Am I expecting too much?

[ncwolfie]

I've been observing for years but have just recently gotten serious about imaging...and with mixed results. Some hit and some miss, mostly miss. Two constants that seem to give me trouble are keeping the image noise down and producing a well defined image (they are in focus but the detail is just not normally sharp when viewed full resolution).

I am realistic about my setup, it is fairly meager compared to most here, but yet I have seen others produce seemingly better results. It's my feeling that I am missing something in the processing aspect. I continue to review video tutorials concerning Deepsky stacker, Photoshop astro techniques etc., but feel I have a wall.

I mostly concentrate on deep sky.

As I said, my optical setup is fairly basic. It is however pier mounted and is as follows. Celestron EdgeHD8 on an Advanced VX mount. I normally use an edge focal reducer for deep sky. My alignment is pretty good I think but I do not currently have an autoguider (that is coming soon). So with that in mind, I rarely push any exposures beyond 2 minutes, normally 90 seconds. Normally I shoot at 800 to 1600 iso, I use BYEOS  and a Batinov mask for focusing precisely.

My camera is a Canon EOS T4i DSLR, unmodded so I know I am missing some reds. But colors are not as big an issue as noise and a lack of what I perceive as sharpness. Maybe I am expecting to much?

I know that images are unique and processed differently, but I have a loose standard protocol I use, based on Doug Germans DSS video tutorials.

It is as follows:

Normally use the default settings in DSS to stack then, go directly to Photoshop.

1.Convert to 16bit, then local adaption to exposure and gamma.

2. Crop just a small outside edge of the image out, to get a good dark edge.

3. I was using a plugin called gradient exterminator, but didn't see that much difference so dont currently use it. So now run "HLVG" plugin filter and sometimes Solar Screech plugin "Colour Corrector"

4. Create level adjustments layer and stretch to a low mid range 105-127

5. Create normally 2-4 curves adjustments layer and adjust each one a small bit to bring out detail, keeping a nice smooth curve.
6. When satisfied, flatten the image.

7. Copy background layer and do match color adjustment. Sometimes use
8. Copy previous layer and run noise filter as needed.

9. Copy previous layer and run a sharpen filter (unsharp mask). I have been using smartshapen lately as it has seems to give me better results than unmask sharp most of the time.
Sometimes I reverse the order of 8 and 9.

That's my basic run through. After that I run different adjustment layers I feel are needed if any, mostly hue & saturation, or contrast or color balance or mixer and sometimes a photo filter. Not all of these normally just what I think helps.

Below is an image of ngc 7331 I took the other night. The image was stacked in DSS (19 lights, 10 each, darks, flats and bias) (5x 90 second exposures @1600 iso, 14x 90 second exposures @ 800 iso) and processed in Photoshop as listed above. As a smaller image it doesn't look so bad, but a closer look shows more noise and reveals no real definition of edges. Is there something I am missing in my processing, should I be producing a better image than this even unguided short exposure, or am I in fact expecting to much?

The full size image can be viewed here at astrobin as well as another image of the veil nebula which looks ok as a thumbnail but bad at full resolution as well

http://astrob.in/213175/0/

http://astrob.in/212804/B/

Any tips or advice would be GREATLY appreciated.

Thanks ,

Randy

[Cozzy]

Randy, I'm sure others will come along soon and give you  more detailed comments.

your comment ref:( It's my feeling that I am missing something in the processing aspect.) is likely true you are missing data, lots and lots of data. A few minutes of images is just not going to cut it, you should be looking at hour/s worth depending on the subject. 

Your work flow is fine as long as it works for you.

hope this helps, keep going.

kind regards

Tim

[pyrasanth]

Get the book The Deep Sky imaging Primer (Bracken) this will tell you what you need to know & yes you need more data. When I was working on my last project the core of M31 I stacked 180 2 minute exposures for the L channel alone and that is through an 11" F2.2 RASA.

[abodee2]

I'm still just learning PS and very new, but i had a go at just simply removing that gradient just for practice from the image you posted. i'm using annies astro actions plugin which I've found cheap and useful. There will no doubt be some very helpful people that certainly have a lot more experience than me!

[ericxbenedetti]

You need more data, that includes more lights more darks, more flats, and more bias frames. My first few images were like that, maybe 30-45 minutes worth of light exposure and 10-15 frames of each calibration. Now I try to shoot 2 hours worth of exposure at split into a couple different ISO's, and I do 30 or more of each calibration frame. Pictures have turned out a lot better.

[Zakalwe]

There's a lot to be said for where you are at...you seem to have a lot of the basics sorted. Focus looks good. Tracking looks good. The star shapes look good. So I would say that you have a great foundation to build on.

Now you need to get data-lots of it. Good quantities of good quality data will make the processing so much easier. Trying to drag detail that's not there will just frustrate you.

Personally, I would get into autguiding as soon as possible. Your discarded subs will decrease and you can start going for much longer exposures, which will make the processing even easier. And again- data, data,and more data. I have a very rough rule of thumb- with my sky conditions I aim for about 8 hours total integration time. That makes my data pretty much "stretch proof"- i.e. I can stretch it enough to get at the faint stuff without the histogram breaking down.

Working with a long focal length scope like the SCT is never going to be easy, but I reckon that you're doing fine. Just concentrating on gather a few hours worth of data before you even think about processing it.

[Garethr]

+ 1 for the above,

It's all about the amount of data ( too a certain degree) without guiding you will always struggle believe me I've tried.

There are occasions when loads of shorter subs work but note the word loads.

Guiding and long steady subs with minimal discards not only helps with the final results but avoids the depressing feeling of deleting poor subs after being up all night

Gareth

Your on the right track though so keep going looking forward to your next image.

[ollypenrice]

I'm astonished that at this focal length (or more accurately pixel scale) your tracking is as good as it is. However, round stars are not an indication of good tracking. The reverse is true: oval stars are an indication of bad tracking. But randomly indifferent tracking will produce round stars because they will have been equally smeared in all directions. You need to autoguide. I'd advise an off axis guider for this scope.

More data and longer subs, as everyone is saying. They're right. Very rarely do I post an image with less than 8 hours and single panel images often run to over 20 hours.

When you sharpen I wouldn't rely on a high pass filter to select what sharpening is retained. It will sharpen the stars as well - which is a bad idea.

I sharpen a bottom layer after excluding the stars ad then use an eraser on the top layer to pass only those details I want sharpening. Only very small parts of images usually benefit from sharpening but, few as they are, they make the whole image look sharp. And this way you never sharpen the noise!

Olly

[gnomus]

I am certainly no expert and I hesitate to offer an opinion.  However, I started with kit very similar to yours (an Edge HD 8 on a Celestron CGEM mount) and I offer my experience, for what it is worth.  I would say that I did not get as good results as you are getting (nowhere near in fact) so hats off to you.  Everyone else has commented on the amount of data required.  The more data, the more the relative noise decreases and the more can be pulled out of the image.  I don't think there is much merit in taking exposures at 1600 and 800 ISO.  This is one of the many areas where knowledge of 'normal' photography may be getting in the way of astrophotography.  (This happened over and over again to me.)  My advice would be to stick with either 800 or 1600 (I don't think it will make much difference).

You are making it very hard for yourself trying to image with such a long focal length scope.  Things started to pick up for me only when I finally accepted this fact (I resisted for some time) and bought a cheap refractor (an ED80 - what else) with a focal length less than a third of that of the Edge (and a price less than a third of the cost to boot).   Nevertheless the ED80 started at f/7 (compared with f/10 for the Edge) and could be reduced to f/6.3.  I was pretty much where you were before getting the ED80 - noise, smeariness (if that is a word) and great difficulty in processing.  For myself, I felt I needed to learn to walk before I could run.  I know of course that I am advocating further expenditure, but we spend a lot of time out there gathering data and our time is precious. 

Guiding is something that I also found helped enormously - but the real 'improvement' came when I started to make things easier for myself. 

Good luck