The PixInsight or Photoshop Black Hole

So what’s better for processing images, PixInsight or Photoshop? No, this isn’t exactly another one of those opinion pieces. Google “PixInsight vs. Photoshop” and you can read for hours. The question may have been answered a million times over with a thousand slightly differing opinions, but none of them might be right for you. The super short answer is that Photoshop and PixInsight can both be used on their own, but in the best use case they compliment each other. Reading through, you’ll see that this doesn’t necessarily mean you have to add two more software licenses to the cost of the shiny new imaging rig you just bought to get started. I’ve only just now transitioned to a PixInsight and Photoshop blend after spending more than three years strictly processing all data with Deep Sky Stacker and Photoshop.

For a little background and history, I’ve been in various levels of IT organizations for very small and very large companies over the last 25 years. I’ve spent most of that time in the end user space, with a big chunk of the last quarter century spent improving user experience. So I think I’m somewhat qualified to say this: PixInsight is the worst thought out end user application ever created that maintains such a large popularity. There is absolutely nothing intuitive about it. Books – very long books – are written just on using the basic functions of the application. Thousands of hours of YouTube tutorials exist. None of the tutorials are the same, and they are quickly out of date. “But that just shows how versatile the application is”, you say. Hogwash. Versatility doesn’t have to introduce a level of complexity requiring a one-hour video tutorial or an entire chapter in a how-to book when learning a single basic function. You shouldn’t need a programmers brain, or require days of research to learn basic functionality of any application; especially not an application at this price point.

So let me tell you what I really think of PixInsight if you haven’t already guessed. It’s freaking awesome!

No, you didn’t misread. No, I’m not contradicting myself. PixInsight is an insanely powerful application, and the data you spent hours acquiring can benefit from it greatly. The usability of the application just absolutely sucks. There’s really not a softer way to put it. The complexities and number of wrong turns you can make from the very first steps causing vague errors that aren’t documented can drive you insane. If you try to speed through learning it, or try to take it on before you’re ready, you’ll be aggravated beyond belief, your wallet will be $250-300 (USD) lighter and your data will look like my back yard yard after the snow melts in spring… I have three dogs.

Well, what about Photoshop? I’ve been using Photoshop for terrestrial photography and digital art for around 20 years, but when I started processing deep space photos I had to take about 80% of that knowledge and dump it out the window. I knew where all the necessary buttons, sliders and functions were and at least that was useful. Other than knowing my way around, I found myself doing things that were completely counter-intuitive to everything I had learned in processing and improving land-based photos. There absolutely is a learning curve with Photoshop and it could be considered steep if you’ve never used it before. Yes, there are long books and thousands of hours of Photoshop tutorials on YouTube, just like PixInsight, but the general intuitiveness of Photoshop is consistent with applications we’ve all been using since the 80’s. On the most basic level there are buttons like OK, and cancel in Photoshop instead of a triangle, a square that is sometimes there, a filled-in circle that is there most of the time and a empty circle that pops up once in a while. All those shapes in PI have a meaning and a specific function, but it’s just another thing to learn.

From a usability perspective, the amount of amateur documentation and tutorials available for both applications is about the only similarity. Let’s take a look at one function – stretching an image. I understand how to perform in both applications, but unnecessary complexity is introduced in one of them.

Stretching an image in Photoshop:

  1. The newbie way – Open the Levels tool. Adjust the mid-point slider gradually to the left and bring the black-point slider to the right, avoiding clipping the left side of the histogram.
  2. The Pro Way – Open the Curves tool and adjust with gradual curves by pulling the line up, then open Levels and bring the black-point slider to the right, avoiding clipping the left side of the histogram. Rinse and repeat. Bonus points for adjusting curves and balancing R, G & B separately.

Stretching an image in PixInsight “automatically”:

  1. Open the Screen Transfer Function tool
  2. Open the Histogram Transformation tool
  3. Hit the “nuclear” button on the Screen Transer Function tool
  4. Check if it looks better with the channels linked or unlinked by clicking unchecking the linked button and clearing and again clicking the nuclear button
  5. Drag the triangle in the lower left of the Screen Transfer Function window to the bottom bar of the Histogram Trasformation window
  6. Clear the stretch on the Screen Transfer Function tool window (If you don’t do this, you’ll double-stretch your image)
  7. Drag the triangle in the lower left of the Histogram Transformation tool onto your image (or select your image and click the square) to actually perform the stretch

There are nuances to the Photoshop stretch, just as there are more advanced ways to adjust the stretch in PixInsight. The steps above represent the simplest form of stretching images in each application. Pixinsight performs great estimated stretches. Nine times out of ten you can just apply that first stretch from the Screen Transfer Function Tool. But don’t forget to open the Histogram Transformation Tool, drag the triangle from STF to the bottom bar of HTT, clear the stretch from STF and drag the triangle from HTT to the image you are stretching. Yes, you can perform a stretch directly from HTT in PixInsight. This example simply lays out the most common method of initially stretching an image for each application, and level of convolution involved.

PixInsight is a pain in the ass to use, but it is the industry standard for a reason. Everything in the application was built for enhancing collected photons that have been traveling for thousands, or even millions of years, before landing on your sensor. Why would you even consider Photoshop when you’re starting out when PixInsight is the bees knees? I found three primary reasons that apply to myself, and may apply to you as well.

  1. Pretty much all aspiring astrophotographers have Photoshop in their toolbox, and at $10/month for the Photographer’s subscription, it’s less panful than a nearly $300(USD) upfront cost of PixInsight.
  2. A whole lot of people staring in astrophotography are coming from some level of terrestrial photography experience. Chances are, they already have Photoshop and are familiar with the functionality.
  3. Photoshop isn’t a click-here, click-there kind of predictive processing application. It definitely has it’s own learning curve, but the names of functions and the steps to perform them are intuitive. Adobe is a huge company that has groups of employees whose entire job is to make complex applications more useable.

I took on the free PixInsight 45 day trial not too long after I started imaging and I realized something about a week into it. It was a piece of crap, or at least I thought it was at the time. I believe there is a level of understanding needed about the data you’ve acquired and what you’re trying to accomplish to be successful in using PixInsight. You need this understanding because your data is your data. Tutorials on YouTube are performed on the video creator’s data and following their tutorial verbatim will not produce the same results. Tutorials will definitely help you learn the tools within, but in general, they won’t help with the nuances, or backtracking over the 100 steps you just performed to figure out where it went wrong. When you start imaging, you won’t have that level of understanding and you’ll find that just acquiring the data is hard enough to learn. That’s OK. This whole hobby is a journey. Astrophotography images are not like any images you’ve taken before. I certainly didn’t understand what my data was and what I was trying to do with it on the level necessary to be successful is such a complex application.

Since PixInsight can do everything from start to finish, I need to take just a quick aside to talk about using PixInsight or Deep Sky Stacker. When you follow the right steps, in many cases PI can get you off the ground in better shape than DSS. Below is just a starting point of three hours on a heavy crop of Sharpless 240, the Spaghetti Nebula. This is a tough, dim and huge target. The data on the left was calibrated and stacked using PixInsight, and the data on the right was thrown into the salad spinner and stacked with Deep Sky Stacker. Both were opened in PI, stretched and ABE applied. The noise reduction in the PI stack is undeniable. Upon closer inspection the stars are also quite a bit tighter with better color in the PI stack.

From other examples I’ve seen, differences between PixInsight and Deep Sky Stacker are almost nullified with an extensive data set (especially mono narrowband), or when imaging under dark skies. The image above shows the best example of the starting point PixInsight can provide you if you only managed to get a few hours on an object, or if you live under light-bathed skies.

Where was I? I’ll show an example of IC1805, The Heart Nebula, processed as a best effort in Photoshop and in PixInsight. The image on the left, processed as an HOO palette from Photoshop, received some recognition and several social media reposts. It’s pretty, but it doesn’t hold a candle to the image on the right, processed in an H/HO/O palette, using both PixInsight and Photoshop. The center of the image (Melotte 15) is a bit blown out to the blue channel, but the definition of the dust creating depth is a world apart. A year ago, regardless of what processing software I was software I was using, I couldn’t have created the image on the right.

It’s really not my place to tell you that you absolutely have to go one way or the other; or use both right from the start. However, I do believe an aspiring astrophotographer just starting to collect those ancient photons should start out with Photoshop to get a grasp of what you’re pulling out of those subframes. A lot of people complete much better work than I do with just Deep Sky Stacker an a base version of Photoshop. If you continue imaging, there is definitely a time in your imaging career that you will need to have both Photoshop and PixInsight. When you reach that point is really none of my business. If money isn’t a problem, you might just want to get them both right away. The PixInsight license is perpetual, so you’ll always have it. However, I can pretty much guarantee that until you have an understanding of how this data works and what you’re trying to accomplish, your Photoshop images will look much better than your PixInsight images. It’s very easy to go too far with an application so powerful.

Keep all of those original subs and calibration frames you spent so many hours acquiring. Once you do switch over, you’ll find yourself wanting to reprocess every bit of data you’ve ever collected!

Clear Skies, Bleary Eyes – KA

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Google photo

You are commenting using your Google account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s

This site uses Akismet to reduce spam. Learn how your comment data is processed.