Julian Shaw's sombrero - with a surprise!

[ollypenrice]

Firstly, see what you think of the image. This is a crop to close in on the galaxy - or galaxies. There are lots!

Personally I love it. I can say that because all I did here was the post-processing of the data collected by Julian Shaw from his New Zealand observatory. This was from a six inch refractor and Atik 460 mono. I'd say that was darned good going from a 6 inch scope. (TMB LZOS 152 apo. A good six inch scope!  ) His sky was also excellent, SQM 21.95 to 21.98. Mount was a Mesu 200, guided by small guidescope.

Now for the big surprise. This was taken at F15 using an Astro-Physics Advanced Barlow. It does rather knock some popular assumptions about F ratio on the head, no? So the focal length was 2280mm and the pixel scale 0.41"PP.

The image also shows what an incredibly versatile photographic scope a large refractor is. It can close in like this on a pretty small galaxy using a small pixel camera or it can do the whole of M42 and the Running Man with a full frame chip.

While I don't know the total integration time (Julian will doubtless come in on this), here's another surprise. Determined not to blow the core, he shot this in three minute subs. So that's F15 and 3 minute subs. Shouldn't be possible, should it?

I have never enjoyed a data set more than this one. Many thanks to Julian for bringing it along.

Olly

 

[gnomus]

An absolute stunner, Olly.

[StargeezerTim]

That is a stunning image!

[Knight of Clear Skies]

Very interesting, I've been wanting to try DSO imaging with a barlow at some point. Would be interested to hear what the total integration time is.

DSO images are often heavily stretched, which hides how great the difference in brightness really is. I've noticed that going from, say, f2.8 to f4 (halving the light gathering) doesn't make as much difference as you might think.

[Knight of Clear Skies]

At the other end of the scale, I also like this Sombrero from the 30-second (single exposures or subs) challenge:

 

[ollypenrice]

5 minutes ago, Knight of Clear Skies said:

At the other end of the scale, I also like this Sombrero from the 30-second (single exposures or subs) challenge:

 

Absolutely.

Olly

[simmo39]

another wow olly.

[Petergoodhew]

F15 .... 3-minute subs .... sounds quite amazing.  I would have expected more like 10-15 minute subs at that FL.  The result is stunning.

Maybe I should re-think my imaging plans!

[Thalestris24]

That's a smashing image, Olly. Probably a testament to having lovely dark skies as much as anything! That barlow is pricey... I've toyed with the idea of trying out a much more modest 2" ED barlow on my 115mm APO just to see if I can pick up some of the smaller targets.

Louise

[juliangeorgeshaw]

- Julian here. The integration time was about 7.5 hours total of which about 5 hours was luminance.

- I tried processing the data myself (with PI) and made a hash of it. Olly very kindly offered to demonstrate what he could with the data and PS. As expected he's done a fantastic job, much better than I ever could and yet... And yet to my debauched PI-eyes it could be improved with a little less colour saturation and a bit more sharpening to bring up the pie crust . To my eyes it's a little soft. I'm still a newbie and this is just my opinion (and monitor), I'm sure others would say "natural" rather than "soft". (Don't worry Olly, no doubt my taste will mature .)

- Imaging with a Barlow? This prompts the question: What's the point? Can a Barlow increase resolution? If not, why not simply take an image without the Barlow (and without all the problems of long focal length imaging) and blow it up? Of course a Barlow cannot increase the resolution *of a telescope*, resolution is limited by aperture. But we are talking about the resolution of the total system = scope + camera. Take a look at this from the Astro-Physics website (scroll down to Advanced Convertible Barlow):

 http://www.astro-physics.com/index.htm?products/products

I took a few subs of the Sombrero without the Barlow and I seemed to get a similar but less pronounced improvement (A-P used an 8600 chipped camera [5.5micron pixels] and I was using an Atik 460EXM [4.54 microns]). 

- I think the improvement in resolution with the Barlow is real and/but it's due to the Barlow effectively shrinking the pixel size. (What does this say about the conventional sampling arguments which say that one should use big pixels for long focal length and that 0.41 arc secs/pixel is too small?). Or perhaps it's just that round stars simply look better than square stars even though both are false representations of what should be points of light? In any case I'm a believer in the A-P Barlow!

- Could one get similar results with a different Barlow? I doubt it (though someone will prove me wrong ). Why? The challenge at 2.4m focal length with DSOs as targets is as much mechanical as optical. For an extra $50 I got the 2.7" diameter A-P extension tube into which one can screw the Barlow, so the Barlow only has to support its own weight. Conventional Barlows are long skinny things which must support the weight of the camera and filter wheel. This works famously well for planetary targets but it is asking for trouble with (relatively long-exposure) DSO targets. (In the attached photo, the Barlow is actually inside the tube labelled "A-P Advanced Barlow".)  Note that the A-P Barlow does not claim to cover big chips. To cover chips you will need a mirror scope I think. But they are lots of trouble and I hate diffraction spikes .  Pricy? $250 for the Barlow + $50 for the special extension tube + postage, customs & VAT is not cheap (esp. at post-Brexit vote exchange rates). But in my experience, unlike a lot of other kit, Astro-Physics kit always works (probably because the guy who makes it also uses it).

- I had intended to OAG but my adapter was wrong ?so I used a little Borg 50mm guide scope with a screwed down fixed focus, mounted on the tube rings. (A-P's Roland Christen is a fan of guide scopes for refractors but is adamant they should be mounted either directly on the scope or using dedicated rings, not the scope rings. I tightened everything up and got away with it.) I used PHD2 for the guiding and a Mesu mount. (This was my NZ Mesu, I have another one in France. I had motor problems with both of them but now they are fixed, enfin, they both work very well ).

- The light weight and high sensitivity of the Atik 460/EFW2 combination helped. (8600 chipped cameras *are* much less sensitive and IMO it does make a difference.)

-  3 min subs at F15? Yes it's true. I am allergic to cooked cores so if I can see a core in a linear sub I reduce the exposure time. Also I don't like images in which the cores of stars are white and the the colour is only around the rim, I'd rather have the read noise. Finally I was conscious of the hazards of guiding at 2.4m with a piddly guide scope attached to the tube rings so I was inclined to keep the subs short. The Mesu worked fine at 2.4m, lesser mounts might not. But even it was affected by light to medium breezes (the scope is completely exposed and long-ish refractors are a difficult load). However 5 min subs probably would have worked too. In any case, as Nigel's excellent 30 sec subs image shows, M104 though small is very bright. Same applies many other deep sky objects. For example this is an image by the excellent and very helpful to me Argentine astro-imager , Ignacio Diaz Bobillo of another small bright object, Ghost of Jupiter using the A-P Barlow and a 130mm refractor (for a net focal length of 1640mm):  http://www.pampaskies.com/gallery3/Deep-Space-Objects/Ghost-of-Jupiter2

Julian

 

 

 

 

 

 

 

[juliangeorgeshaw]

[steppenwolf]

Quite a remarkable image for such an optical train and all from 3 minute exposures too. I like it!

[Thalestris24]

4 hours ago, juliangeorgeshaw said:

- Julian here. The integration time was about 7.5 hours total of which about 5 hours was luminance.

- I tried processing the data myself (with PI) and made a hash of it. Olly very kindly offered to demonstrate what he could with the data and PS. As expected he's done a fantastic job, much better than I ever could and yet... And yet to my debauched PI-eyes it could be improved with a little less colour saturation and a bit more sharpening to bring up the pie crust . To my eyes it's a little soft. I'm still a newbie and this is just my opinion (and monitor), I'm sure others would say "natural" rather than "soft". (Don't worry Olly, no doubt my taste will mature .)

- Imaging with a Barlow? This prompts the question: What's the point? Can a Barlow increase resolution? If not, why not simply take an image without the Barlow (and without all the problems of long focal length imaging) and blow it up? Of course a Barlow cannot increase the resolution *of a telescope*, resolution is limited by aperture. But we are talking about the resolution of the total system = scope + camera. Take a look at this from the Astro-Physics website (scroll down to Advanced Convertible Barlow):

 http://www.astro-physics.com/index.htm?products/products

I took a few subs of the Sombrero without the Barlow and I seemed to get a similar but less pronounced improvement (A-P used an 8600 chipped camera [5.5micron pixels] and I was using an Atik 460EXM [4.54 microns]). 

- I think the improvement in resolution with the Barlow is real and/but it's due to the Barlow effectively shrinking the pixel size. (What does this say about the conventional sampling arguments which say that one should use big pixels for long focal length and that 0.41 arc secs/pixel is too small?). Or perhaps it's just that round stars simply look better than square stars even though both are false representations of what should be points of light? In any case I'm a believer in the A-P Barlow!

- Could one get similar results with a different Barlow? I doubt it (though someone will prove me wrong ). Why? The challenge at 2.4m focal length with DSOs as targets is as much mechanical as optical. For an extra $50 I got the 2.7" diameter A-P extension tube into which one can screw the Barlow, so the Barlow only has to support its own weight. Conventional Barlows are long skinny things which must support the weight of the camera and filter wheel. This works famously well for planetary targets but it is asking for trouble with (relatively long-exposure) DSO targets. (In the attached photo, the Barlow is actually inside the tube labelled "A-P Advanced Barlow".)  Note that the A-P Barlow does not claim to cover big chips. To cover chips you will need a mirror scope I think. But they are lots of trouble and I hate diffraction spikes .  Pricy? $250 for the Barlow + $50 for the special extension tube + postage, customs & VAT is not cheap (esp. at post-Brexit vote exchange rates). But in my experience, unlike a lot of other kit, Astro-Physics kit always works (probably because the guy who makes it also uses it).

Yeah, probably a waste of time for me to try with my equipment and skies. Just a dream... Sigh

Louise

[juliangeorgeshaw]

I had another look, this time on my on my iPad. Now the colour looks great. On my monitor (a Dell Ultrasharp) the halo surrounding the galaxy looks greenish and the rest of the image has a magenta-ish cast.

So I take back what I said about the colour  .

I would still be tempted by a bit more sharpening (I know, typical rookie's error ).  

[juliangeorgeshaw]

17 minutes ago, Thalestris24 said:

Yeah, probably a waste of time for me to try with my equipment and skies. Just a dream... Sigh

Louise

Hi Louise

I didn't mean to be a wet blanket .

I meant to say that most Barlows are designed with visual use or planetary imaging in mind whereas the A-P Barlow (and in particular its dedicated extension tube) was designed by a DSO imager (Roland Christen). I think other Barlows would work optically but you might have to figure out a way to make the imaging train rigid enough for multi-minute, long focal length exposures. However note that even with a Barlow your 115mm APO's focal length won't be that long so it might work. 

As for the dark NZ skies, yes they certainly helped but, as Nigel's image proves, M104 is a bright object. I would try it and see. I only tried it myself because I admired Ignacio's images.

Julian   

[alexbb]

Lots of details in the pic, good resolution. Very nice!

[ollypenrice]

Julian, regarding the sharpening I would say that, using Unsharp Mask, I have taken this just past the noise floor already. (On the galaxy itself. I haven't sharpened anywhere else and wouldn't.) This technique can take the image no further, believe me. I always try for more but all it did was sharpen the noise. I tend to go just a tad beyond the noise floor under the influence of some good imagers who feel that a little grain in an image adds to its visual depth by establishing an apparent foreground.

However, that's using Unsharp Mask. I think that starting again and using Deconvolution instead might find more sharp galactic detail. I'm not very expert at deconvolution and need to study Warren Keller's opus on Pixinsight.

A bit worryig that some screens reveal problems. My right hand screen has a high gamma setting so I can use it to look for things which don't show on my 'nice' left hand processing screen but clearly this didn't cover all the bases.

I did compare this with the Hubble Sombrero (being a massochist) and I doff my hat to your capture. There are many tiny galaxies discernible in yours in which Hubble finds some detail, but you have 'em!

Olly

[carastro]

Interesting thread, and excellent image, the detail on the galaxy is almost Hubble quality.  Yes I have always been told not to use a barlow because of the focal ratio implications, but this certainly makes me wonder whether using a barlow in a dark site for tiny object might be worth a try.

Carole 

[swag72]

This is a super image and opens up a big debate and makes many of us think! My belief is that this would not be possible in anything other than a dark sky........ I think if you want the best data then you need the best conditions. I am coming round to the idea that there is absolutely NO compromise on dark skies..... if we don't have them (as I don't) then there are things I need to accept I can not do and the only way I can counter this is by shear quantity. 

I am discovering these days just what my minimum limits are for data sets and it is considerable...... That's my compromise.

[Petergoodhew]

11 hours ago, Thalestris24 said:

Yeah, probably a waste of time for me to try with my equipment and skies. Just a dream... Sigh

Louise, I've tried a couple of times with my Altair Astro 115 ED/APO and a 2" ED Barlow - so far without success.  I had a very heavy QSI683wsg dangling off the end of the Barlow - and my whole setup just wasn't solid enough (and not helped by the thumbscrew attachment of the Barlow to the focuser).  I'm curently working on changing to a screw-thread attachment to the focuser, cannibalising by Planostar flattener, which should help - but I suspect I also need to find an approach like Julian has for screw-threading the Barlow too before I can succeed.  Focal lengths of 1600+ are pretty unforgiving!

[Allinthehead]

Stunning. Always a favourite target

[RichLD]

It's all about the skies! My favourite galaxy and a fabulous effort by both parties - bravo!

[sharkmelley]

That's a great result with excellent processing by Olly!

I really do love to read about unconventional approaches to astrophotography so this was a great thread to read.  This approach certainly goes against perceived wisdom but it produced a lovely image and intrigued me enough to do a few quick (and approximate) calculations.  Since noise is always the limiting factor in astroimaging, it's instructive to look at the sources of noise.  

With a pixel size of 4.5microns, F-ratio of 15 and 21.9 mags/square_arcsec skies a 3 minute luminance exposure will probably collect 10-15 electrons from the skyfog.  Skyfog noise is the square root of this i.e. 3-4 electrons of noise from the skyfog.  The read noise of the Atik 460 is around 5 electrons.  The dark current and hence thermal noise is vanishingly small.  So the sensor read noise is the dominant source of noise but not by a huge margin.

Longer exposures would help but you would have to make them a lot longer e.g. 15 minutes or more before the noise difference becomes noticeable in the final stack.  On the other hand, reducing the read noise would make a big difference.  A modern CMOS camera like the ASI 1600MM has a read noise much closer to 1 electron.  So using such a camera the skyfog noise would easily be the dominant noise source, making a visible difference.

With modern CMOS cameras Refractor+Barlow imaging is therefore quite viable for LRGB and one-shot colour imaging from a noise point of view, especially for brighter objects.  I'm obviously assuming that the optics are well corrected so the stars are not distorted.  However it's probably not ideal for narrowband imaging because skyfog noise levels are much lower with narrowband.

The downside is that with such a slow F-ratio it takes a few hours of imaging to gather sufficient data for a good image.  A big aperture reflector (with the same effective focal length) would achieve the same image quality in a shorter time.

All in all, this example is a great demonstration of what is achievable. Thanks to Olly and Julian for taking the time to write it up.

Mark

[Adam J]

5 minutes ago, sharkmelley said:

That's a great result with excellent processing by Olly!

I really do love to read about unconventional approaches to astrophotography so this was a great thread to read.  This approach certainly goes against perceived wisdom but it produced a lovely image and intrigued me enough to do a few quick (and approximate) calculations.  Since noise is always the limiting factor in astroimaging, it's instructive to look at the sources of noise.  

With a pixel size of 4.5microns, F-ratio of 15 and 21.9 mags/square_arcsec skies a 3 minute luminance exposure will probably collect 10-15 electrons from the skyfog.  Skyfog noise is the square root of this i.e. 3-4 electrons of noise from the skyfog.  The read noise of the Atik 460 is around 5 electrons.  The dark current and hence thermal noise is vanishingly small.  So the sensor read noise is the dominant source of noise but not by a huge margin.

Longer exposures would help but you would have to make them a lot longer e.g. 15 minutes or more before the noise difference becomes noticeable in the final stack.  On the other hand, reducing the read noise would make a big difference.  A modern CMOS camera like the ASI 1600MM has a read noise much closer to 1 electron.  So using such a camera the skyfog noise would easily be the dominant noise source, making a visible difference.

With modern CMOS cameras Refractor+Barlow imaging is therefore quite viable for LRGB and one-shot colour imaging from a noise point of view, especially for brighter objects.  I'm obviously assuming that the optics are well corrected so the stars are not distorted.  However it's probably not ideal for narrowband imaging because skyfog noise levels are much lower with narrowband.

The downside is that with such a slow F-ratio it takes a few hours of imaging to gather sufficient data for a good image.  A big aperture reflector (with the same effective focal length) would achieve the same image quality in a shorter time.

All in all, this example is a great demonstration of what is achievable. Thanks to Olly and Julian for taking the time to write it up.

Mark

I agree, was going to ask how long the integration was, but then read further down the page. A bigger aperture Newtonian at the same focal length, hence faster F-ratio would have required a much shorter total integration time. Beyond the F-ratio arguments its a great image!

[wimvb]

14 hours ago, ollypenrice said:

Julian, regarding the sharpening I would say that, using Unsharp Mask, I have taken this just past the noise floor already. (On the galaxy itself. I haven't sharpened anywhere else and wouldn't.) This technique can take the image no further, believe me. I always try for more but all it did was sharpen the noise. I tend to go just a tad beyond the noise floor under the influence of some good imagers who feel that a little grain in an image adds to its visual depth by establishing an apparent foreground.

However, that's using Unsharp Mask. I think that starting again and using Deconvolution instead might find more sharp galactic detail. I'm not very expert at deconvolution and need to study Warren Keller's opus on Pixinsight.

A bit worryig that some screens reveal problems. My right hand screen has a high gamma setting so I can use it to look for things which don't show on my 'nice' left hand processing screen but clearly this didn't cover all the bases.

I did compare this with the Hubble Sombrero (being a massochist) and I doff my hat to your capture. There are many tiny galaxies discernible in yours in which Hubble finds some detail, but you have 'em!

Olly

I actually found these recipes/tutorials of good use

http://www.pixinsight.com/examples/M81M82/index.html

http://www.pixinsight.com/examples/M81M82/index.html

Adding a luminance mask to protect the background, helps keeping the noise down.

The trick is to find the best settings for deringing and regularisation.