Should I move to PixInsight?

[Midnight_lightning]

Hi, 

I'm now in my 3rd year of astro-photography and have done pretty  well initially using DSS, Star Tools, PSE for LRGB imaging.

I am currently using APP for stacking/registering, which I really like, and Affinity Photo to replace PSE, much better in my view but has more limited tutorials specific to astro. 

However, more recently I am doing more HST and HaLRGB, and am struggling with post processing. 

I am using APP for stacking registering and am very happy to stick with this, it is very good. I do a bit of post-processing in Star Tools (Stretch, LP, Decon, colour) on each channel and then move to Affinity to merge the channels and fully post-process.

I'm not sure this is the best approach, I find Star Tools great for LRGB but lacking for HST etc.

I could probably do it all in Affinity given time but it's a steep learning curve and I am wondering if PI is the way forwards?

I have been put off PI in the past by a, perceived at least, steep learning curve and difficulty to master, I'm not very good at remembering complex processes and settings especially when I only get to use them infrequently.

Any help, advice, thoughts will be gratefully received, especially with what PI could help me with. I particularly seem to struggle with developing and merging channels without introducing a lot of noise, and then getting the colours I want and am not sure whether PI would help with this or if I would still be needing Affinity photo for a lot of work? Ideally I would like to be using Affinity for minor tweaking. 

Thanks

Jon

  

  

[PeterCPC]

All I can say is that I have moved away from PI to APP and much prefer it. To me PI is far too complex to get to grips with properly. I use Affinity/PSP for tweeking.

Peter

[tomato]

Jon,

I am in exactly the same boat, I have been imaging for about 3 years and use APP (which I think is great), Star tools, PS and GIMP. I too have been considering moving into PI but have refrained for much the same reasons as yourself. Certainly a lot of the best NB images posted on this forum are processed with PI, but equally I believe there are a few very accomplished imagers who don't use it.

I'm retiring soon so I will probably buy it as a tool to try and keep my brain active...

 

Steve

[fwm891]

I use PI almost exclusively now - yes there's a learning curve as with any software. Watch some of Harry's video demos to give you a flavour of things. http://harrysastroshed.com/pixinsight/pixinsight video html/Pixinsighthome.html

You can even put all your flats, darks, bias and lights (LRGB's) into a single module and it will align and integrate them or you can do them individually...

Regularly upgraded too without extra fees. I'm a convert, it just takes a little time. Get the basics in place and you'll soon have a work flow without harming original data - think PS adjustment layer on steroids.

[wornish]

PI does have a fairly steep learning curve mainly because it has so many processes and scripts and can do lots of very advanced stuff.

If you only plan to use it occasionally then creating your own set of "how to" notes certainly helps.   I have used it for over two years now and I am still learning.  

 

 

[spillage]

I am a little bit behind you and only a couple of years into AP. From the beginning I have only used DSS ,startools and occasionally gimp. I did try PI LE and the PI trial and initially found it very overwhelming to understand compared to startools.

Startools to me is a very more user friendly bit of software that has some hidden talents and can produce some great images, its main fault not that its a fault but more of a good point is that its not as in depth as PI. In startools shrinking stars is easy seems to work okay..sometimes have issue with ringing. PI though seems to do a more consistent job but to me it feels like you have to run several processes to achieve it.

Waffling a bit here but I see it like installing a program on your pc. Startools seems like your windows .exe install and PI is like compiling the software in linux.

I will keep on with PI but I will say there is allot to remember.

[MARS1960]

I have used PI for about a year, it's easy to get an image, what isn't so easy is getting a great image and i still haven't.

Learning curve is vertical with overhangs but fortunately places like http://www.lightvortexastronomy.com/tutorials.html and books like Warren Keller' Inside Pixinsight exist.

I'm happy i bought it.

[Adreneline]

I use PI and PS. I find PI is best for pre-processing, colour calibration, and MLT noise reduction with a suitable mask before either merging channels or taking the image non-linear using many, many small iterations in HistogramTransformation. Once non-linear I like to use PI again to further reduce the noise levels using MLT and ACDNR processes. Personally I like to then use PS for colour manipulation using levels, curves, layers and masks; I have had some success using the colour manipulation processes in PI.

I'm not sure the learning curve for PI is any more or less steep than PS if either is to be used to its full potential; they are very different in their approach but as previously stated the likes of the LVA tutorials for PI and 'Dark Art or Magic Bullet' by @steppenwolf are both invaluable resources. I've found Annie's Astro Actions to be a really handy addition to PS. I've also begun to find that PixelMath in PI can be really powerful, especially if you scour the PI forum where you are likely to find that some very clever person has already written and shared the code.

I've learnt that less is often more and simply because all these processes and tools exist they don't have to be used on every image; as with all of them it is knowing when to stop - and I'm still trying to figure that out for myself!

Good luck with your deliberations.

[steppenwolf]

This is a tricky question to answer as so much depends on previous experience!

I have used PhotoShop for years and still use my trusty old PS3 version. I have always calibrated, aligned and stacked my images using MaxIm DL and then moved the master images into PhotoShop for the rest of the image processing. However, I have recently started using PI as I was curious to see what all the fuss is about and I find it a real pain to use in comparison with PhotoShop - but for calibration, alignment, stacking and more recently deconvolution, it is really good and worth the money for just this preliminary processing but after this stage, I move straight back into PhotoShop!

Will I ever go exclusively PI? Maybe, although it will be a very gradual process as currently I feel that I am making good use of the best attributes of these two software packages but will continue to explore PI at my leisure.

[sloz1664]

Pixinsight, like any other powerful software packages are difficult to master. But you do not need to master it you use it effectively. As previously stated, there are tutorials on the web that will get you started. Harry Page's beginner tutorials will give you a good understanding of the basics and If you have a particular target you would like to see processed in Pixinsight; there are many available on YouTube. Before you purchase, have a look at the various tutorials and see if you understand the processes and more importantly the process flow. If so, you know what you have to do. Personally I think Pixinsight is the way to go. It is designed primarily for astrophotography and, in my mind, it can only get better.

Steve

[wimvb]

10 hours ago, Midnight_lightning said:

I'm not very good at remembering complex processes and settings especially when I only get to use them infrequently.

PixInsight has 4 workspaces where you can keep individual processes iconized. That is, you process an image, and save an instance of each individual process on the workspace. At the end you save all the icons as a set. Next image to be processed, you load that set again. It will have all your processes at the settings you had before.

[Midnight_lightning]

Firstly, thank you all for responding, the information provided is very useful.

I watched all of Harry's videos yesterday and found it looks more straight forward to use than I had feared, although definitely a learning curve to go through. Importantly, I can see how I can immediately follow the tutorials whilst incrementally learning the package and will take out the 45 day trial initially.

I particularly like the Stretch and combine facilities which are areas where I struggle and appear relatively straight forwards. I can also follow Harry's workflow initially which is very useful information. I haven't looked at Steppenwolf or light vortex yet but will do over the weekend.

From what I saw in the tutorials it appears that the APP Calibration is more sophisticated/comprehensive than PI but I could be wrong.

I couldn't see any multi-session facilities in PI, is it possible to load all the Darks, flats, lights, bias, dark flats from several nights and have PI work out which calibration files to use for which lights from different sessions?

It's not a show stopper for me as I expect I will still use APP for calibration if I start using PI for Post-processing.

For anyone reading this who is starting out in AP or only using single shot/LRGB I highly recommend Star Tools.

It is extremely easy to use, you can leave most things at default settings and get good results most of the time but it is also very powerful and deep if you want to go into the complexities. My only gripe with ST is that support for HST / HaLRGB processing isn't great - although can be done. Beginners will really benefit from ST as you will have a very steep learning curve just learning to collect data. ST will allow you to get good results with minimal effort, just at a time when you will really need it.

Thanks again.

Jon 

 

[steppenwolf]

18 minutes ago, Midnight_lightning said:

I couldn't see any multi-session facilities in PI, is it possible to load all the Darks, flats, lights, bias, dark flats from several nights and have PI work out which calibration files to use for which lights from different sessions?

Yes, PI will do this for you with the caveat that your Flats will all need to have been captured at exactly the same focus position and camera rotation. There is a script called BatchPreprocessing that will do this for you. However, I use library Bias, Dark and Flat frames because I work from a fixed observatory and if you don't have that luxury, I would recommend using library Bias and Darks but take all your Flats in one session.

[Scooot]

I started with PixInsight knowing nothing else, well almost nothing, so anything was going to be a steep learning curve. I like using it now, but I’m still learning. 

I’ve used many of the guides & tutorials already mentioned but the light vortex tutorials mostly, and found them to be extremely helpful. However I’ve recently discovered an unadviseble procedure in the light vortex calibration process, which is to pre-calibrate darks with the master bias. This link explains why it’s a “No No” and a better process, which you might find helpful  if you do decide use PixInsight to calibrate & integrate. 

https://pixinsight.com/forum/index.php?topic=11968.0

 

[wimvb]

33 minutes ago, Midnight_lightning said:

From what I saw in the tutorials it appears that the APP Calibration is more sophisticated/comprehensive than PI but I could be wrong.

If you use the batch preprocesser (bpp) script, much of the sophistcation if PixInsight is hidden. But if you use the individual processes, you have all of its power available to you.

The intermediate way is to calibrate in bpp, and do the star alignment and integration seperately. This gives you more freedom to adjust parameters in order to optimise snr of the final image, as well as the ability to use drizzle and large scale pixel rejection (efficient way to remove satellite trails).

[Midnight_lightning]

10 minutes ago, steppenwolf said:

Yes, PI will do this for you with the caveat that your Flats will all need to have been captured at exactly the same focus position and camera rotation. There is a script called BatchPreprocessing that will do this for you. However, I use library Bias, Dark and Flat frames because I work from a fixed observatory and if you don't have that luxury, I would recommend using library Bias and Darks but take all your Flats in one session.

I use library Darks and Bias but take Flats/Dark Flats for each session. So for my 3 sessions on M31 I have one set of Darks/Bias but three dedicated sets of Flats/DF's for each filter. With APP I just load them all up and it uses the correct calibration frames for each session. 

I usually take Flats next day using SGP and have never considered the focus position, I just use whatever the last focus position was from the night before for all filters. Is this an issue, given focus changes with temperature I'm not sure how I would decide what specific focus to use for each set flats ?

I set up each night in the same place but there will be minor positioning differences including potential for very minor camera rotation. 

 

[wimvb]

18 minutes ago, Midnight_lightning said:

I usually take Flats next day using SGP and have never considered the focus position, I just use whatever the last focus position was from the night before for all filters. Is this an issue, given focus changes with temperature I'm not sure how I would decide what specific focus to use for each set flats ?

The small refocus adjustnents shouldn't matter. Any remaining vignetting will be corrected in dbe during postprocessing.

18 minutes ago, Midnight_lightning said:

I set up each night in the same place but there will be minor positioning differences including potential for very minor camera rotation.

Best to take fresh flats each session when camera orientation may have changed. Dust bunnies may also have changed position. Positioning changes (goto inaccuracy?) don't matter.

[RayD]

To add to Wim's comments, regarding camera rotation, this will be taken care of during the registering process but may result in the need to crop your image depending on how much rotation you are talking about.

[sloz1664]

If you have APP for image calibration & stacking then I'd stick to it. It is the core process of APP and does a great job in processing your data.

Steve

[Grierson]

As a Pixinsight novice can I just put a word in for Asto Dude’s 12 videos for absolute beginners. I havn’t the experience to say whether they are better or worse than others mentioned above but I am finding them very useful.

[harry page]

On ‎24‎/‎11‎/‎2018 at 09:07, Midnight_lightning said:

I watched all of Harry's videos yesterday

You must be mad

Harry

[Midnight_lightning]

Harry, they are great, and having downloaded the trial this morning I watched all the Novice ones again, this time  running my M31 data through them.

I did have an issue which I cant explain, don't know if you could help? Image below.

EDIT- JUST TRYING AGAIN AND LOOKING BETTER, MAYBE USER ERROR - WILL POST TOMORROW IF I SOLVE IT 

My M31 RGB showed a red bias when combined, I tried the Linear Fit but it didn't make any noticeable difference. So I processed through to Colour Calibration hoping something would fix it but it didn't. I have processed this data in Star Tools without issue so am not sure where I went wrong?

The videos are great though, meant I could quickly start getting to grips with PI.

Jon

 

 

[steppenwolf]

2 minutes ago, Midnight_lightning said:

Harry, they are great,

I concur! 

[sloz1664]

2 hours ago, harry page said:

You must be mad

Harry

I thought that was a pre-requisite of astro imaging anyway

Steve

[wimvb]

40 minutes ago, Midnight_lightning said:

Harry, they are great, and having downloaded the trial this morning I watched all the Novice ones again, this time  running my M31 data through them.

I did have an issue which I cant explain, don't know if you could help? Image below.

My M31 RGB showed a red bias when combined, I tried the Linear Fit but it didn't make any noticeable difference. So I processed through to Colour Calibration hoping something would fix it but it didn't. I have processed this data in Star Tools without issue so am not sure where I went wrong?

The videos are great though, meant I could quickly start getting to grips with PI.

Jon

 

 

This looks very much like a colour gradient, even though it doesn't show in the background. Have tried dbe?