What is processing, what is art?

[Topographic]

Inteested in some opinions.

If you look at Astrobin there are a wide range of processed images for every target, take the Crescent nebula as an example.

My image https://www.astrobin.com/ofnh00/ Compressed version below

Now this is just StarTools with the H(H+O)O preset. The rest is adjustment of the sliders. Compare the colours with some of the others or my image adjusted with Gimp (below), a simple colour adjustment.

So what is the correct colouring? Is there a correct colouring?

We all see 'photoshopped' images, on the internet, in magazines, some are clearly intended to mislead. This is across all genres of photography. I am not pointing the finger at Photoshop users, the term above is simply used to describe images that have become false through manipulation by any computer programme.

I am interested in when we should reveal how an image has been 'tweaked'. Are some images/imagers presenting a false impression of what another photographer can achieve, particularly beginners? When does the image become art rather than a 'true' representation of a target.

Thanks for reading.

David

[firebrand18]

Great topic of discussion (or 'contention'!) depending on how you look at it. Your post made me smile as I have been thinking the exact same thing this week and mulling asking the team here on ST for their thoughts.

This can certainly open a Pandora's Box of viewpoints and I fully agree with you; there is nothing wrong with creating beautiful art and interpretations of the spectacular objects we all image but somewhere along the line you should be indicating that you have indeed "taken liberties" to achieve the end result, so non-astro aware folk on Astrobin, Instagram or other mediums are not fooled into thinking this is what certain objects really look like.

In the end, 'Color' is the biggest culprit; creating monochrome images (like in the good 'ole days) pretty much keeps the 'embellishment' aspect out of the equation; but as soon as you start playing with color manipulation, all heck breaks loose.

I am exclusively a StarTools user, never having dabbled in Pixinsight, Photoshop or other software to process my images, and I strive, as Ivo likes to say, keep things "Documentary" as much as possible.

This week in particular, I am imaging 10+ hours each on the Heart & Soul nebulae using the great Optolong L-Extreme duo-band filter, and no matter how much I drag RGB and Saturation sliders around in the Color module, play with HDR, Superstructure, etc., I ONLY get the normal Red and Cyan-Teal Bi-Color variation as expected with this filter.

Then I jump onto Astrobin and get floored with seeing the exact same targets taken with the exact same filter with completely different Color balancing (SHO, different color/hue etc); beautiful indeed and makes me wonder why I can't get this in StarTools? I am a relative beginner in AP (2 years) but heck, I can't be that much of a novice; something is going on.

Nothing wrong with Art and interpretation; just explain to educate the viewer what they are looking at.

Maybe I'm off-base here and have lots more to learn (always) but would love to hear others comment on this!

Let the fun begin!

Cheers.

Nick

[Mike in Rancho]

Well, indeed the opinions are bound to be all over the map here, both as to what is art vs documentary, and even as to where the line in the sand is for "manipulation" while still maintaining the bulk of photographic integrity. Are "corrections" allowed if you believe you have introduced an improper artifact, for example?

Even within the bounds of documentary astrophotos, the range of decisions affecting the final outcome is vast, and doesn't necessarily start entering the realm of art. Field of view and rotation; amount of integration; stacking settings; how much to stretch, how much to saturate, and so on. Even the bulk of ST's modules, which help us stay documentary, can produce very different outcomes, like contrast and HDR settings. But those are decisions on how and what to reveal, and even though one may like the aesthetics of it, has it really become art? To me, no, not yet. I think of it as art when stuff is created that really was never there, such as hue manipulation (rather than balancing) or creating a phony SII channel in order to achieve a SHO-appearing image when you really don't have S, H, and O.

I am not biased against colors, even if someone makes a gaudy final image (though I will still likely find it too gaudy for my taste!) I see it merely as a decision on how to display the acquired data. I consider most DSO's to be effectively invisible even in luminance/grayscale. This is why we amass long integration, and then have to crazily stretch the image to reveal the structures - which are real and actually there. Same with color, and saturation is the chrominance analogue to stretching. Those wavelengths are there and were truly captured, even though they would be similarly invisible to the live view human eye. So, how much to we want to "stretch" the colors?

Of course, "accurate" balancing of the color channels is probably another story, and we would need Ivo's expertise on that. (for which he'd probably link us to some university paper )

[firebrand18]

Good comments indeed and I agree with your "hue manipulation (rather than balancing) or creating a phony SII channel in order to achieve a SHO-appearing image when you really don't have S, H, and O."

In the end, how much you decide to stretch your data, how much Contrast, HDR, Sharpen, De-Convolve, Shrink, Noise reduction is of course subjective and to taste and nothing wrong there, especially when shooting through the murk of Bortle 8 skies; you need all the help you can get.

Ultimately, its staying within reason and if you follow StarTools standard methodology and flow, will achieve a pretty decent "documentary" result; start playing with "creative" masks in each module or funky compositing, well, you'll start veering down the yellow brick road! If you do and say so when posting, all good! The viewer is not deceived.

Love to hear Ivo's and others thoughts as I have seen an increasing number of comments recently from the regulars here that have hinted at this.

No university papers please!

Cheers.

[Russ.Carpenter]

I'd like to mention that there is an excellent book on the topic of narrowband colors, written by professional astronomer Travis Rector. The title is Coloring the Universe and the publisher is University of Alaska Press. In the introduction, Rector states: "This often complex process of making images that are both aesthetically pleasing and scientifically useful is at the heart of this book."

Another brief mention: 90 percent of the time, I prefer the processing solutions in StarTools to their analogs in PixInsight. However, there is one strategy in PixInsight that is really useful with narrowband images: tone mapping using Pixelmath. I wish Ivo could incorporate something similar in StarTools.

Russ

[admin]

It's quite heartening people here are across (or at least interested) the difference between documentary photography vs art!

The true art, when it comes to astrophotography, is in the deliberate, targeted framing and interpretation of the data. Mastering this is (should be) the goal of the astrophotographer.

Where astrophotography becomes astroart, is when that interpretation impinges on reality. This is usually due to inexperience or not understanding the tools used, and rarely a true attempt to deceive (though there have been instances).

On detail and documentary photography

There are generally two classes of algorithm operations;
Restorative algorithms/operations; these tranform the data to conform to an established ground truth.
Enhancing algorithms/operations; these transform the data in a way that is discretionary.

These can further be categorised based on where they get their data from (from the dataset itself, or from some external source).

	Restorative		Enhancing
Intrinisc	Deconvolution Color Wipe	HDR Sharp Contrast	Shrink
	DBE (sample setting-based background removal)	Heal (hot pixels, dead pixels)
Extrinsic	Photometric Color Correction		Heal (dust donuts) Neural Hallucination (TopazAI, StarNet) Synth Operations with hand-drawn masks

Some operations exist somewhere in between. Purists will want to stay in the Intrinsic/Restorative category as much as possible. Most AP-ers are happy to venture outside that category (myself included) and I would argue you can still call the resulting image documentary (portraying reality) in most cases. However, when you are enhancing your dataset with data that is extrinsic to it, that's when things are decidedly no longer documentary - it no longer accurately reflects what was recorded (e.g. no longer portrays reality).

All that aside, a good rule of thumb is asking yourself then quesiton; "would I do this to a terrestrial negative (DNG) and still call it a photo with all its detail intact?"
If not, then; "to compensate, can I articulate to my audience what is going, so they can understand why things look the way they look?"
If the answer is "no" on both accounts, then you are likely engaging in art, rather than photography.

On coloring and documentary photography

This is part of a post I wrote a while back on CN;

bobzeq25, on 04 Aug 2021 - 01:38 AM, said:

I will note I find the concept of "documentary images" hard to grasp.

The concept is not too hard to understand. When processing an image, simply ask yourself the question "is this step I am performing helping me to convey reality better?"

It's a rather simple 'yes' or 'no' question, but to be able to answer it, it requires you to;

• understand your tools
• have a goal for your final image (and understand it!)
• communicate that goal to your audience so they can understand what they are looking at

To illustrate, take these 6 versions of an image by Courtney Smith (from Unsplash);

post-194246-0-13487200-1628041203.jpg (463.45 KiB) Viewed 5271 times

So, which ones are documentary? "The answers may surprise you"

Top left: Original image, which is pretty straightforward documentary. It depicts a nature scene in Utah. The photographer could have picked any (useful) exposure, depth of field, or color balance, without having to explain herself.

Top middle: A black and white version. Still pretty documentary. Color information is now absent, so your audience will lose some important cues (for example the great variety in chlorophyll concentrations in the vegetation).

Top right: The red and blue channel swapped. It may (or may not) surprise you to learn that this is still documentary, you just need to tell the audience what you did here. It may well be that you wanted to make it easier to see chlorophyll concentrations in the vegetation (humans and many animals are great at seeing things that stand out from green, such as a predator, and also very good at ignoring shades of green for the purpose of picking up important detail - taking the "shades-of-green" out of the equation makes it much easier to analyse what's what).

Bottom left: The original image with a fractional differentiation filter applied to it. This filter specifically makes it easier to spot faint structures, whose presence are very real. Still documentary, but you will have to tell your audience what they are looking at and how to interpret your image.

Bottom middle: A picture that looks plausible, but does not depict reality. Artificial detail has replaced real detail. This is not documentary. Note that if I had chosen to mask out the bush and left it black (or white) this would have helped the image a lot; it makes it clear that something was there, but that it obviously detracted from the goal of the image (either for the purpose of the presentation, or for the purpose of processing - you will often see apodization masks that mask out bright star cores to process and visualize their faint exoplanets for example).

Bottom right: Manipulation of the individual color channels in a non-linear fashion. This is not documentary. Hues vary all over the place depending on brightness and convey nothing about reality. The color information has been compromised by inappropriate use of a tool, and no amount of explaining will help your viewers adjust to this.

As you can see, there is really nothing prescriptive or dogmatic about the notion of performing documentary photography, and you have loads of leeway to do as you see fit. Just don't "lie" about the reality you recorded to your audience, and - preferably - help your audience interpret your image where needed. Just doing the latter can make an image go from ugly/bizarre at first glance, to actually brilliant.

Hopefully this arms you with some ways to separate the wheat from the chaff when perusing tutorials, videos or evaluating images by other astrophotographers!

[firebrand18]

Thanks Ivo for adding your valued insight into this astro "art form" medium we are all obsessed with. Great topic of conversation which I'm sure can be endlessly debated.

[Topographic]

Thanks for people's contributions.

I directed people to a particular target in Astrobin because one celebrated image was aesthetically very pleasing, an award winner, yet the clouds of nebulosity were almost terrestrial, billowing and fluffy, contrasting with the data in nearly all the other images. Of course there were lots of effusive compliments. My thoughts were, 'wait a minute, there are less images and net imaging time than I collected, so what gives'. I played around but couldn't get anywhere close. Cue feelings of inadequacy followed by...'no, someting isn't right'. I checked the processing programme that had been used.....Photoshop.

Now I admit it is a beautiful picture but feel that people are being deceived. it also makes me wary of other images processed with Photoshop, I have the same misgivings over some of the practises in PI. I don't pretend to be able to stand up and argue the relative merits of the PI approach versus the StarTools Ivo Jager system. However sometimes I go by what feels right, hence I use StarTools, nothing to do with the price of PixInsight, obviously

For my second image (above) I stated on Astrobin that I had used a simple colour ammendment with GIMP (cheapskate ). I don't pretend to be an expert but I want to provide as much information as is reasonable, it may help someone. Or not