Horse Head in Ha

[Paul Haese]

This is part of a much larger project but thought some of you might find it interesting to look at the Horse Head nebula in Ha light.

This is 13 x 10 minutes.

There is still some noise left and as such I think it needs another 6 or seven subs to smooth it out. Perhaps a lot more.

Your thoughts are welcome.

[multiweb]

Amazing details. Processing looks a tad over sharpened to me though. Nice shot. :thumbsup:

[AlexN]

I'd hate to say it, but imaging at F/8 through a Ha filter, you really may want to consider 20 minute subs... The area below the curtain is still pretty dim in your shot, and there is lots of detail to be had down there...

On a brighter note, the detail you've captured in the horse head and in the curtain, including the streamers, looks very - very nice! Looks like you've got that guiding and collimation under control! Was this guided with the OAG or with the ED 80?

Good going mate.. More exposure required, but a fantastic start!

[Paul Haese]

Alex, the actual stacked image was quite a lot brighter but also a little too noisey for my liking. I did a little curves work just to hide the noise a little.

The trouble with longer subs is that something can go wrong and you lose that sub. I could do longer subs in my luminance layer though and that would brighten up the eventual image.

The guiding was conducted through my TSA102 which is now riding on top of the RC via a set of Parrallax Rings.

Marc, yeah probably is a little over sharpened. I did not do much too it though.

[pmrid]

In many ways, this image is very close to the one that won the Royal Observatory Greenwich's Astro Photo of the year for 2009 (with one obvious difference being this is in Ha and it was colour). But the interesting thing it deminstrates is the significant enhancement in contrast and hence in detail that Ha seems to make possible. The winning image was praised for its "delicate wispy detail around the horse's head and mane". Having compared the two in that limited respect, it seems to me that Paul's image is even more subtly detailed and executed. Makes me want to have a play with narrowbanding.
Peter.

[TrevorW]

The new camera seems to be producing the results Paul

[spacezebra]

Stunning image Paul.

Cheers Petra d.

[Paul Haese]

Here is a reprocessed version

[strongmanmike]

Yes that's pretty good Paul, a little dark perhaps but that was your intenton by the sounds of it and besides, it is a no brainer to alter if you want.

Mike

[AlexN]

The repro is brilliant mate.. A little noise is readily acceptable in my opinion, especially if it allows more fine detail in an image...

Great work indeed.

I think you really need to spend more time refining and using the OAG built into your QSI583. The chances of something going wrong in 20 minute subs are minute providing that your guiding is tight and well controlled... I've shot nothing but 20 minute exposures for the last few months worth of imaging and not lost a single sub... Granted, im working with 1/5th the focal length, but I'm also working with a mount thats 1/10th as capable as yours...

[Paul Haese]

Alex, yeah that is true. My guiding is getting pretty tight now and I should give this a try. Working with the OAG is a little more difficult because the stars are so much dimmer and a little harder to find too.

[rat156]

Hi Paul,

The horsey is very dim and could do with the 20 minute subs mentioned here.

I hate to say it but the star shapes are a little elongated and the circular diffraction pattern around the brighter stars is not symmetrical. I'm not sure what this means as I'm no optical expert, but it must be affecting the star shape off centre.

I thought that the OAG in your system is in front of the filters? Why are the stars dim? Being where it is there should be plenty to pick from. I have the advantage of being able to change the orientation of the guider to the main chip, but I try not to change it too much as that means I have to recalibrate it.

Nice detail though. I'll have to wait another month or so until Orion rises out of Melbourne's murk.

[Paul Haese]

Stuart, there is elongation only on the upper right and lower left of the image at full size. The off center ringing is odd as I have not had this before myself. Finding the sweet spot has been a pain in the butt, but I have had it before with the 40D which has a bigger sensor and I will find it again. My adapter for the flattener is due to arrive from the States next week. This will mean I can sort collimation and have a truely flat field. I might have tightened up the collimation screws a little too much too. Still quite happy with the detail of the head despite some "small" issues.

[Paul Haese]

Sorry neglected to answer the OAG question.

Yes the OAG is in front of the filters but it is still quite dim with regard to stars. Sometimes I can use the OAG and sometimes I cannot; it just depends on the field.

[rat156]

What are you using as a guide cam?

Cheers
Stuart

[telecasterguru]

Paul,

I like the repro, it is very, very good. The contrast is noticeably cleaner.

Frank

[Paul Haese]

Stuart I am using the star shoot autoguider from Orion. It is very sensitive but I think the f8 optics is what is the problem really.

[rat156]

Hi Paul,

The f-ratio of your optics has very little effect on the intensity of point sources. For diffuse objects, then the image is spread over more pixels for longer FL, and unless you have a huge mirror, then that generally relates back to f-ratio. I'm not going to start another "f-ratio myth" argument, but most of this doesn't hold for point sources, where it is the size of the Airy disk which matters.

Sometimes I struggle to find a decent guide star, but I'm usually looking for sub 1 second guide exposures to run the AOL. If I were doing normal guiding then I'd be going out to 2-3, maybe even 5-6 seconds, anything faster than that and you'll end up "chasing the seeing".

You're on a G11 aren't you? If so then get the Ovision worm, run PEC training (I used pempro, not necessary, but worth every cent, getting cheaper as the $A goes up). Then you'll easily be able to guide at up to 10 second guide exposures.

You really have to use OAG at the focal length we're imaging at, even the slightest flexure stuffs things up completely.

Cheers
Stuart

[bloodhound31]

I'm not really a fan of B&W images, but Paul, this is stunning work mate. Well done.

Baz.

[Paul Haese]

Thanks guys.

Stuart I am using an EM400. However, I am sure I don't have flexure. However there is something going on with the way I have the camera squared to optical path. Possibly even that final touch of collimation.

In the images below you can see that center top left and bottom right are all circular but top right and bottom left are definitely elongated. Well at least that is what I am thinking. Can you see what I am talking about? Personally I reckon it needs to go just a little to the right with the collimation.

[gregbradley]

Your images continue to impress with this setup Paul.
A little bit more of fine tuning your setup and it should be a reliable
imaging machine.

Elongation in one corner and not the others to me indicates sag/tilt.
Collimation would be all corners wouldn't it?
I got a touch of this myself on the weekend as I had to space out an adapter
that screwed in too deeply into my filter wheel causing some slight tilt.

Are you using all screw adapters?

Modern cameras plus filter wheels plus OAG plus adapters adds up to a lot of weight which is a strain even on an expensive Tak focuser.
So if there was any slop in the focuser to start with it will add to the problem.

You can always use a guide scope. I use an Astrotech 66ED scope and an SBIG ST402ME (nice camera but cheap fittings like the power connector which often is "touchy" and the autoguiding cable won't disengage from its socket oh for a FLI guiding camera!). It works fine although I only do 10mins with it.

As far as not being able to have the image brighter because of noise;
you clean up background noise by either:
1) light gaussian blur on the background only.
2) use a plug in like Noise Ninja which does a really nice job and hard to beat
3) Use a Noel Carboni action

You can even blur the noisy areas manually by using the blur tool and rubbing it on the offending areas. Works well with noisy rim areas of dim galaxy shots.

Just moving up the black point is another way but you lose most of your detail as a lot of faint stuff is towards the black end of the histogram.

When you image LRGB on this use either no filter or a clear filter for the luminance. No need for UV/IR coated luminance filter which reduces a small amount of transmission on this dim object.

Greg.

[Garyh]

That`s a nice HH Paul.
Would love to see it again with more subs! :thumbsup:

[Paul Haese]

Good points Greg. I checked my images from 50 seconds (the ones I used to setup framing) and the stars are the same shape. That certainly is not flexure, so it means something is not right on the image plane.

I just have to go through and check a few things now. One for certain is to ease off the collimation screws a bit and recollimate but keep the screws loose. I am thinking that the secondary might be warped a little, so just need to check that.

Greg I am not using screw adapter between the focusor and camera connection. The OAG is incorporated into the camera, so I need to get a handle on this I think. Other people have got that sorted. The spacer that I put in to get focus might also be affecting the stars too. It might need shimming.

Everything is an evolution and I thank you all for your comments and suggestions.

[gregbradley]

Ashley at Precise Parts in Miami is great for creating a custom made anodised aluminium adapter. He could make one for your camera to fit your focuser. Putting a nice camera like that in a screw tightened focuser is asking for trouble.

Yes 50 seconds shows no differential flexure but it would also still show tilt which affects any exposure no matter how short.

A heavy camera in a screw tightened focuser is unlikely to be square.

But as you point out it may well be something else but even if collimation is out tilt still appears to be present. Perhaps both problems exist.

Greg.

[AlexN]

Funny you should mention that Greg.. I was just about to say the same thing regarding Ashley at Precise Parts.. I received a few adapters a couple of days ago that he made up, all are brilliant.

I figured much the same thing.. Heavy camera setup with a 2" compression ring style focuser was probably a nightmare in the making... I got him to make solid screw in fittings for damn near everything, and the one push fit part of my whole setup now is the 2" nose that slides into the focuser, its 3.1" long so it sits firmly against the wall of the focuser tube and is very unlikely to sag about... the OAG screws onto that, flattener screws onto the rear of the OAG, CFW screws to Flattener and Camera screws to CFW... It all looks solid as a rock... If anything the only thing I see giving me grief is the actual focuser its all being put into :D

[Paul Haese]

Well as it turns out guys great minds think alike. Precise parts has machined up an adapter from the camera to the flattener. It is on its way to me as we speak. The problem with the stars is only recent. One of my early images (lagoon nebula) taken with the QSI showed only mild elongated of stars (see attached) caused from field curvature. It would seem therefore logical everything stems from something I have done in recent months.

Can you see what I mean?

Yes another screw in adapter is needed I think to attach the focusor to the camera, but I will need to think on this a little first as to how this will be done.

[gregbradley]

Interesting. The only thing I might add there is sometimes a sag or tilt problem can be slightly erratic as different angles of the scope for imaging different objects can cause different stress.

Hopefully your adapter sorts it out.

Greg.

[Paul Haese]

Greg its a matter of narrowing down the culprit. It could be something moving at one time and nto at another. Time will tell I suppose. Thanks.

[Phil Hart]

i am a fan of black and white images, particularly of this region so i really like this shot. and this is a great focal length for the region too - lots of promise for really attractive details (especially the complex area under the horse's chin). i agree that sharpening artefacts look like they're just starting to show at the top of the head.. but this could perhaps be masked out rather than reducing sharpening across the whole image.

on the topic of longer subs, i'm prepared to be a little contentious and say that i think you might not notice much difference going to 20 min subs, but then i'm biased by too much experience with DSLRs with low read noise. i know that with DSLRs, the read noise and therefore sub-exposure length becomes irrelevant even with greater than 1-2 minute subs. but don't take my word for it.. have a look at the middle row of images in this test:

http://www.philhart.com/canon_test

now i know that CCDs have higher read noise than DSLRs, but it's surely lower with modern cameras than it used to be. and the signal is low with HA so that also increases the desired sub-exposure length.

so in theory, yes longer subs will be better, but in practice i think the difference is so low that i'd go with paul on this one.. you'd only need to lose one 20 minute sub due to problems and you'd have lost more than you'd gained. ie a little increase in total exposure time is worth more than the increase in sub length?

having said all that, i'd love to see an objective test that does show a difference between 10/20 minutes subs with a modern ccd.

phil

[Phil Hart]

having spent the weekend surrounded by people imaging the horsehead, i should retract my comment about the sharpening artefacts.. the brightening around the edge of the horsehead is clearly a real feature. not the first time i've mistaken features for artefacts.. i was once cloning out bok globules from a nebula before i realised my mistake :)

anybody listening should also probably take my comments about sub-exposures with a grain of salt too. if you're chasing really faint stuff then photon statistics start to count against you with shorter subs.

i'll go and crawl back under my rock and hand back over to the experts..

[AlexN]

I did that the first time I imaged M16 Phil... The Bok Gobules just looked like random sharpening artefacts or maybe dust on the sensor, so I cloned them out :D haha.. Yes the brightening behind the horses head an mane are real, and seen in many of the best photos of the area..

As for exposure lengths, I did think it was strange when you mentioned that 20 min subs wouldnt make a big difference over 10.. I thought "How does having twice the data per sub exposure not make a difference to the depth of an image? 60x10min subs may be 10 hours data, just like 30x20min, but surely 30x20min subs would give a deeper image"

How did you go imaging this weekend? got anything to post...

[Paul Haese]

There is an idea getting around the traps now that the background should have a weight of 1000 ADU. Most of my current images have something around 380. That probably would indicate that I need to capture my images for longer, but I need to balance this against saturating my stars. In reality I should have gone with a pixel size and well depth quite a bit larger.

So for now I need to keep the subs around 10 minutes. I think 20 minutes will saturate and flood the CCD too much.

Phil, that bright patch behind the head often looks like an artifact. There might be a star behind the head section that is illuminating the surrounding gas just enough to birghten that spot.

[AlexN]

Paul, for Lum/Ha data I wouldnt worry about saving the stars from saturation, worry about that in your RGB frames.. worst comes to worst take a series of 3 or 4 minute shots just for preserving star colours then selectively mask them if in you need to...

Lum and/or Ha should be purely as much data as you can get in order to give the target brightness and detail.. preserving star colours is another matter all together.

[Manav]

Beautiful image quite majestic in B&W

[jase]

Looking good Paul. The first and second images, I think the highlights could be made to pop a little more, but data looks great. Will be good to see how you handle a RGB blend with such a data set.