Processing Help Please - M83 using a QHY8 and DSLR

[Screwdriverone]

Hi All,

As some of you may know, I am testing out a QHY8 on my scope and am having difficulty processing the FITS subs as well as the RAW CR2 files from the DSLR.

I have uploaded 2 sets of data taken from the same night (first was 20 mins with the QHY8) and then the DSLR was put in to take 20 mins of subs while the scope was pointing at M83.

Vignetting is horrible on both, I am thinking maybe the primary is too low in the tube as I recall setting it right down on the screws when I took out the longer screws before I fixed the DSLR focus problem with a low profile 2" adaptor....? Any advice on this would be good.

So, what I am asking is if anyone can use their skills to play with the data and remove the vignetting, stretch, process and get a result so I can determine the pros and cons of the QHY8 or the sticking with the DSLR moving forward.

I am planning on a light box as well as a peltier coolbox for the DSLR perhaps to remove some of the noise, but obviously, using the DSLR for so long its been easier to capture subs as Backyard EOS is quite schmick and I know what I am doing with ISO and sub length etc....

The files are listed below, anyone who wants to play around with them, please feel free and if possible, document the steps and software used so I can try to reproduce your results at my end.

No flats, or darks are in the stacks or in the ZIP files as I wanted the comparison between the two cameras to be of 20 mins of SIMILAR data without any complications like different darks or bias etc....

Excuse the noise, the dust and the vignetting, I am working on fixing these, the main point of this exercise is trying to process similar data from the two cameras the same way and see which one looks better.

Obviously, I have the DSLR already, so I need to justify the expense of a $1500-$2000 camera based on its results with this side by side using the same scope and same data capture operator.

I look forward to some help and hopefully some results and tips please...

Files:

Zip of 10 x 2 min Canon CR2 data (60mb)
http://dl.dropbox.com/u/70660066/2nd%20April/1000D%2020%20mins%20CR2.zip

Zip of 4 x 5 min QHY FITS data (125mb)
http://dl.dropbox.com/u/70660066/2nd%20April/QHY%20Fits%2020%20min.zip

My stacked tiff for Canon from DSS (92mb)
http://dl.dropbox.com/u/70660066/2nd%20April/m83%201000D%2020%20mins%20no%20dark .TIF

My stacked tiff for QHY from DSS (143mb)
http://dl.dropbox.com/u/70660066/2nd%20April/m83%20QHY%2020%20mins%20no%20dark.T IF

Cheers

Chris

[allan gould]

Chris
I had a go at processing your file. I took thew qhy8 zip and then stacked in DSS. Took the autosave file into PS CS3 and just did levels and curves until it was almost right. The followed this procedure to remove vignetting
(http://www.iceinspace.com.au/forum/forumdisplay.php?f=35), then touched up the dust spots with healing tool and removed hot pixels as well. Highlighted the galaxy and increased saturation by 20 % and that was about it.
Hope the JPEG is OK
I then downloaded the stack from the canon but there was no colour in it. I processed the same in PS as previous image and the second jpeg is the canon result.
You really need a lightbox

[Screwdriverone]

Cool!

Thanks Allan,

Seems you are working on your old camera's data once again!

Thanks for the processing flow too, will give this a burl.

Could you check what settings you put into DSS for me please? I have done that step so many times I am not sure whats the best bet to get to this.

JPG is fine, I am simply seeing whether this data is usable as I get frustrated with noise and after removing vignetting stars go blobby or I get grainy noise etc, so its good to know someone who knows what they are doing can get results.

Cheers

Chris

[allan gould]

Stacking parameters were for DSS
Standard mode
Kappa-sigma clipping
bicubic alignment
Tiff files final format
cosmetic.. detect and clean hot pixels set at 1 pixel and 50% threshold and same for cold pixels

OK?

Noise can be removed with Noiseware (free)

[Screwdriverone]

Thanks Allan,

I assume Kappa sigma was set to 1, with 5 iterations (I have been using this from info gleaned elsewhere)

I appreciate the help.

Have a great Easter

Chris

[allan gould]

Chris, that's correct. Sorry I should have included that information.
Glad that old QHY8 of mine is still going well, my new QHY10 has just gone back for repairs.
And same to you, have a great Easter break.

[Screwdriverone]

Tell me about it :lol:

I think my primary is too low which might be making the vignetting worse. I dont recall it being this bad before when I took some shots of M42 without the coma corrector in play.

I might wind it up and see if the next night I can get out improves it.

Priority one is building a lightbox, I agree.

Thanks

Chris

[Screwdriverone]

I have officially hit a brick wall.

I can use Startools to now remove dust bunnies using "heal" but have been struggling to remove noise and vignetting easily, all combinations make the image worse. grrrrr

I hope someone can help with their processing flow here, I have tried curves like Allan suggested, but photoshop drives me insane and I end up closing it in frustration.

I may end up just giving this all away considering how annoyed it is making me. I can SEE the data I want, I just cant process. FULL STOP.

Most of the day has been spent twiddling, reading, trying things and what do I have in return? nothing.

I used to design multilayer circuit boards with high level cad programs and have used autocad easily, yet I DONT KNOW how to drive friggin Photoshop to get what I want. Even my 17 year old who drives photoshop like a demon, doesnt know how to use the tools to clean up this crap.

its effing pissing me off.

is it really this hard?

[Screwdriverone]

Update:

Anyone who still would like to process my subs is more than welcome, however, Allan has helped me come to the realisation that I need to concentrate on better data by using better tools during capture rather than wave a processing magic wand to try and make the sins evaporate to reveal some phenomenal shot. It aint gunna happen.

I am sorry for the *****ing and moaning about all this. Here is my new plan.

1) Build a light box - eliminate the vignetting issue
2) Stick with the DSLR and forget about a CCD until I have ALL the data I need to process to get a good result.
3) Take darks ( I do this now, but need to use them properly and make sure they are at the same temp as the lights)
4) Build a cooling box for the DSLR using the copper box and peltier design.
5) Get a Hutech IDAS LPR filter to kill the sky garbage that floats over my house.
6) Remove all the dust bunnies.
7) And finally......but probably most importantly.....capture ALL the data I need for a quality shot and then not be impatient to create a masterpiece with only 10-20 mins of subs and expect things to jump off the screen and all fall into place.

Rather than spend $1500-$2000 on a CCD and find myself on another learning curve with still no flats, having to sort out coma again, introduce a new learning curve for capture programs and processing, I should be able to buy the LPR, flats box bits and the camera cool box bits for less than say $400 and then wring the DSLR's neck for all its worth before assembling the data to create some decent images.

Once I have achieved that much, then a newer, cleaner and more sensitive camera will simply make things all that much easier.

Thanks to those who have helped and have attempted to extract something from my subs, I do appreciate the help.

Sanity has returned somewhat.....;)

Cheers

Chris

[midnight]

Chris, I totally agree with this. I'm sorry I can't help contribute to your immediate need but I am following a similar path. Ie, get the most out of my 40D including setup, guiding, dew, lights/darks etc, post processing until I am comfortable with this b4 upgrading.

There are some truly fascinating images with a DSLR out there.

I am a strong believer in grass roots learning. Get a feel for the machine and what goes on underneath;)

Good luck, I agree with your approach and appreciate the frustration. I have just returned home from a long (since Dec) but intermittent stretch away from home so looking forward to getting back out there again.

Cheers,
Darrin...

[Screwdriverone]

Thanks Darrin,

Its quite hard to resist the temptation to upgrade to try and eliminate beginner's sins.

I think its been invaluable to me to have Peter's QHY8 to test out, as the two side by side stacks look quite similar, although cleaner with the QHY8 due to its cooling.

It certainly has helped me see there is no silver bullet that can solve all my questions or fix the missing pieces in my set up or skills.

Good to know I'm not the only one in this boat.

Cheers

Chris

[midnight]

My philosophy on this is simply upgrading to a dedicated CCD with limited skills could be risky financially if you are not out regularly and/or still trying to get some good fundamentals padded down first. This would really only apply if you already have an imaging base (eg a DSLR etc)

When (I hope when and not "if") the time comes and I am comfortable, you never know what might be on the market in 6 or 12 or 18 months time when upgrading.

I must admit though, I have been tempted on many occassions to look seriously at the QHY line. If I can't do good justice to my 40D with the gear and skills I possess, then upgrading imo, would potentially be not an effective use of my money going in too early. Just my view on it.

Good luck Chris!!

Darrin...

[RobF]

Hi Chris

The processing mountain is a fairly steep one to climb early on, but if you have any streak of "computer nerd" in you it is a fun pursuit (it is for me anyway week nights when I can't justify setting up).

The vignetting issues are significant on an SW600 and I've never really succeeded correcting for them well enough with a lightbox. I bought Pixinsight ages ago just for the DBE function, but it's so powerful in so many other areas I think it's worth including in any astrophotography budget. I suppose a larger secondary would be a good investment, but I'm time poor for hardware work having 2 small kids.

I didn't really "cotton on" to the benefit of cooled CCD until shooting the Rosette from suburbia a few years ago. Up until that point satisfying results were possible for almost anything collected with enough care and enough exposure time through a light pollution filter. However the Rosette is quite a bit more diffuse than many other commonly imaged objects, and too my chagrin even with many hours of data, the signal to noise with was quite poor. Fundamentally, a faint object, taken against city skyglow and noisy (summer temperatures) DSLR sensor wasn't good.

In recent years I've been shocked just what is possible from very dark skies with a cooled CCD and good tracking. Only trouble is most of us only experience that once or twice a year if we're lucky. Must of this comes down to what you really want to achieve yourself against your time and budget constraints. What types of objects do you like, how often will you travel, what temps will you set up in?

M83 is a relatively bright object. You might find a fainter galaxy in Leo is a more telling test of QHY8 versus DSLR this time of year. On M83 I would expect similar levels of detail as both sensors have a bayer matrix in front. The main difference SHOULD be background noise and sky detail providing tracking was equivalent in both cases. You really need at least around an hour of data for each to be fair. I noticed Peter Ward commenting in a recent thread that an IDAS LPS isn't really much benefit for a cooled CCD versus DSLR. Must confess I usually still use my IDAS from Brisbane but did my original with/without experiments on DSLR and haven't checked since.

Don't forget to have fun with the journey. We all want a great image early on for encouragement, but each foundation stone is essential - gear set up, skies, exposureme time, tracking, calibration, processing. I still use to be blown away with a noisey DLSR shot of a galaxy I could never hope to observe visually in my old 4.5" Newt :)

[RobF]

I'm guessing these were taken recently with the moon up from the amount of background glow?

[RobF]

Here's my quick effort....
(on the QHY8 data)

p.s. Startools is an amazing bit of kit for the price too I should have said before!

[Screwdriverone]

Hi Rob,

Thanks for giving the processing a go, I have successfully "removed" the dust with heal in Startools, but then have difficulty cleanly removing the vignetting without introducing gradients.....

I hear what you are saying about dark skies and CCDs. Trouble is, I am time poor like you said and dont really feel like spending an hour or more each way packing all things into the car (chair, table, scope kit, scope, tripod, PC, cables, power tank etc etc) and then unpack and reverse it all to find that when I get there, its windy/cloudy/crap seeing/ and so on. OR I have left a bit of kit behind and cant take photos.....grrrrr

So, its the backyard for me for a while, I have been able to master drift aligning with PHD (thanks Trevor Gerdes) and now with my lower profile 2" adaptor, the DSLR is quite coma free, so I have been making progress of sorts since the monsoons stopped.

I was toying with spending $1.5K on extending the decking out the back next to the pool, putting a tilt/slide off roof shed obs there with the scope on a pier and therefore maximise sky time rather than the hour it takes me now to set everything up. This then turned into redirecting those funds into a new CCD, so with Peter's generous loan of his QHY8, I had a chance to try it out side by side with the DSLR.

Then the same processing hurdles came up with either camera, so I have now decided to stick with small steps, such as the light box (and check the primary as well) to fix the vignetting, a LP filter to help the DSLR during glowy nights and also build a copper peltier box to reduce some of the thermal noise so prevalent in the 1000D.

This will be a LOT cheaper and allow me to grab some of the brighter galaxies and nebulae so once the limits from the burbs are reached, I can still venture off to dark skies and/or buy a cooled CCD when the funds are more plentiful. In the meantime, I am in processing school until I can learn the tools inside and out.

I have reset my expectations to a lower level and regained some perspective from some sage advice here and via PM from you very patient lot, so thanks to all who have helped me plan for what the next step should be for me...:)

I am (and probably always have been) a computer nerd, which is why I have been frustrated with my incompetence in driving processing programs to get some results out of what I recognise now is essentially cruddy data. But, its all a learning experience!

Cheers

Chris

[RobF]

At least you've had a look at what's on the other side of the fence though thanks to Peter, rather than having to make a leap of faith. Reckon we're lucky with modern tech at our disposal that there often is a cheaper road that still delivers a significant % of the benefits of spending wads of cash.

At least you sound like you're having fun again :) :thumbsup:

[Screwdriverone]

Yeah, it was a bit touch and go for a while there, I wanted to smash things. 5 days of watching progress bars go nowhere and give no progress on anything other than dust bunnies was tipping me over the edge.

I have had a revelation.....;)

Cheers

Chris

[alistairsam]

Hi Chris
I'm a bit disappointed you've given up on the ccd route but you've got to stick with what works best for you
If you listed all the problems you had with the ccd so far, don't they seem like more of a back end problem, in IT speak?
Eg, the vignetting would require the correct placement of your primary and the correct sized secondary, dust bunnies would require cleaning, coma corrector the right spacer, software issues with win 7 and good capture programs or more know how with the programs and so on?
I guess my background in digital imaging and years of photoshop made my transition to the ccd very easy in terms of processing, so I can see the pros straight away.
I kind of feel you would have better results if you fixed the root causes regardless of the camera, unless you've accounted for all that already and as mentioned a faint object would have provided a good baseline.
The advice you're heeding to is sound, so good luck and it'd be good to see your dslr pics.
Cheers

[Screwdriverone]

Hi Alistairsam,

I havent given up on the CCD route, for now anyway, it is just that as you say, there are a few things to fix first before I can justify the expense and eliminate the frustration I am getting with BOTH cameras.

Here is something I have been able to extract from my DSLR data using startools and slightly in Photoshop (layer blend the galaxy to the background.

It isnt going to win any prizes, but at least it looks a lot better than what I have been churning out from this 22 mins of data previously. I did use some darks and bias frames with this and it seemed to get a better result from the stacking, then I worked on the vignetting with Wipe and had a play with the masks in Startools to keep the galaxy isolated while I played with the gradients, seems to have worked better? There are still things I dont want in there, but this will do for now....

Lots of things to improve on, but hey, its a start...and after all, officially, my first galaxy shot......Yay! :P....:D

Cheers

Chris

[Poita]

I'll have a poke at it tonight.