First Processing with PixInsight - NGC2070 Ha

[DavidTrap]

Finally nailed my backside to a chair for long enough to go through some PixInsight tutorials and videos (having a day off helped) so I can now present my first offering for your appraisal. It's definitely complicated, but seems very powerful.

The image was taken through my Tak at F6. I took 60 & 900 second exposures through a 5nm Ha filter. I calibrated and stacked in Maxim as usual. These were registered and HDR combined in PixInsight. I then applied a histogram transformation and HDRWavelets. (I can see H rolling his eyes at all of this HDR stuff already!) Next was ACDNR and a further histogram stretch, followed by an Unsharp Masking.

The full size image is found here (http://users.tpg.com.au/dctrap/astrophotos/NGC2070_HDR_900_and_60_sec_Ha%20ful l.jpg).

It certainly isn't my best data - it was captured in poor seeing with lots of moisture and a bit of wind about. I wasn't 100% happy with guiding or focus either. (More excuses than a tomcat going to the vet!)

Hopefully I'll get a LRGB image up in the near future and then a HaLRGB combo.

Thanks for looking,

DT

[Octane]

That looks fantastic! Very 3D feel to it!

There's some weird stuff happening in the background. Is that due to heavy use of some kind of minimum filter? Like, splotches, almost?

Also, I've got nothing against HDRWavelets in PixInsight; it's the clown vomit filter applied to terrestrial imaging which makes head explode. :P

H

[RickS]

David,

Have a look at ATWT for noise reduction (even on linear images) and also for sharpening. The wavelet based tools are amazing.

I'm processing a HaRGB Tarantula right now. Perhaps I should have chosen a less popular target. There can't be many photons left out there :lol:

Cheers,
Rick.

[DavidTrap]

Thanks Gents,

H - I think I may have been a bit heavy handed with the noise reduction...

Rick - there are many things to try in PixInsight, I just need some hints on how to use these things - the lack of manual is a little frustrating!

Anyway, here is a rough and ready LRGB - very ROUGH!!!

DT

[RobF]

Love the Ha David - coming together very nicely.
Heck, you should have seen my first few PI disasters. Your way out in front with those. Certainly a different interface, but some amazing functionality in there.

[RickS]

David,

Here's some info on using ATWT for noise reduction:
http://pixinsight.com/forum/index.php?topic=3184.msg21961#msg21 961

Here are some pointers to more info:
http://pixinsight.com/forum/index.php?topic=3268.msg22460#msg22 460
RobF's list of links is very handy.

There's lots of good stuff on the forum, but it's harder to find than in a manual...

The rough and ready LRGB looks pretty good!

Cheers,
Rick.

[Hagar]

This image looks quite good David. I have a copy of Pixinsight but struggle with it. For me it seems very complex and doesn't seem to give me anything I can't get easier in CCDStack and Photoshop.

[DavidTrap]

Thanks Rob, Rick & Hagar,

Hagar - I'm sure PS has similar functionality, but I'm getting an impression that Pixinsight has a lot of the tools built in that you need to add/buy plugins for with PS.

DT

[Ross G]

Hi David,

Your photo looks very good.

I am just learning Photoshop, but I am starting to read some very good things about PixInsight.

If you had the time, it would be good to see a comparison of the same shot done with both processing systems.

Good luck.

Ross.

[DavidTrap]

Thanks Ross,

I really think it's a case of "more than one way to skin a cat". Outcomes will be similar, but the journey will be different. I'm sure many images will pass through both programs.


DT

[jjjnettie]

That looks fabulous David. :D

[gregbradley]

Well done on taking the plunge and climbing the learning curve. PixInsight has a lot to offer.

Nice image.

Greg.

[h0ughy]

nice work David. Pixinsight is not photoshop and i think you will soon realise that. it is quite a powerful programme specifically for astro

[DavidTrap]

Thanks JJJ, Greg & Houghy. It's certainly a learning curve. I agree that it is an astro-specific program which should deal with our data quite well, but I also can see that some of the "artistry" of image processing will be achieved with PS.

DT

[dugnsuz]

The Ha version is beautiful David
Doug

[DavidTrap]

Ta Doug.
Hope to get around to a reprocess of the RGB soon, and hopefully add some more data this new moon weekend.

DT

[alan meehan]

Good shot David and nicely processed ,keep with Pixinsight its a great
programe.
AL

[DavidTrap]

Thanks - Allan. Just need to find a few more days off!

DT

[CoolhandJo]

Nice One

[hotspur]

sorry for not looking at this earlier,but have been a bit busy.

Well done David!,it has not taken you long to become one of the best
imagers on these pages! Your colour is the best I have seen-you will be
featured in world wide publications very soon! going by these great examples.

cheers Chris

[DavidTrap]

Thanks Paul & Chris - although I think I've got a little way to go before I'll see my images in print. You can hide an awful lot in a 200kb image!

DT

[Stevec35]

Not bad David. Good luck in your future endeavors with Pixinsight. My usage of it has so far been pretty limited. I just wish the @#$** thing had a proper manual.

Cheers

Steve

[DavidTrap]

Thanks Steve,

The lack of a manual is challenging, but I wonder what has been more use to most PS users - the official Adobe manual or tutorials?

DT

[Stevec35]

Good point. I guess my real objection to it is that it looks nothing like Photoshop. I do like some aspects of Pixinsight though. The deconvolution routine is very effective.

Cheers

Steve

[Poita]

For once I really like the Ha image much better, usually I'm not taken by B&W Ha stuff, but in this case it seems to have way more detail and tenuous whisps than the colour image.

[Hagar]

No David Pixinsight isn't photoshop. It is supposedly made for astro imaging. It is a very complex program that is only supported by very basic tutorials which don't cover a poofteenth of the controls available in the package. The tutorials only cover the basic use of the some of the controls and processes.
At least the photoshop help system gives a wide breakdown of it's controls and processes.

Like it or not, Pixinsight has been sold as a super package and it may well be but until it has a decent user manual which spells out all control commands etc. it will still lag misserably behind Photoshop.

It is great to have a program that is well supported by the manufacturer, has reasonable literature with it and as a plus has people writing very detailed tutorials on it's use. For me Pixinsight was a total waste of my money until it can provide detailed support like photoshop.
PS. A forum doesn't constitute detailed support either.

Sorry I'm not as enthusiastic about it as you are but so far what I have seen just doesn't do it for me. Photoshop is still industry standard and I'm sure Pixinsight uses many of the base control used in photoshop.:shrug:

[Octane]

Doug,

I prefer to still think of PixInsight as pre-processing software. There will always be a place for Photoshop for me. Neither is a replacement of the other, and, each fulfils its own niché.

It is actually really quite straight forward to use the basic commands. And, if you hover the mouse over the functions, it tells you all about them.

I used it for two days and was able to produce my Rho Ophiuchus luminance mosaic without resorting to any documentation or tutorials. Just pressed buttons and it all started to make sense!

When the help documentation is released, it is going to be incredibly comprehensive.

H

[Hagar]

Firstly Sorry David for apparently hijacking the thread.

Now H, Yes i understand what you are saying but a program without a manual is next to useless for an old fart like me. Be it a hole in my memory or some other failling facet. The crux is a program costing 170 Euros should come with documentation. In fact it is apparent that the programmers are in fact far better supported than fools like me who paid for the program and are still waiting for a real documentation source. Programmers have a full set of VCL documentation.

I must find out if I can transfer the licence and sell the dam thing to some other fool.

[Octane]

It's cool, Doug. I do understand where you're coming from. :)

Sorry to continue the hijack!

H

[DavidTrap]

Hijack away gents - tangential thoughts often provide very useful information for our obscure hobby!

DT

[rcheshire]

The easiest way to learn PI is to follow the video tutorials and ask questions of the developers and module authors.

I have never used PS and started with Pixinsight from scratch when there was exponentially less documentation than today. Now each tool is linked to its docs. The manual is in-built you might say.

PI runs on Windoze, MacOS, Linux (I use the 64 bit version) and OpenBSD. It is updated periodically (automatically) and there is no charge at this stage for version upgrades. Version 2 however, may be a different matter - fair's fair. The energy and time that Juan and Vicent put in is exemplary and worth the money.

The videos are the best starting point. Word of caution - you might want to import images from your favourite stacking program (DSS) rather than do the calibration in PI. I'm quite comfortable with master frame creation and calibration in PI these days. It's not that difficult if a little time consuming. There are pitfalls if importing calibration frames that have been processed in other programs - not always.

There is virtually a recipe tool set in PI to get started. Covered in Harry's videos.

[Hagar]

I see what you are trying to say Rowland but to be honest there is No detailed or other instructions to speak of.
An example, Noise reduction. Opening this gives you 3 options.
Option 1 ACDNR
Option 2 GREYCstoration
Option 3 SCNR
I can hold the mouse on any of these options until it is burnt into the screen and it doesn't tell me jack about what it is. Even opening one of these options and I am greeted with a dialogue box with the same heading and still no information. This may be well and good for someone who wrote the code behind it and knows what it is about but in the real world this is truely pathetic when a mouse over can be used to give some detail or explanation.

I also like the reasoning on the web site for not having a manual. stating it would be of huge propotions. So what, at least it would be something to use and work from. To say the developers have expended a lot of time and effort to develope the program, I have no doubt they have but the product is only half done without some real reference material. We in the astro community seem to be ready to accept half built software and hardware without being able to make adverse comment. In an electronic world a document of 20 GB is acceptable for download. Why shouldn't any software developer be held accountable for reasonable instruction manuals rather than just relying on other people to create video tutorials which to say the least are very basic and cover about one hundreth of what some of the controls and dialogue boxes can do. If this is all you need to do then why program all these controls and provide nothing to say what they do.

As I said in the previous posts. Just another waist of money and not real cheap at that.

[RickS]

And then there's option 4, ATWT ;)

There is an "Officially Unofficial" reference guide at: http://blog.deepskycolors.com/PixInsight/

I found it easy to learn the basics from the tutorials and then pick up new things from the forum and experimenting. I also found that reading the Berry & Burnell book (Handbook of Astronomical Image Processing) gave me enough background to understand the basics of many of the algorithms and parameters.

Occasionally I would find a PI manual useful for reference but it doesn't bother me too much. I can understand why others, like Doug, might find this a major obstacle though. I think it's probably a matter of personal learning preferences...

Cheers,
Rick.

[Hagar]

Thanks Rick. This link looks to be the goods but I am amazed thyat a resource like this isn't listed in the pixinsight home page. Perhaps it highlights a hole left by the project managers.
When time permits I will have a good look at this. Thanks Mate.

Now what the hell does ATWT mean. I am certainly technologically challenged with abreviations like this. I also had trouble with FFS so feel for Unipol. That gave me another one to forget.

Thanks Rick, most helpfull.

[RickS]

Doug,

ATWT is the AtrousWaveletTransform (you can see why the name gets abbreviated!) It's wavelet based so it allows you operate at different physical scales. You can use it to do noise reduction, even on linear images. It also does sharpening.

You are right. It's surprising that the PixInsight web site doesn't reference the reference guide even if it's not official.

BTW, I thought FFS might have been a Fast Fourier Transform minus one? :lol:

Cheers,
Rick.

[RobF]

I think many of the criticisms leveled at PI are quite valid, particularly the (very gradually improving) documentation black hole.

But, there are many things it does that I can't do or can't do better any other way, and that's why I bought it. That and a belief that the list of such features has and will continue to grow. Just a tool though in the end, like so many others. I am left handed after all.....;)
It's certainly not the fastest processing production line I've ever had, but I often find myself going back to scratch with PI if I've been tempted into a processing shortcut someother way.

[rcheshire]

True, documentation is a downer and there have been many complaints about this. Still it is improving. I would go so far as to say that because PI has been developed and is used by a scientific community, that there has been an assumed level of knowledge. This was evident in the DSLR_RAW processing saga.

As far as I can tell, increasing popularity among amateurs has taken PI by surprise and the pressing need for documentation now exceeds development efforts. I think they have employed a new coder/programmer.

Be patient. If I can get my head around it without the docs and a non technical background - there are some processes that require external study - it should be a no-brainer for those with the relevant expertise.

ATWT is the Swiss army knife - used in conjunction with the Screen Transfer Function for realtime feedback the simultaneous sharpen noise reduction process is awesome - linear and non-linear - add HDRWaveletTranform, LRGBCombination and LocalHistogramEqualisation, some curves and your practically done.

If you want to avoid the extensive calibration process use DSS. I suspect that they will automate calibration at some stage, although the philosophy is very open source - the user should have control over parameters and processes at every step.

The bottom line is 'experiment experiment experiment' - each image/data set is different. I'm a convert;)