Layer in PI

[Simon Hicks]

Lots of really good discussion about astrophotography and astroart etc. And lots of valid views from all sides. I'm especially glad that Max played devils advocate because I really think that has opened up the discussion...and I share many of his points and concerns.

I think there is a wider point to be made as well though. There are people out there who want to create pure scientific representations of the physical objects in the sky, some who want to enhance but not change the data, some who are happy to enhance the data a bit more and do some restricted modifications, and some who want to produce out and out astroart.

The point that giving people tools to modify masks easily (paintbrushes etc)....and how they use them is up to them, is a good one. I have on a couple of occassions resorted to modifying a mask in PaintshopPro (shock, horror!!!) and then copying it back into PI. Its a crappy 8bit jpeg mask which is not nice, and just having to use PainshopPro is not very satisfactory.....but sometimes, its just the easiest solution. I would really love to be able to do everything in PI.

The other point is that I want PI to be used by as many users as possible. It is a commercial package, and the future development of it relies on sales revenue...lets not beat around the bush. It is well known that one of the obstacles to people using PI is its different (better) approach. When people find 'astroart' type features missing that they are used to then they get frustrated. The world is dominated by Photoshop users. It would be good for PI to have some of these features to help the Photoshop community delve into PI and rapidly make progress....even if that is more along the astroart direction. If they get hooked on PI, because they are making good progress with it, then they will have a chance to learn about the 'proper' way to process data to maintain as much of its scientific integrity.

That's my two pennies worth. 

[vicent_peris]

Hello again,

I'm pushing hard this discussion. There is, IMO, a misunderstanding of what's "art" and what's "science", and the relationship between them. And this leads, IMO again,  to a wrong concept of what's astrophotography and what's photography in gerenal.

The clone stamp tool is very well classified into the Process Explorer. Lasso, magic wand and so on are not artistic tools, but painting tools. Why you have the prejudice in astrophotography to think that to make an artistic work you need to paint?

My opinion is that we, astrophotographers, are completely biased. We are obsessed in taking such amount of light, such amount of detail, and such amount of deepness, that we forgot the person that's going to look at our works. We use technology to acquire our data, and technology to process it. And finally we think we make science by applying an À Trous Wavelet transform, because we are going to make visible those tiny details.

I made some months ago a non-astro work:

http://www.astrofoto.es/Galeria/NonAstro/Hand1/Hand1_en.html

Why we can think in an artistic result easily? And why we think primarily in a scientific result looking at this photo?:

http://www.astrofoto.es/Galeria/2008/ngc7331caha/ngc7331caha_en.html

Both images have been acquired and processed using exactly the same techniques. Really I don't need the use of the lasso to make an artistic work. The main problem I see in astrophotography and in general photography is that, at the time being, there isn't a real concept of what image processing is. There isn't an idea of image processing as a way itself to communicate with the public. And there's the wrong concept that you are not making an artistic work unless you take the lasso in your hands.

Please, make your own experimentation. Process one image until you know what's it; control all of its aspects: noise, contrast, dynamic range, etc. And last, and not untill you get your image under control, try to do completely different approaches to the image. There's not only one way to process the same image! Try to do 3, 4 or 5 versions of the same data. Think what you want communicate in each version. You will see that each version speaks to the public different documentary and aesthetic aspects.

Regarding this, I have a second version in my webpage of the hand photo:

http://www.astrofoto.es/Galeria/NonAstro/Hand2/Hand2_en.html

I also have three printed versions of this photo, wich were presented in 2006 in an exhibition named "Astrophotography and music, the art of interpretation":

http://www.astrofoto.es/Galeria/2006/LuzCenicienta/LuzCen_en.html

Last Christmas I was in Paris visiting the exhibition "Delacroix and photography". In the beginning of photography, painters were shocked by this new technique, as it was an "attack" to painting. Painters saw those photos as a reflection of reality made by a machine, one copy, and aesthetically false because of its accuracy (opposed to painting). I'm sorry to say that, in many aspects, we are still in the daguerrotype era. This becomes worse if you think that we are in the digital era, where we can have really good data and techiques to achieve our goals. To me, we will not be really in the digital era until we think in image processing as a way of expression itself. We must raise over the concepts of "extracting scientific data" and "artistic representation of nature objects".


Regards,
Vicent.

[Cheyenne]

I think that what this really boils down to is that there are several different goals that people want to achieve.

At the one end, is the pure scientific recording of data.  Just the bits, remove known noise and look at the results.  Maybe play around with false color to help bring out some of the contrasting information.  The goal is to document a physical collection of data.  May make a nice image for some scientific journal, maybe it might have some merit that the general public may go "ooohhh  aaahhhh" over.  But the real goal is to have the data to answer questions.  Where those pixels just as bright last year?  Has there been some small change in that structure over time, etc.

Next is trying to eek out the raw image and enhance it enough so that it's pleasing visible image, the goal is to get an cleaned up image of what the "camera" sees, maybe with some reproducible steps that tries to emphasize certain detectable features in an algorithmic fashion.  In other-words, if given the same set of input data and the transformation information, the end result can be reproduced by anyone.  There are some artistic qualities to this process, but they tend to lean more towards choices of what algorithm is going to be applied.

Finally there is the purely artistic goal, to capture what the mind's eye sees.  It's applying a full artistic hand to the data.  It could be something as simple as adding diffraction spikes to the stars, or maybe it's emphasizing some stars and de-emphasizing others really bring out a particular shape.  But basically it's adding something that wasn't recorded, or selectively removing recording information that would not be removed in an algorithmically way.

Each has it's purpose, none are "wrong" or "right" for that matter depending on what the goal is.

The only real issue comes up when one tries to substitute one type of result for another and claim that it's "true".  (i.e. say adding a few new bright pixels to a galaxy and claiming it's a super nova)

[Simon Hicks]

Vincent,

I think you make some excellent points. We do use the term 'scientific' far too liberally. In reality the only stage that could be used in any scientific way (I think) would be a stacked, calibrated, background subtracted image. Any wavelets or even stretching is an aesthetic action which we use to change the data into something that distorts the data into a form that looks good, i.e. it reveals an image of what we want to see. The unstretched data is very difficult to see! I guess we are all trying to 'see' the wonders in the sky, to appreciate them and enjoy them.....and by adjusting the data we are 'creating' our own distinct distorted version. Essentially a very un-scientific process.

And I certainly don't use any of my data for any scientific purposes. I use it to create nice images.

It's therefore strange how we as a community (me included) have such a shared inate sense of what is right and wrong, permissible or non-permissible in our image processing. We 'know' when an image has been processed 'wrongly'....even though none of us have ever seen the detailed colours and structures of a faint nebula with our naked eye. But at the same time we all accept careful use of Atrouswavelets, deconvolutions, stretching, masking, saturations etc as perfectly valid.

So if we are distorting the data to fit our interpretation of what we want it to show....does that mean we are creating art? And are we artists? Definitely too big a question for me to answer.

By the way, your Hand pictures are excellent...the intense texture and contrast in the skin, the heaviness of the flesh, the hand covering the one breast, and the other exposed really grab your attention....and then the cat's paw (which I didn't notice at first!) is such a humorous and clever addition to the scene.

Cheers
          Simon

[vicent_peris]

Simon,

I agree with you, but would change the word "distorsion" by "interpretation". To me, a linear image is a distorted image visually speaking, as is a fully processed imagen scientifically speaking. A linear representation as the one below is completely distorted visually: bright stars are not white circles, and the galaxy core is not a white plateau.



In other words, we are not distorting data, but changing proportions between image properties to better communicate it visually. The difference between the visual and scientific fields is the search for a different meaning of the original data.

This leads to your question: yes, we're making art. With some constraints, but it's art. This is because our goal is the expression and communication of beauty.

On the other hand, being documentary photography, we must communicate also a nature phenomena, having the responsibility to communicate scientific facts. So here are our constraints... But I think there are the same constraints that, for example, when making an interpretation of a music score.

Thanks for you comments on my Hand pictures... What I made to take the photo is to put the model in the armchair an let the time to pass through... I acquired a serie of photos, while talking with the model, and in some of the photos the cat went on the model legs.

I work by creating a "base scene" and letting things acting alone and naturally. What's really important here is intuition. Think that in a really good work, 1% is genius and 99% hard work. But the intuition included in this 1% is as essential as the hard work. The final photo is a crop of one of the frames. When I saw that small crop inside the photo, I saw at the very first moment the complete work. I think this is really important not only in this class of works, but also in astrophotography. And not only at the time of processing the image, but also when acquiring or even when selecting your targets.

To make the processing talk itself, my idea was to "not illuminate" the scene. The photo was acquired in a really cloudy day, so there was very little contrast in the whole image. In this way, the image becomes more manageable, and the processing can "extract" the light from the objects themselves. The *linear* raw image is below. As you see, it has very low contrast. In fact, my major handicap now is that I need a very good camera... the image was acquired with an Apogee Alta U16 and I get tons of noise during processing due to the low contrast of all the structures.




Best regards,
Vicent.

[Simon Hicks]

Hi Vincent,

Firstly, thanks for sharing the info about the Hand composition, that was facinating.

Re your example of the linear galaxy image being a distorted image visually....I'm not too sure I'm convinced by that. It certainly doesn't look visually like I would see it throught the eyepiece of a telescope. But that's simply because it would be much dimmer....the stars would be nice pinpricks etc. However, the sensitive camera, long exposure, stacking, etc mean that this is more like what I would visually see if everything was brighter, or I had eyes the size of a football pitch to gather more light. When I visually look at the sun on a clear day, I don't see a sharp sun-sized disk, my eyes get saturated....I just see a huge yellow glow. Is that not equivalent visually to a flat topped galaxy?

The difference between the visual and scientific fields is the search for a different meaning of the original data.

I like this definition.

Cheers
         Simon

[vicent_peris]

Hi Simon,

Re your example of the linear galaxy image being a distorted image visually....I'm not too sure I'm convinced by that. It certainly doesn't look visually like I would see it throught the eyepiece of a telescope. [...]

You are the first to say that. :-)

The unstretched data is very difficult to see!

I'm not talking about visual looking of celestial objects. Forgot this; to me, astrophotography has anything to rely on night eye vision. As I said, this is a XIX century heritage.

I talk about image expression and visual content communication. As you said, an unstretched data is impossible to see; but a linear stretched data will show only a tiny part of what you want to communicate visually. Linear stretched data will show you also parts of the object in a way that is not representative of the data the image contains. And worse, linear stretched data will look always the same for one image, because it hasn't any aesthetic basis. Thus, from a pure aesthetic point of view, to me a linear image is a distorted one.


Best regards,
Vicent.