Pixinsight

[Gina]

Yes, I thought that too.  My images so far with the ASI1600MM-Cool don't seem to show any gradient other than a slight amount of vignetting which I get rid of with Flats.  Mind you, the DSOs I've imaged so far have been very high in the sky and light pollution is very low here.  It may be a different matter with DSOs lower down or with the moon out.

[Jokehoba]

16 hours ago, wimvb said:

With DBE there seem to be two philosophies: few and large samples or more and smaller samples.

Juan Conejero (Mr PixInsight) advocates few samples per row so as not to pick up nebulosity as background.

On the other hand, in Alejandro Tombolinis examples, he can use a large number per row.

Whatever you choose, always keep and examine the background model. If that has some kind of structure or colour variation, you have most likely picked up nebulosity as background.

It's a tedious job examining, moving, resizing or deleting 400+ samples in an image. But in my experience, it pays off. Processing gets so much easier if you get rid of gradients.

http://wimvberlo.blogspot.se/2016/08/note-on-dynamic-background-extraction.html

Totally agree.  I used to cover the image in samples but nowadays I'm more selective. As an example, IC1396 - as originally processed in 2014, and same data re-processed in 2016, having gained more experience of PI and the power of DBE especially.

[Gina]

Yes, In the Orion example above I too thought there were far too many samples in the area of the sword where there is a lot of nebulosity, apart from the M42, Horse's Head and Running Man.

[harry page]

Hi

If you are removing a gradient ( which by it nature is smooth ) well chosen samples are the way to go , you  can very often

remove a gradient with very few sample ( It can be done with 3)  , I often only use say 10 or 20 --As you might have heard me say

Quality not quantity

If you are trying to correct other problems with DBE ( like uneven background ) It can be useful to use more samples

Other than some narrow band images , I have never seen a image that does not benefit from s good DBE ,  It is only when you have achieved a flat evenly illumined image

can you get the best from your Image .

Regards

Harry

[wimvb]

Just did a quick 'n dirty DBE for Atreta (Frank).

The gradient was really nasty. It consisted of severe vignetting, mostly at the bottom and right, plus a band at the bottom

Vignetting was cleaned up very nicely by using "division" as the correction method

I couldn't for the life of me get rid of the band at the bottom with DBE, no matter where I placed the samples. Ended up using the CanonBandingReduction script, which is kind of a cure-all for this sort of problems.

Then applied a second dosage of DBE to flatten out the background some more. CBR can cause some mild background variation, which can be cleaned up by DBE.

 

BTW, now that I look at what I'm writing, I realise that some of the steep learning curve in PI must be to remember and figure out what all these acronyms mean.

[Oddsocks]

On 6 September 2016 at 16:39, gnomus said:

Mr Keller has a book coming out later this month - https://www.amazon.co.uk/gp/product/3319256807/ref=oh_aui_detailpage_o04_s00?ie=UTF8&psc=1 - I have no idea if it will be any good, but I have ordered it based on the strength of his videos.

 

Looks like the publishing date has slipped again....Had an email from Amazon this morning saying expected delivery now 20th October.

[wimvb]

9 hours ago, Oddsocks said:

Looks like the publishing date has slipped again....Had an email from Amazon this morning saying expected delivery now 20th October.

Strange. I got an email from a swedish book site, saying that it's available for delivery. That was a few days ago.

Just cancelled my amazon order and placed an order on the Swedish site. It says delivery estimate 2 - 5 days. Let's see if they can manage that ...

[Marky1973]

I've also had a dispatched email from Wordery via Amazon... Seems odd... Might need to contact the seller and check....

[Filroden]

The book is available in ebook format via iTunes. It may we'll be available through other stores. It comes with a free sample of the first couple of chapters.

[wimvb]

According to the isb number it's the softcover that's available, and the one I ordered.

[Oddsocks]

3 hours ago, Marky1973 said:

I've also had a dispatched email from Wordery via Amazon... Seems odd... Might need to contact the seller and check....

 

2 hours ago, Filroden said:

The book is available in ebook format via iTunes. It may we'll be available through other stores. It comes with a free sample of the first couple of chapters.

Thanks Mark, Ken

My order for the book was via Amazon's own book shop so perhaps their supplier is the one that is late.

I do have a few eBooks on computing subjects but always find it frustrating having to flip between the eBook app and the live program when trying to follow an eBook for directions, just too set I'm my ways I guess, although thinking about it just now it occurred to me I could read the eBook on my iPad while working on the program on the Mac....Hmmm

[wimvb]

Springer has a release date of Oct 11 on their site for the softcover. The ebook seems available for immediate download.

[Gina]

I too have had an email from Amazon - "Estimated arrival date: October 11 2016".

Coming back to the software itself, how good is PixInsight at rejecting poor frames?  I captured 168 120s light frames in Ha overnight, probably half of which are either totally useless or of poor quality.  Should I go through all those frames looking at each individually - a long and tedious job - or could I leave it up to PI to discard duff and poor quality frames?  Using the BPP script to calibrate the frames.  The poor and duff frames were due to bands of cloud passing over during the imaging run.

[Filroden]

The SubFrameSelector process does a great job of measuring each sub against various measures and you can then compare them on a graph a deselect ones which fall outside your criteria. You can even apply weights back into the header files whic it will use in integration so the better subs contribute more. 

Its described in detail here: http://www.lightvortexastronomy.com/tutorial-pre-processing-calibrating-and-stacking-images-in-pixinsight.html#Section5

and it's covered in the book though he does it at an earlier stage in the process.

[Gina]

Thank you Ken - I'll carefully read through that   Meanwhile, I'm just trying BPP to see what is makes of my data.  I have 100 bias, 66 dark, 100 flat and 166 light frames.  It rejected 29 lights as not having a star match and is now integrating 137 lights having been through all the calibration stages.  I know BPP only does a rough job and that better processes are available for proper processing.

[Gina]

The result from BPP was rubbish!  Not really surprisingly as it only rejected 29 frames due to having no stars - it accepted all the rest - good or bad - with the same weighting!  I shall use the SubFrameSelector process you linked to above and do the job properly.  I'm sure this will be a lot better than for me to select frames that look good and dump poorer one which might have some value.  I can use bias and flat frames I already have in the camera sensor measurement script but will need to take a couple of darks as they need 10:1 exposure ratio (it says).  Have to say, I like this scientific way of choosing frames over pure guesswork

[wimvb]

As long as bpp star registration tool can detect stars, it will accept the image. There are settings for number of stars and iterations, which will affect acceptance of a sub. Far better is either subframe selector as already noted before. I usually load calibrated subs in the blink tool, and evaluate visually. Then I move the best calibrated subs to a new folder and register/integrate from this.

Either way, the process of selecting subs isn't automated. The DSS way of rejecting x% subs is arbitrary anyway.

[Gina]

Thanks Wim   There's no doubt that PixInsight excels over DSS etc. in so many ways.

[derrickf]

Hi Gina,

if you haven't already seen it, Jim Morse <jkmorse57@gmail.com> updates a PI cribsheet and makes it freely available to those that request it from him. Calling it a cribsheet doesn't really do it justice - the current version is about 70 pages in length and contains huge amounts of info about workflows, process settings etc. Jim included a spreadsheet that can be used to build sub-frame selector expressions with the latest version of the document.

 

HTH

 

Derrick

[jkmorse]

Hi folks,

I am brand new to the forum but somewhat of an old hand at PI and, as Derrick indicated above, I have a workbook (started out as a cribsheet but its now up to Rev 41a and 70+ pages so workbook is a better description).  In any event, A couple of people saw Derrick's note and have asked for a copy which I was and am happy to provide free of charge (just my way of giving back to the community).  Just drop me a line at jkmorse57@gmail.com if you are interested in the workbook or if you have questions about PI generally.  Though I have no affiliation other than as an extremely happy user, I have to say nothing else comes close for astro imaging (and believe me, I have tried most of the alternatives in my time).

And Derrick, thanks so much for the kind words

Best,

Jim

 

 

[Singlin]

e-mail sent.

[ollypenrice]

28 minutes ago, jkmorse said:

Hi folks,

I am brand new to the forum but somewhat of an old hand at PI and, as Derrick indicated above, I have a workbook (started out as a cribsheet but its now up to Rev 41a and 70+ pages so workbook is a better description).  In any event, A couple of people saw Derrick's note and have asked for a copy which I was and am happy to provide free of charge (just my way of giving back to the community).  Just drop me a line at jkmorse57@gmail.com if you are interested in the workbook or if you have questions about PI generally.  Though I have no affiliation other than as an extremely happy user, I have to say nothing else comes close for astro imaging (and believe me, I have tried most of the alternatives in my time).

And Derrick, thanks so much for the kind words

Best,

Jim

 

 

Whoa, it is perfectly obvious that other graphics programmes come close because some of the best astrophotos ever taken have been done without Pixinsight.

What I think each of us can reasonably say is that a particular graphics programme works for us because we understand it, because its tools and interface suit us, because we like the results we can acheive with it. You find this in PI. I find it in certain parts of PI and I love to use those parts, but I am far more at ease using layers in Photoshop. I have no intention of asserting that Ps layers are better  but merely that they are better for me.

These programmes are tools. Some prefer one, others another. It depends on who you are and how you think and work.

Olly

[jimboozle]

email sent - thanks for the offer!

[jkmorse]

Hey folks, 

Jim again.  I received a bunch of requests but am flying home this evening so it will be later tonight before I get a chance to respond.  But I will answer all tonight (and may just take a [removed word] at responding to Olly too ?).

best,

Jim

[ollypenrice]

12 hours ago, jkmorse said:

Hey folks, 

Jim again.  I received a bunch of requests but am flying home this evening so it will be later tonight before I get a chance to respond.  But I will answer all tonight (and may just take a [removed word] at responding to Olly too ?).

best,

Jim

 I'm bursting with curiosity to know what the removed word might have been! I hope it was juicy!!!

It's good of you to publish your workflows. I'm going to buy Warren Keller's new PI book as well. I'm not a software fundamentalist and I'm certainly not monogamous! (As an imager...)

Olly