At what point is it cheating?

[JimothyC]

This evening I've been working on some data I took of M31 a couple of days ago. I've been playing around with Pixinsight and trying to get my head around some of the concepts, a lot of which are very different from Photoshop. One of the features I've enjoyed using is the Photometric Colour Calibration, which takes the co-ordinates of the image you've been shooting, plate solves the image and tries to correct the images colour to what they should be. It's a very cool feature and makes up for some of my lack of technical skills when it comes to taking exposures. 

I'm pretty pleased with the result and it's definitely and improvement on my first effort at M31. 

 

However I'm still a bit disappointed in the lack of colour when compared to images such as this one published on Wikipedia. 

https://en.wikipedia.org/wiki/Andromeda_Galaxy#/media/File:Andromeda_Galaxy_(with_h-alpha).jpg

So I thought I would experiment. I downloaded the high res version of the image, added it as a new layer in Photoshop over the top of my original and aligned the stars as best I could. I then changed the blending mode to Colour, so I was only using the colour information from this layer.  I then tweaked the opacity till I liked the result. 

I think it looks much better, but is it still my image? The detail is all mine, but some of the colour data has come from somewhere else. Isn't this much the same as the Photometric colour calibration?

Either way, it still feels like cheating. 

[saac]

I've seen a few images that have been produced collaboratively by sharing data and I think that is fine. For me, the key test is to ask "does the final image use data which was not present in the original image (being it from single source or collaborative)".   In the pursuit of representing what was there, then so long as no information is added, then I personally think that is fine.  If the pursuit is to produce an artistic representation then adding information that was not present in the original image is pushing more towards an artistic outcome so I guess it depends on what your intentions are. 

I guess the only other thing to add is the need to declare the process used to create the image  - simply be honest as you have with the second image. 

 

Jim 

[ecuador]

If you used the image to calibrate your own color information, i.e. find the appropriate curves for each color channel, then I'd say it is fair game. But this image contains data that is not yours. Nobody will mind, but I'd personally say "in this image, Luminosity is my own data".

[teoria_del_big_bang]

This is a question I have asked myself time and time again.

As a newbie I am still finding my way and so far only just dabbled into AP.  There are so many aspects to this I never even thought about when I started out.  I was quite surprised that obtaining the data was not even half of the process (although obtaining data and obtaining good data is quite different) and that the processing can be so involved. There is so much you can do with Pixinsight  and photoshop that I also wonder when it the image a true representation of the images you took and stacked and when does it become more made up by the processing and / or your imagination or wishes ?

I am sure this question will ring a wide range of opinions as it is probably very subjective and often down to the user to whether they are happy with the final result.

Steve

[Daniel-K]

does it matter aslong as your enjoying yourself? 

[SteveNickolls]

You might want to watch this from the Astro Imaging Channel, "The Ethics of Astronomy". Enjoy.

Regards,
Steve

[Stargazer33]

As Jim says, I think as long as you declare what you have done and credit the additional data and don't try to pass the image off solely as your own work, then it's okay.

Personally, I think all the processing that we do is "cheating". After all we never see through the eye piece what we produce in our images. I think that is why so many people new to visual astronomy are so disappointed at what they see at the telescope.

If you then add in Ha, OIII and SII; wavelengths that we cannot see ourselves then you just have to accept what imaging truly is. As I believe Sara has posted about in the past. It's one person's representation of what they imagine the object looks like.

[JimothyC]

22 minutes ago, Daniel-K said:

does it matter aslong as your enjoying yourself? 

You're quite right. I'm really enjoying this new hobby of mine. 

 

14 minutes ago, Stargazer33 said:

As Jim says, I think as long as you declare what you have done and credit the additional data and don't try to pass the image off solely as your own work, then it's okay.

Personally, I think all the processing that we do is "cheating". After all we never see through the eye piece what we produce in our images. I think that is why so many people new to visual astronomy are so disappointed at what they see at the telescope.

If you then add in Ha, OIII and SII; wavelengths that we cannot see ourselves then you just have to accept what imaging truly is. As I believe Sara has posted about in the past. It's one person's representation of what they imagine the object looks like.

It appears that with astrophotography that there is a very fine line between helping the viewer to comprehend the structure of objects and artistic interpretation. I guess the choice of which direction to take is entirely my own, however I'm not suggesting that I'll continue my new technique into the future, it was more of a thought experiment if anything. Why does it feel ok to get colour data from an automatic tool like Photometric Colour Calibration whilst doing it manually from another image feels dishonest? I think I'd rather strive to get the colour data in my original image more accurate through technique at exposure time if possible.  But I expect that's a lot of practice and probably a few more coins to save to get there. 

[MarsG76]

As long the detail and data is all yours, than its not cheating.   

[Stargazer33]

If it helps you learn and encourages you, with images that you are happy with, rather than getting frustrated about not being able to reproduce others efforts, then it can never be bad.

Obviously getting the result out of your own data is always more satisfying, but it can be a long difficult road to learn some of these techniques. Either it is hard to grasp the technology or, as is probably more the case in the UK, we rarely get a long period of decent weather and therefore image data to hone our skills.

[IanL]

Cheating? No, provided you don't claim data that you have sourced from elsewhere is your own, and provided that data is made available for you to use (via a creative commons or other open license for example). If the colour data was lifted from Wikipedia, there will be license details associated with the image telling you what you can and can't do.

Using photometric colour calibration is absolutely not cheating. It is a scientific basis for setting the colour balance of the image and absolutely not subjective (unlike what you did subsequently which was to lift extra data to make the image pretty). OK your eye is not sensitive enough to see Andromeda in colour, but if it was (or Andromeda was much brighter), the PCC process gives you a true representation of what you would see.

But actually, the main problem is that you didn't need to add the extra data in the first place. I took your original image (jpeg so low quality, try it on the original data for better results). Using the PI CurvesTranformation process, I just selected the saturation curve (the "S" at the right hand side of the row of icons near the bottom of the process dialog), and set this curve:

Applying it to the original image with just your data gives me this, which is more like what you were hoping for and a more accurate colour representation that the combined version you created, no extra data needed:

Just keep working on the processing, lots of tutorials out there and every time I work through one I add something new to my workflow.

[cuivenion]

Try this:

http://www.lightvortexastronomy.com/tutorial-example-m31-andromeda-galaxy---dslr.html

The whole site has excellent pixinsight tutorials. This one deals with a m31 in particular.

[JimothyC]

2 minutes ago, cuivenion said:

Try this:

http://www.lightvortexastronomy.com/tutorial-example-m31-andromeda-galaxy---dslr.html

The whole site has excellent pixinsight tutorials. This one deals with a m31 in particular.

Thanks everyone for all your advice. I’ve been working through those tutorials. There’s quite a lot of content there and I hadn’t seen that one. Cheers

[JimothyC]

39 minutes ago, IanL said:

Cheating? No, provided you don't claim data that you have sourced from elsewhere is your own, and provided that data is made available for you to use (via a creative commons or other open license for example). If the colour data was lifted from Wikipedia, there will be license details associated with the image telling you what you can and can't do.

Using photometric colour calibration is absolutely not cheating. It is a scientific basis for setting the colour balance of the image and absolutely not subjective (unlike what you did subsequently which was to lift extra data to make the image pretty). OK your eye is not sensitive enough to see Andromeda in colour, but if it was (or Andromeda was much brighter), the PCC process gives you a true representation of what you would see.

But actually, the main problem is that you didn't need to add the extra data in the first place. I took your original image (jpeg so low quality, try it on the original data for better results). Using the PI CurvesTranformation process, I just selected the saturation curve (the "S" at the right hand side of the row of icons near the bottom of the process dialog), and set this curve:

Applying it to the original image with just your data gives me this, which is more like what you were hoping for and a more accurate colour representation that the combined version you created, no extra data needed:

Just keep working on the processing, lots of tutorials out there and every time I work through one I add something new to my workflow.

Yes that’s pretty much exactly what I was after. I’ll give that a try tomorrow. Thanks

[wimvb]

I too think there's much more in your data just waiting for you to pull out.

Pcc doesn't always give the best "artistc" image with colour contrast. Try traditional colour calibration and Mark Shelley's arcsinh stretch. Before you stretch, extract luminance and process that seperately. Then combine in lrgb combination.

[Joel Shepherd]

Well, who's to say that the other image you found is "correct"? ?  After all, none of us have ever seen Andromeda that close with our own eyes.

That said, play with saturation. For RGB, I saturate after combining the (linear) channels, and before stretching or adding luminance. Stretching reduces saturation. so increasing saturation beforehand results in better color afterwards. Also, after combining and stretching, use an inverted range mask (so you strongly protect the high signal areas) and desaturate the background. Apply a little smoothing and fuzzing to the range mask so that the transition from desaturated to signal areas is natural. That will blacken your background and help the colors pop a little more.

Overall, I'd say for your first pass at PixInsight, that's a good effort. Stick with it. I've been using it for two years now and continue to learn with every new project, and the results have gotten better as well (at least in my opinion!).

[Tim]

It is definitely cheating when people take other's images and try to pass them off as their own. I despise this, and cannot see the point, at all. What do they hope to gain?

Sometimes, I use this technique. Find an image I admire, then try to figure out how the image was captured and processed, and then try to replicate it using my own captured data. Occasionally I use the admired image as a layer over mine to see how it compares, but never, never, never incorporate any of that data i to my own data.

This was the technique I used to try and get close to Hubbles Crab nebula photo.

https://www.astrobin.com/28442/?nc=user

[Scott]

On 06/08/2018 at 21:12, Daniel-K said:

does it matter aslong as your enjoying yourself? 

Not quite Danny. Using someone's data without their permission is definitely wrong. So is not declaring it even when permission is given.

[Mr Spock]

Using an image without permission from the copyright holder is theft.

[Stub Mandrel]

It's not cheating if you are open and clear about what you have done and have permission/licence to use the image.

Strictly the image off Wikipedia is under the CC by A (Creative Commons by Attribution) licence so you ought to link to the licence and give the name of the original photographer as well as your own, although giving a link to the original image will satiasfy most people.

https://creativecommons.org/licenses/by/2.0/

 

There are people on SGL who get great satisfaction out of processing data from large telescopes like the Liverpool scope, and a recent APOD was an amateur re-process of Hubble data. All legit if the original source is clearly attributed.

 

That said, my personal view is that I would be cheating myself if I used someone else's data to improve my own image. I do occasionally test my skills on correcting someone else's image, but I always keep these in a temporary directory of their own I empty regularly so I can't accidentally assume they are my own at some point in the future!

 

[newbie alert]

From a overall visual point of view I still much prefer your original rendition..its only until you zoom in on the second you see abit more structure in the dust lanes..im sure you can pull this out..

 

[JimothyC]

Thank you all for your feedback, both ethical and technical. I'd like to make it clear that I wasn't proposing using other peoples data without attribution or permission when publishing. I'm very sensitive to issues of copyright and take a dim view of people ripping off other people's work. That said, i'm a computer programmer and an open source advocate, so I'm also a huge fan of collaboration and building on top of other people's work for the benefit of all, when licensing allows that is.  I guess it's all a matter of context and perception. 

I've taken the technical feedback on board and had another go at processing my data. I tried the M31 tutorial from LightVortextAstronomy but I wasn't happy with the final result., so instead I used some of the lessons learnt from those and other tutorials and some of the suggestions from above and created this. 

This is much closer to what I was trying to achieve. I'll hopefully add some more data to this project in the future and improve things further.

If anyone else wants to try processing this from the raw data they are more than welcome to. I've uploaded the original post integration image as a zip file, apologies for the massive size (750mb!), I thought about resizing it first, but I thought I'd leave it as is for people to process as they like. The large size is down to 2x drizzling and already large dslr image. I'm not sure I'll worry about drizzling in the future but I was following a tutorial. 

 

drizzle_integration.zip

[newbie alert]

Thanks for the file..i might take you up on the offer if i have enough room on the laptop..I'm feeling  astro starved atm but fear not it won't be at your level!

[ecuador]

In case you want more feedback, you really shouldn't drizzle with a DSLR, except perhaps if you are undersampling and at the same time stacking more than 100 exposures, which is not a common thing.

A quick run through Star Tools (typically takes me 10 minutes, similar to the process I describe here) gave me this:

It actually took me a few extra minutes as the colour module didn't do as well as it does automatically, because the stars seemed to have a lot CA around them (doesn't look like an APO) which threw it a bit. But overall I think it is nice enough for a quick 10-15 mins of processing. 

[JimothyC]

Thanks for the tips. You're quite right about drizzling, I think I'll be out of diskspace in a month if I keep that up. 

I've had a quick scan through your tutorial page and it looks very handy. I'll buy myself a copy of Star Tools and give it a proper go through. Thanks