Coming Soon ...

[MarkS]

Coming soon ...

... is my Spring experiment - the Bananascope Deep Field image.  I already have 30 hours in the bag (360 x 5min) and tonight is another clear night!  Plus I have 8 hours that I took last year that I could potentially add in.

[Whitters]

Eager anticipation

[Rocket Pooch]

Blimey

[JohnP]

Well I'm voting APOD & I haven't even seen it yet? look forward to it.... John

[The Thing]

I bet it will be a real stunner. It probably already is...

[MarkS]

To be honest I haven't processed any of it yet so I've no idea how it will come out.

I grabbed another 6 hours last night (so I have 36 hours).  Hoping for another 6 tonight, making 42. 

Mark

[Carole]

The mind boggles, looking forward to seeing the results Mark.

Carole

[RobertM]

I'm really looking forward to seeing this marathon effort too, how long have we to wait ?

[MarkS]

I'm really looking forward to seeing this marathon effort too, how long have we to wait ?

Dunno - I've a bit overwhelmed with data right now - already over 540 five minute subs and tonight is another imaging night    
Some people wouldn't even have 500 planetary frames in an AVI for Registax.

This is 6th successive night of imaging since Friday at DSC and then I also did a few nights last week.  I've never had such a long spell of imaging. 

My PixInsight script keeps crashing with "Out of Memory" so I'm going to have to process in chunks of maybe 100.

Mark

[MarkS]

Only got 4 hours of usable data last night because of the moon.

If tonight is clear I'll be able to breach the 50 hour threshold and after that the moon brightness will prevent any further work.

Was hoping to batch process some data during today but my PixInsight trial version refuses to start - it can't access PixInsight's license server on the net

Mark

[Rocket Pooch]

This had better be good, good dynamic range, perfect colours, nothing washed out, no noise and round stars

[Mac]

This had better be good, good dynamic range, perfect colours, nothing washed out, no noise and round stars

At least set Mark a challenge,
this is a normals days processing for him. 

The final image will be interesting,

You never know but NASA might come knocking on your door with an image processing job.

If as you said you are going to do them in batches of 100,
it would be interesting to see how each batch off 100 improves the final image.

Or whether up to a point, adding more data gives you that much more improvement,
obviously where there is faint data the more the better, but at what point do you draw the line?


Mac.

[RobertM]

Oh dear, I am dissapointed, only 50 hours !  You mean we'll have to put up with noisy data

[JohnP]

Mark - If you are using PI will be useful to try the animate galaxy feature - looks like a really useful feature - came across it on this post on PAIG..

http://www.progressiveastroimaging.com/viewtopic.php?f=13&t=6155

[MarkS]

Mark - If you are using PI will be useful to try the animate galaxy feature

I think you mean annotate!

yes I'll definitely give that a try.

[JohnP]

Well spotted... Of course I put that in as a Friday morning test to see who was awake...! 

Now an animate galaxy feature would be funky - show them all spiralling off in different directions... might take a bit more processing power than my humble laptop though...

Just get your image processed this weekend!!!! Never known such a thread in anticipation of an image!

[mickw]

OK getting impatient now, hurry up 

Is the banana deep field just a relatively blank area or are you imaging a known area for a hell of a lot longer - Or must we wait and see 

[Mac]

Or must we wait and see

Tut Tut,

Good things come to those who wait.

Mac.

ps im still waiting. 

[MarkS]

It's a area I've imaged once before.  Plus it's an area where I can leave the scope running all night without it hitting the tripod legs or similar.  
I've haven't stayed up all night imaging 7 nights in a row  
I set the start the imaging run off at 10pm and then go off to bed, switching it off at 6am.

I may be able to get a preview done this weekend.

Mark

[Rocket Pooch]

Markabs Chain thingy

[MarkS]

Markabs Chain thingy

Yes - that's the one.

[MarkS]

Can't believe this.  After many abortive attempts (because PixInsight kept crashing with "out of memory" problems in Windows XP - XP isn't officially supported but most things work) I finally got it all processed bit by bit in PixInsight only to find that the final stack was complete rubbish.  Something had gone wrong in an early calibration stage resulting in a image with a nasty red cast and with truncated pixel values.

So I'm going back to my trusty old IRIS and beginning all over again ....

To prevent loss of dynamic range I'll calibrate and stack the 635 subs in groups of 8 in IRIS and then perform the final integration of 80 substacks in PixInsight.

AARRGGHH!!!!!

[MarkS]

I finally got a sensible stack of this data overnight and had a quick look this morning. 

First impression is that the image is very much smoother than anything I have seen before, which means that I can stretch it much further to bring out lots of galaxies in the background.  But stretching the image by 3 times more than usual also means that the gradients become 3 times worse.

The gradients in this image really are a major issue.

I'll post a preview tonight but I'll probably  have to do a few more complete processing runs in order to get this anywhere near to a state I'm happy with.

Mark

[JohnP]

First impression is that the image is very much smoother than anything I have seen before

blimmin' hope so after 50 hrs....

[MarkS]

Well here's the preview:


Larger version here:
http://www.markshelley.co.uk/Astronomy/2012/markarianstretched.jpg

I've stretched this to a ridiculous degree and you can see the gradients I refer to.  Lots of hot pixels as well.   This is going to take an awful lot of work to turn into a reasonable image

The photon noise and the gradients from the sky background level are by far the main limiting factor with this image.

[Rocket Pooch]

ouch

[Mac]

I gave up counting at 40 galaxies 

Hmm that does look like a challenge, but you always did enjoy such things

I'm sure you have this under control.
Looks very good though,

Mac.

[JohnP]

This is going to take an awful lot of work to turn into a reasonable image

Blimey you are not kidding.. Why do you have such issues with gradients etc? Is it due to local light pollution? Also I know you said you have pushed it a lot but am surprised at the amount of noise given your dark location & number of exposures... Is this the full stack or just a subset?

I really hope you can sort all the issues out cause this will be a stonking image - Its a great FOV.


John

[Carole]

Wow, I got up to 63 galaxies.

I see what you mean about the gradients Mark, it will be a lot of work to sort it all out, but the end result will look fantastic.

Carole

[MarkS]

There are a few reasons for the gradients:

1) Gradients are always a problem for the Canon on the Tak.  The F/2.8 light cone is very wide and causes "shadow zones" with strong gradients on the CCD where the light cone hits the CCD housing.  The CCD is at the bottom of a "tunnel" - not an ideal position!  Robert had it even worse with his Canon on the F/2 Hyperstar.  The light cone hitting the walls of the housing also accounts for those extra diffraction spikes. 

2) Markarian's Chain is not particularly high in the sky so it is in a region where light pollution sky gradients exist in any case.

3) It is much worse in this image because the low noise levels have allowed me apply much greater stretch to the dynamic range. The more you stretch the data the more you reveal the gradients.

4) PixInsight does a weighted stack which measures the noise of each stack and optimally weights it.  This has allowed me to throw in a whole load of dubious subs that I wouldn't normally use.

5) I used a dew shield for this image but the flat I used was a earlier flat taken without a dew shield.

I need to do a partial stack containing only the better data, deal with its gradients and then use it as a template for gradient removal on the rest of the data.

Although I live in a relatively dark place (compared with London) light pollution is very definitely the main limiting factor for broad band imaging.  In this particular image the photon shot noise of the light pollution is a factor of 5 greater than the read noise and the thermal noise.

This means that if I did the data acquistion with the same equipment from somewhere truly dark, I would be able to reveal fuzzies 5 times fainter - plus I would not have gradients to deal with!

How depressing ...

[Rocket Pooch]

Have a beer...

[mickw]

Beer

The answer to, and cause of all of life's problems 

[RobertM]

It may take a little longer than you imagined but it will work out ok.

First off I would take flats with the dew shield and in a suitably dark place.  That should deal with optical defects but then you will have the gradient to deal with.  I would hope that's fairly linear so should be an easy second step (says he hopefully).

Robert

[Whitters]

I count 77 Fuzzies.  keep at it mark

[mickw]

I think it's looking very good, you know what the issues are and how to deal with them so it's not really a problem - just time consuming 

The end result will be worth it though