe a few thousand). So I had another pop at M1:

This is only:

B: 45 seconds (3x15s) (Although I had many more they just would not align with any other sub)

G: 105 seconds (7x15s + 1x5s)

R: 75 seconds (4x15s + 3x5s)

L: 415 seconds (26x15s + 5x5s)

 

 

Ken, amazing for such short time, good detail in there. I got a multi coloured blob from 60 minutes of 30s on M1.

Nige.

[Filroden]

1 hour ago, Nigel G said:

Ken, amazing for such short time, good detail in there. I got a multi coloured blob from 60 minutes of 30s on M1.

Nige.

The individual subs were sometimes shocking. Blobby stars with slight bulges in which ever way the wind/passing cars would send the scope. I guess it shows the power of stacking. As each bad sub looked different to the next, they only really contributed where they were more consistent. I just wish I could figure out why my blue subs wouldn't align. The process could detect the stars but they could not be matched to the stars in any of the other subs (L, R or G) or even the combined images. They visibly look fainter in blue than the other filters but other than their actual size, about the same number of stars could be seen in R G and B so they should have matched. A puzzle for another day.

[Filroden]

17 hours ago, Filroden said:

 A puzzle for another day.

Whilst I still haven't solved the puzzle I was able to manually align 11 blue subs and redo my image. Not that 11x15s B subs adds much more data but I think the colour balance might be better (if not a little too red?). In theory, I could now go back through all my subs as some of the RG and L subs also didn't align. However, no matter what I do, this will never be a great image with such little exposure time and the problems I had getting focus and from the wind.

Version 1.1

Edit:

Version 1.2 with the green slightly boosted (to balance the red/purple hue). It probably needs more noise reduction, particularly the background.

Edited January 16, 2017 by Filroden

[rotatux]

13 hours ago, Filroden said:

However, no matter what I do, this will never be a great image with such little exposure time and the problems I had getting focus and from the wind.

There's also the fact that with your SCT, the longer focal length makes imaging more sensitive to (bad) seeing, hence probably more difficult to focus. What focus technique do you use ? I recall you are software-based, but do you also use a bahtinov ? My f/4.5-5 130PDS is way easier to focus than my f/12 Mak, and long time since imaging with the latter but I remember using a bahtinov mask a few times and it was nice.

Rather than boosting the green I would rather put in a bit less blue. Anyway I like both. It's a pity this target requires a long focal / SCT or MAK to get a big enough image, otherwise it's really nice.

No clear sky here since end of december and still eager to try my new christmas 2" UHC filter on heavy LP. And as time passes, my targets are getting out of my (limited) sight

[Filroden]

2 hours ago, rotatux said:

What focus technique do you use ? I recall you are software-based, but do you also use a bahtinov ?

I have a mask for the SCT but now that I use SGPro I find it much more accurate to use the focusing tool within the software. I take continious 2s 2x2 binned exposures and SGPro analyses the star widths in the entire image (HFR). I adjust focusing until this number is minimised (always making my last adjustment in the direction that lifts the primary mirror). However, the focusing is very course. Even a slight turn can make a large difference. And even on a very heavy tripod, the whole system takes a long time to settle from any vibration. If I was serious about using the SCT I would add a new focuser to the visual back that had a finer control and could eventually be automated (which would reduce vibrations enormously).

3 hours ago, rotatux said:

Rather than boosting the green I would rather put in a bit less blue.

It's a tough decision to make on this target. I've seen RGB images that show quite a range of colours. My histogram in v1.2 has a balanced red/blue mix with the green slightly lower than both (it was much below in v1.1). The background in both v1.0 and v1.2 is more neutral than in v1.1 which had a red/purple cast (because I'd overdone the green reduction). So I think the stars and background are correct. However, I have so little data that even 15-30s extra in one channel is making a big difference.

I have more time later today so I may revisit the other subs I took (I took 60x1s of each channel as well as more 5s and 15s images) to see if I can align them and get a closer balance of total exposures in each channel. I may even have to dust off DSS. One of the problems I find with the SCT images is that there might only be 50-60 stars, so finding alignment is harder. With my refractor, I more often have 800+ stars so there is more to match.

3 hours ago, rotatux said:

And as time passes, my targets are getting out of my (limited) sight

I know that feeling. And the targets moving into my sight now are the galaxies in Leo and Virgo which either need the SCT to get real close and personal or a slightly longer focal length on the refractor (probably 800-1000mm).

[The Admiral]

3 hours ago, rotatux said:

No clear sky here since end of december and still eager to try my new christmas 2" UHC filter on heavy LP. And as time passes, my targets are getting out of my (limited) sight

Tell me about it! I've had one good night after Christmas and another early in the New Year, and that's it, else the Moon's in the way. Even then, I almost put off imaging the Orion region, thinking I might wait until it appeared at a more sociable hour! Glad I didn't now. I think in the UK at least, you've got to grab any opportunity when it appears. That's something Ken and Nige seem up to doing.

Ian

[lune lupine]

Hope this is the right thread - just picked up this lens (samyang 16mm F2) : no packaging and a bit scuffed on outside but lens etc looks ok in daytime - I presume this lens is ok for astro photography? hoping to take to dark sky area later in year.

Edited January 16, 2017 by lune lupine

[happy-kat]

Yep this is the right thread if you do not have a wedge on your mount, I can't remember :-)

[The Admiral]

2 hours ago, lune lupine said:

Hope this is the right thread - just picked up this lens (samyang 16mm F2) : no packaging and a bit scuffed on outside but lens etc looks ok in daytime - I presume this lens is ok for astro photography? hoping to take to dark sky area later in year.

Looks reasonable from this review http://www.lenstip.com/380.1-Lens_review-Samyang_16_mm_f_2.0_ED_AS_UMC_CS_Introduction.html , but you might want to stop down to say f/4 to sharpen up the edge resolution. It'll be pretty wide field. A quick Google for "Samyang 16mm astrophotography" I found http://www.astropix.com/HTML/I_ASTROP/EQ_TESTS/rokinon_16mm_lens_test.html and it looks promising.

Looking forward to your results.

Ian

Edited January 16, 2017 by The Admiral

[Shaun_Astro]

I've been following this thread for a while, there are some amazing pictures here!

I've been struggling with my mount trying to take photos for a while and have finally got a reasonable image with my DSLR and a 58mm lens on a Star Discovery mount. Previous sessions were ruined by bad focus, the mount being set to the wrong day and the battery having ran out a bit, causing star trails imperceptible at the time.

I only managed 60x30s subs and 10 darks before the lens froze over. Processing done in in light room and photoshop. The original had very bright streetlights in the bottom right which I managed to eradicate thanks to a tutorial vid by Doug German, a great channel!  I'm getting a 130pds tomorrow so hopefully should be able to produce better results in the future.

 

 

Edited January 18, 2017 by Shaun_Astro