Heart in Ha with Canon 200mm

[Adreneline]

This is 26 x 180s of Ha taken with the Canon 200mm + ASI1600 + Astronomik 6nm Ha filter - unguided.

Testing out my revised setup - the WO mount can be made to snuggly fit the 200mm as well as the 135mm

Couldn't resist adding a starless version.

Adrian

Edited October 13, 2020 by Adreneline
Starless added.

[nfotis]

interesting setup - doesn't the EAF tilt the lens as it rotates the focusing belt?

Since I have lots of good Canon lenses, I am interested in your experiences with EF lenses and astro cameras.

N.F.

 

 

[Adreneline]

42 minutes ago, nfotis said:

doesn't the EAF tilt the lens as it rotates the focusing belt?

Well the tension is at a minimum and the amount of rotation is also minimal. The WO tube ring combined with the ZWO camera support make the whole thing very rigid all the more so because I have a reasoable amount of seperation between the two support.

I have glued a toothed belt to my Samyang to pretty much negate backlash and provide excellent repeatablility but I've not been able to bring myself to take that step with my Canon lens yet.

[ollypenrice]

Great! Personally I'd be inclined to combine the starless and starry versions in search of one with nice tiny stars. (I'd do it in Ps by pasting the linear original on top of the final starless in layers and changing the blend mode to lighten. I'd then gently stretch the top layer till the stars started to appear and stop when I had them at the desired size. Only the stars will appear in the blend if you choose blend mode lighten.)

Olly

[Adreneline]

3 minutes ago, ollypenrice said:

Personally I'd be inclined to combine the starless and starry versions in search of one with nice tiny stars.

Thanks Olly.

The plan was to grab some OIII and SII when the UK weather permits. Now Starnet is very accessible within PI using a Mac I've started partially stretching the image, removing the stars, and then stretching the starless image a little less than it could go. I then combine the starless versions. I've found putting the stars back proves to be tricky though because the finished result tends to look like the stars have been 'dropped' onto the image and any sense of 'depth' is lost. I've tried various methods including your luminosity layer approach but it never seems to look right.

I think I must be doing something wrong!  

Adrian

[ollypenrice]

2 minutes ago, Adreneline said:

Thanks Olly.

The plan was to grab some OIII and SII when the UK weather permits. Now Starnet is very accessible within PI using a Mac I've started partially stretching the image, removing the stars, and then stretching the starless image a little less than it could go. I then combine the starless versions. I've found putting the stars back proves to be tricky though because the finished result tends to look like the stars have been 'dropped' onto the image and any sense of 'depth' is lost. I've tried various methods including your luminosity layer approach but it never seems to look right.

I think I must be doing something wrong!  

Adrian

Hmmm, it has worked quite nicely for me but processing is a funny business. Maybe the choice of curve is critical. If the stars look 'dropped in' (and I know just what you mean) it might be worth trying a hand-shaped curve which rises quite steeply at the bottom and flattens to a straight line early.  This should give softer stars.

Olly

[Adreneline]

20 minutes ago, ollypenrice said:

it might be worth trying a hand-shaped curve which rises quite steeply at the bottom and flattens to a straight line early.

Haha! Now I find Curves are tricky things as well! I have to say I've never really embraced Curves in PS - I think I'm intimidated by them.

Just to be clear are you suggesting applying a curves transformation to the image once it has been 'recombined' or just to the star layer before recombining?

Adrian

[ollypenrice]

28 minutes ago, Adreneline said:

Haha! Now I find Curves are tricky things as well! I have to say I've never really embraced Curves in PS - I think I'm intimidated by them.

Just to be clear are you suggesting applying a curves transformation to the image once it has been 'recombined' or just to the star layer before recombining?

Adrian

I'd be applying the curve to the still-linear original while it was pasted onto the processed starless. In blend mode lighten it won't be visible at all because none of it will be brighter than the processed layer below. If you now stretch it in Levels it will produce a pure log stretch, steeper at the bottom and leveling off higher up. You won't see this curve but it's the same shape as the curve in PI's histogram transformation. If this is giving hard-edged stars which look dropped in it may be because the curve is too steep where the outer edges of the stars lie. I'd call the curve below a 'soft star curve' because it will brighten the fainter outer parts of the stars by far more than it will brighten the cores. It would need several iterations and it might not work at all! 

I'm a great fan of the custom curve. For example, many people are scandalized by my Ha curve, which is the opposite of the above. It gives very high contrasts which I think are advantageous when you're aiming to combine Ha with red. In fact it isn't 'my' curve at all, I learned it years ago from a video by a famous imager, and I know Tom O'Donghue uses it as well.

This gave me...

I found this aggressive stretch gave a far more punchy result than a log stretch which, while looking nicely natural, seemed a bit soft and featureless. You should have a play with curves, and with pinning them at a fixed point and then changing the curve above and below the fixed part. That's superbly useful.

Olly

Edited October 14, 2020 by ollypenrice
typo

[Adreneline]

Wow Olly! That is really helpful and thank you for taking the time to explain the process.

I will give it a go and see how I get on.

Adrian

[Adreneline]

5 hours ago, ollypenrice said:

You should have a play with curves, and with pinning them at a fixed point and then changing the curve above and below the fixed part.

I've had a go Olly using your method and this is the result ....

 ...... using this Curve

The Background is my starless, stretched image - the top layer is my linear original image (DBE'd) and then stretched with the curve shown above.

When I came to select a blend option I used Lighten.

To my eye it seems to have worked inasmuch that the stars don't look 'dropped on' to the image.

What's the verdict?

Thanks again for your help.

Adrian

 

 

[ollypenrice]

Looks good to me.

Olly

[Xiga]

Hi Adrian

I too have played around with trying to add stars back in over a starless layer and i feel your pain, it's really not easy to do well. I've had most success doing it on galaxy images, where you have a central contained object of interest, surrounded by background sky and stars. For images of extended nebulae, it's much, much harder to get right. I also don't like my images to have too much star reduction, so for these types of images i've reverted to good old star reduction, but using the one from Images Plus. Mike Unsold has made IP free to use now, and the star reduction process is currently the only thing i use it for. I've found it to be much better than any other type of star reduction i've tried to date (be it Carboni's or Annie's versions, or the minimum filter in PS). The default settings work quite well, so it's a 1-click operation 😀 Here's how i go about it:

1. Open your image in IP and choose Special Functions->Star Size, Halo, Shape Reduction...

2. Hit Apply!

3. Press Done and Save your Image as a New Image. 

4. Close, Re-Open the New Image and Repeat. (I'm not familiar with the UI so i'm sure there is an easy way to do 2 passes without having to close and re-open. If anyone knows please let me know!)

Often 1 pass is not quite enough, so it's worth running 2 or more passes, until you find the one you think has gone just too far, and then apply a lower opacity to it in PS. The routine also seems to apply some sharpening, which you may or may not like (i personally like it). If not, it wouldn't be too hard to mask out the areas of highest contrast. The stars are tightened up quite a bit too, they certainly come out a lot better than Carboni's or Annie's ones. However, if you find they are just a little too tight for your liking, then i've found that running Carboni's 'Less Crunchy More Fuzzy' routine over it at an opacity of ~40-50% often recovers that. 

Anyhoo, i ran 3 passes followed by Carboni's Less Crunchy on your Jpg. The result is below, just in case you find this useful.

HTH.

 

Edited October 15, 2020 by Xiga

[Adreneline]

Thanks Ciarán - that’s really interesting. I tried IP several years back but decided not to buy. I too have tried Carboni’s and Annie’s Actions with varying degrees of success. I have found Annie’s star reduction useful in the past to reduce bloat on B and OIII images but since I fitted a Hoya UV filter on the lens (Canon and Samyang) all the bloat problems have been sorted.

I certainly like what you’ve achieved with my image - it seems to sit half way between my original and Olly’s approach and I agree the slight sharpening is beneficial as it helps to ameliorate the undersampling - and lack of guiding.

Thanks very much for the help and advice. I shall download IP and give it a go.

Adrian

Edited October 15, 2020 by Adreneline