Give me a reason to stay....please!!!

[rdryfoos]

To Juan and Vicent,

Over the years both of you have resisted many of the things PS can do--saying at one point "why would you want to do that with the data" It was specifically about isolating certain features of an image for processing.  Well--I have stayed with PI, being very loyal and refraining from using other software--trying to see your vision.  But now I am presented with a big problem.  An accomplished imager just posted an image of Markarians Chain taken with an FSQ 106--a 4" refractor--wide field.  BUT, he inserted long focal length data captured with a bigger scope into the galaxies.  It looks amazing--and at full resolution the galaxies look like they were captured with a big scope.  I have been asking people on this forum for years how to do this in Pixinsight.  No one has answered me.  You ask why I would want layers?  You ask why I would want high pass filtering?  well--I really want to add high resolution data into targets in a widefiled shot.  I know it can be done...I know the guy who does it....in Photoshop. 

So I ask you....Juan and Vicent….how can this be done in Pixinsight?   I know it can be done....I know Vicent, that you know how to do this.  You are too good a processor not to know.  Juan was supposed to get back to me on this.  I understand he is busy, but that just proves an unfortunate truth---there is no easy way in PI, otherwise he would have just told me instead of having to spend a lot of time and effort working it out.   

Very sincerely,
Rodd

[pfile]

this can be done with staralignment and gradientsmergemosaic. have you tried that? im pretty sure i explained how to do this before but i cant remember.

layers arent coming. juan even implemented it and posted screenshots of layers in action several years ago, and then he changed his mind.

rob

[Pierre]

Did a composite of NGC 7331 using an Orion 80ED, SW 120 ED and a C11. Just as Pfile stated. It just works.
Pierre

[rdryfoos]

this can be done with staralignment and gradientsmergemosaic. have you tried that? im pretty sure i explained how to do this before but i cant remember.

layers arent coming. juan even implemented it and posted screenshots of layers in action several years ago, and then he changed his mind.

rob

  I don't care about layers.  Telling me it can be done, through some hyper complex process is sort of the point.  Who the heck can figure it out?  not me.  I have no idea how to do it.  And I have done mosaics in PI (not with 2 sides combined--only one side--linear.  But still.  If you have a work flow that you can list the steps (no explanation needed)  but a list of steps would be great.  I remember you telling me it might be able to be done using that method--I forget your words, but it seemed like a try this sort of thing--not a definite way.  I did not come away from reading it saying--oh yeah, people use PI to do it.  It might not have been you, though.  But someone had advise for me along those lines. 

[rdryfoos]

Did a composite of NGC 7331 using an Orion 80ED, SW 120 ED and a C11. Just as Pfile stated. It just works.
Pierre

I kind of expected that in CAN be done.  Not really surprised.  But where is the tutorial---or at least a hint at how its done (just naming tools in PI is kind of like telling some one to build a house using the tools available at Home Deport--or whatever store you have that sells tools and supplies)  After all--I asked Juan, and he said he would have to work on it and get back to me.  He was busy getting the new version of PI out, so I understand.  But the fact that he had to work on it was a bit unexpected.

[pfile]

i dont know maybe juan wants to build a special tool for what you want to do, not sure.

it's not a hyper-complex process. even if you are using PI, you need to register the high-res images to the low res images somehow. how do we do that in PI? StarAlignment. i know you know that. well it turns out with the right guidance (previews) StarAlignment can definitely register images of disparate image scale.

GradientMergeMosaic is used to merge together images, usually at the same image scale, and only blend the edges. but it stands to reason that it might be able to blend images which are insets of one another.

that was my entire thought process when i decided to try using both tools to merge two images of different scale. like Pierre, i tried it and it worked. it even worked on nonlinear images.

since GMM will produce masks, you might even decide to use one of those masks to "punch out" the wider field image and remove the low-res galaxies completely before running GMM again. but you might need to know some basic PixelMath to get that done.

if you want to post the images one of us can have a try at it.

rob

[Pierre]

All I have done with the set I mentioned was to select the lights and did a StarAlignment.

Do you have a copy of Warren Keller's book "Inside PixInsight"?

[tdgm]

Hi Rodd

Have you looked into the blending script which is supposed to be exactly like using high pass filters etc of Photoshop in PixInsight. Adam Block is a big user of the script maybe try to ask him.


Steve

[Juan Conejero]

Hi Rodd,

As you call me here, let me give you my personal vision on this subject. Please don't take the following as an attack, but only as the truth about my idea of what is astrophotography, which is also the underlying philosophy behind the PixInsight project. I am not trying to convince you or anyone else; this is only my personal opinion.

You don't really need reasons to stay here. What you really need is to ask yourself why you are doing astrophotography, or in other words, what do you think is astrophotography and what can it give to you personally.

PixInsight has not been conceived and designed, and is not being developed, as a tool to paint images. If you want to paint, PixInsight is probably one of the worst tools you may find for that task. Photoshop and similar applications are much better for that. Photoshop will always win to help you build absolutely stunning and wonderful pictures, simply because nothing can compete with arbitrary manipulations to give you exactly what you want to achieve, when and where you want to achieve it.

An accomplished imager just posted an image of Markarians Chain taken with an FSQ 106--a 4" refractor--wide field.  BUT, he inserted long focal length data captured with a bigger scope into the galaxies.

There are different ways to do this, and in some cases it could be more or less justifiable, depending on the context where the image is going to be used and how it will be presented. But this is painting, and the resulting picture as a whole is not a fair representation of the acquired data. It has been arbitrarily manipulated without a global criterion based on physical properties of the represented objects. It has been painted this way just to make it look 'nice' without any documentary criteria.

Yes, this can be done in PixInsight, and some users have described procedures that should work well. However, nothing in PixInsight can compete with Photoshop to achieve this. A couple layers, a few brushes, a put this here and move that there, et voilà, job done.

PixInsight is a tool to help you develop your astrophotography through the knowledge of image processing. Astrophotography, with all the technical and artistic challenges involved, is a path of personal growth. The why and the how are much more important than the final product.

[rdryfoos]

Hi Rodd,

As you call me here, let me give you my personal vision on this subject. Please don't take the following as an attack, but only as the truth about my idea of what is astrophotography, which is also the underlying philosophy behind the PixInsight project. I am not trying to convince you or anyone else; this is only my personal opinion.

You don't really need reasons to stay here. What you really need is to ask yourself why you are doing astrophotography, or in other words, what do you think is astrophotography and what can it give to you personally.

PixInsight has not been conceived and designed, and is not being developed, as a tool to paint images. If you want to paint, PixInsight is probably one of the worst tools you may find for that task. Photoshop and similar applications are much better for that. Photoshop will always win to help you build absolutely stunning and wonderful pictures, simply because nothing can compete with arbitrary manipulations to give you exactly what you want to achieve, when and where you want to achieve it.

An accomplished imager just posted an image of Markarians Chain taken with an FSQ 106--a 4" refractor--wide field.  BUT, he inserted long focal length data captured with a bigger scope into the galaxies.

There are different ways to do this, and in some cases it could be more or less justifiable, depending on the context where the image is going to be used and how it will be presented. But this is painting, and the resulting picture as a whole is not a fair representation of the acquired data. It has been arbitrarily manipulated without a global criterion based on physical properties of the represented objects. It has been painted this way just to make it look 'nice' without any documentary criteria.

Yes, this can be done in PixInsight, and some users have described procedures that should work well. However, nothing in PixInsight can compete with Photoshop to achieve this. A couple layers, a few brushes, a put this here and move that there, et voilà, job done.

PixInsight is a tool to help you develop your astrophotography through the knowledge of image processing. Astrophotography, with all the technical and artistic challenges involved, is a path of personal growth. The why and the how are much more important than the final product.

Please don't take this as an attack...I would not believe it if I didn't read it with my own eyes.  How can using  2 scopes to capture data of an astronomical target be in the remotest sense "painting" ?  That is the most ridiculous thing I have heard out of the PI camp.  really.  Its beyond the pale in truth.  And it is insulting as well.  If using 2 scopes of different focal lengths is painting, why then is not using different filters?  or narrowband filters combined with RGB data?  Combining data from 2 different focal lengths is no different than using noise control to remove noise--or a sharpening tool to sharpen structures.  in fact--it is LESS painting, because you are not manipulating anything.  You are taking data in its raw form and combining it with other data in its raw form--the data represents the same structure--truthfully.  I find it disturbing that when a tool or technique is brought to the attention of the PI team because people that process images think it is good--but PI doesn't have it--then the answer is it is not scientific, or its painting, or somehow it is not being truthful with the data.   It is a sad thing, but I am sorry I committed to PI......I now understand what was told to me in the beginning when I knew nothing about processing.   This is personified in the fact that I argued with Vicent at a seminar regarding using G2 star for color calibration instead of a relative standard, which was the only way to do it in PI at the time.  He thought my opinion was laughable--refused to discuss it further.  Ruined the seminar for me.   Low and behold--a tool for that very approach came out.  Now, I did not have a real problem with the color calibration, but using a relative approach certainly flew in the face of the ...pardon the French..precocious attitude of the PI camp that only scientifically valid techniques are appropriate for them.  Well, I tell you now...adding high-resolution data (not imaginitive painting, or altered data--but truthful data)to a widefield image is not painting, and PI would be MUCH stronger for it. 

So--it can be done in PI....that's good to know.   

BTW...I'll fill you in on a little secret that you probably already suspect....mabe it causes you to grind your teeth at night, but 99% of PI users are just trying to make "pretty pictures".  I will fill you in on an ancillary secret that you should ponder...just because we make pretty pictures does not mean we are not staying true to the data, and portraying what is actually there.  Astrophotography is nothing more or less than landscape photography in principle.   Not many want to paint--they want to represent.   At least I do.  The fact that the creator of PI thinks that adding high-resolution data to a widefield image is "painting" and somehow to be eschewed........quite frankly worries me.  I find it disturbing to the point of being scary.  I will proceed under the assumption that this was all a sleep deprived hallucination.....it has to be...........and all is well again.
Rodd

[rdryfoos]

i dont know maybe juan wants to build a special tool for what you want to do, not sure.

it's not a hyper-complex process. even if you are using PI, you need to register the high-res images to the low res images somehow. how do we do that in PI? StarAlignment. i know you know that. well it turns out with the right guidance (previews) StarAlignment can definitely register images of disparate image scale.

GradientMergeMosaic is used to merge together images, usually at the same image scale, and only blend the edges. but it stands to reason that it might be able to blend images which are insets of one another.

that was my entire thought process when i decided to try using both tools to merge two images of different scale. like Pierre, i tried it and it worked. it even worked on nonlinear images.

since GMM will produce masks, you might even decide to use one of those masks to "punch out" the wider field image and remove the low-res galaxies completely before running GMM again. but you might need to know some basic PixelMath to get that done.

if you want to post the images one of us can have a try at it.

rob

  When I get appropriate data I will consider it.  But, it would be more helpful to have a workflow.  I am familiar with the mosaic tutorial--and that IS very complex.  It takes hours to really get through it.  Now I am supposed to take that and modify it somehow to insert rather than attach.  I am very giving with my knowledge, having listed a work flow to a novice of 20-30 steps just because they inquired.  I post linear, calibrated stacks for folks living in the north during the summer, just so they can have some data to savor.  the fact that adding high res data to a widefield image is proving so difficult to get some advice, and more disturbingly that it is somehow frowned upon by the PI gods, really sucks.
Rodd