New star flare texture - without the lens flare

[fsgregs]

Hi. I've created a new star flare texture which displays behind all stars in Celestia, without any hint of a lens flare. There is just a glow behind the star. The density and size are similiar to the existing flare.jpg, but without the brighter lines representing lens flare effects. I've also created a smaller flare.jpg texture that does not extend as far out into space. It is about the width of a typical corona. I am guessing, but I suspect our sun and stars would have virtually no "glow" behind them at all except for their coronal atmosphere, if you were to back far away from them to 1 au or more and look at them through a mature solar system. That is because I suspect there would be very low levels of nebula gas near them to light up.

Anyway, you can get both textures at my website (click the WWW link below this post). Rename your existing "flare.jpg" texture, which can be found in your Celestia "textures" folder to some other name such as "flare-old.jpg", and paste my flare texture into the same folder. Then launch Celestia as usual and visit any star.

Enjoy.

Frank

[bh]

Frank...can you post a screenie?

[fsgregs]

OK, Bob. Here is the new flare behind Vega. This is the smaller of the two flare choices I have on my site.




Here is the same view with the default flare.jpg that comes with Celestia.



In the future, I may dim the outer edges of the flare more so that it is less opaque, and even add a bit of random dimming here and there so it is not so evenly round, but you get the idea.

Enjoy.

Frank

[t00fri]

Frank,

I would be interested to learn about the physical effect from which you think your above "flares" will arise in ~ empty space.

Usually the familiar spikes seen around stars are associated with diffraction effects caused by the secondary mirror suspensions in photographs through telescopes. Note that for diffraction effects to be strongly visible the light source must be "small" enough in size at the observer's position. Yours is pretty big, though...

Similarly, the rings seen in telescopes of radius R, say around "point sources" of starlight, are well known diffraction effects associated with the maxima of

Code: Select all


                            |J1(R*theta)/(R*theta)|



in the telescope's focal plane. Here,
theta is the angular distance from the center and
J1 the familiar Bessel function of order 1 that describes the circular diffractive light pattern around the inner bright disc ("Airy disc") of the star.


How is the "dusty halo" in the upper picture supposed to arise in ~empty space?


Bye Fridger

[fsgregs]

Ah, Fridger, I love how you come right at a topic.

If you compare the two images I posted above, you'll notice that the flare I edited "glows" the same and is basically the same texture that you have included in the Celestia distribution, minus the lens flare effects. It is also smaller in radius than the default flare texture. Note that the default texture also has a "glow" surrounding all stars in Celestia when you view them with the default flare. Thus, if you have a problem with star glow around my flare edit, I wonder why you don't have a problem with the star glow in the default flare seen in image 2, which is even more prominent when used in Celestia? (compare both images)

It was my intention to get rid of the lens flare effects and reduce the flare glow image to a size that might resemble the corona around most stars, so that stars viewed in Celestia look as they might actually appear if viewed through a spaceship window (without cameras or telescopes). In photos I've seen of the Sun taken from space, the corona is a true gas haze that has enough mass to be visible, if one blocks out the light from the central star. If that is not true, and no trace of a corona would be detectible in visible light in space, then we should not be using a flare texture in Celestia at all. We should only see the central orb. In my opinion, the lens flare effects really reduce the simulation appeal of the program. For many of us, Celestia brings us into space. The stars should look like they really do as seen with our own eyes, not as seen only through telescopes or cameras.

Conversely, if a coronal glow might be visible but in your opinion, would be much smaller in diameter and dimmer than my edit, I would be happy to reduce the coronal glow to 1/2 of the edited value, or more.

I'd appreciate your advice.

Regards,

Frank

[granthutchison]

In my opinion, the lens flare effects really reduce the simulation appeal of the program. For many of us, Celestia brings us into space. The stars should look like they really do as seen with our own eyes, not as seen only through telescopes or cameras.

I dunno about you Frank, but the lenses of my eyes produce both halo and flare when looking at a bright object against a dark background. (It's largely a product of the semi-crystalline nature of the lens, which scatters a little light internally.)
The corona of the Sun is so much dimmer than the solar disc it seems quite unrealistic to try to make both visible at the same time. (Not, before anyone jumps on me, that I don't find that an interesting concept as an add-on ... just that the argument from realism tends to support a little simple flare, and militate against trying to show the corona.)

Grant

[t00fri]

Ah, Fridger, I love how you come right at a topic.

If you compare the two images I posted above, you'll notice that the flare I edited "glows" the same and is basically the same texture that you have included in the Celestia distribution, minus the lens flare effects. It is also smaller in radius than the default flare texture. Note that the default texture also has a "glow" surrounding all stars in Celestia when you view them with the default flare. Thus, if you have a problem with star glow around my flare edit, I wonder why you don't have a problem with the star glow in the default flare seen in image 2, which is even more prominent when used in Celestia? (compare both images)

Indeed, close to 3 years ago, before I was even a member of the Celestia community, I wrote two Bug reports (unanswered as often) to Chris about those flares in the distribution.

Yet, I was quite sincere above. I sure know about diffraction effects and all that, but I am less educated about the appearance of corona effects as seen from empty space. So my post was a /real/ question, not a fake one...

I agree with most of what you said. My guess would be that the coronal effects are significantly weaker. But I am not sure. Perhaps Grant knows more about this...

Bye Fridger

[granthutchison]

My guess would be that the coronal effects are significantly weaker. But I am not sure. Perhaps Grant knows more about this...

The corona has about a millionth of the surface brightness of the solar disc, so you wouldn't be able to see both at the same time, right next to each other. (Think of a bright white adobe building in desert sunlight, and you're trying to look through the doorway into the artificially-lit interior - the brightness difference there is only about a thousand-fold, and you'll struggle to see anything but a black doorway.)

Grant

[fsgregs]

Grant:

If coronas really are not visible at the same time as stars, then we really should reduce the flare effect around the Sun and all stars in Celestia by a major amount. The glow of the default flare texture is neither realistic nor particularly pleasing with those bold lens flare stripes.

I am a complete novice on the photo-touchup program I am using (Paintshop Pro) so I am not sure I can do a good job on the flare texture, but I would like to drop it to just a hint of a glow, perhaps 60% less in density than my edited version and perhaps another 50% smaller in radius, with no trace of lens flare. There will be some hint of haze glow around stars, but nothing like we have now. I agree with you that our eyes probably do see some "haze" around bright objects as an artifact of our vision, and I agree we need some glow so that you can tell when you are approaching a star (it tends to glow as you close on it). That's a nice effect.

I'll post my new flare texture result in the next week or so after I learn how to do it, unless someone wants to download my edited flare texture as a starting point and drop the radius and density for me???? HINT!

Regards,

Frank

[granthutchison]

The glow of the default flare texture is neither realistic nor particularly pleasing with those bold lens flare stripes.

You don't get flare lines when you look at a bright source in the dark? I certainly do. My only change would be to make them much less symmetrical, but it's an awkward thing to do well.

Grant

[t00fri]

The glow of the default flare texture is neither realistic nor particularly pleasing with those bold lens flare stripes.

You don't get flare lines when you look at a bright source in the dark? I certainly do. My only change would be to make them much less symmetrical, but it's an awkward thing to do well.

Grant

Grant,

yes I was also thinking of physiological effects in the human eye that presumably lead to flare lines. But then the question arises
at which level the simulation should stop its attempt for realism? Should we try to really include generic human observer biases or restrict to realistic modelling of the light sources only.

To carry this eye issue one step further into the extreme, we all know that after looking for a while into bright light sources, humans will see (colored) ghost images and other such effects for a limited time...We would never think of including such effects either, I guess.

Bye Fridger

[Guest]

I made up an alternate flare jpg for Delta Capricorn, which has a small dyson swarm (and lots of artificial planets);





the only thing is, if you set celestia up to display this, it then appears around every star.
an Altflare command might be nice...

[eburacum45]

Thats me again;

[granthutchison]

yes I was also thinking of physiological effects in the human eye that presumably lead to flare lines. But then the question arises at which level the simulation should stop its attempt for realism?

All we're trying to do is to communicate to the user "This is a very bright source", while otherwise disturbing the portrayal as little as possible. (Yes?) For realism we'd need to burn a hole in their retinas, so we've already decided there are aspects of reality we can't (and shouldn't) portray. A simple diffuse patch of light around the star (as Frank proposes) is okay, but can I think easily be confused with a physical property of the star itself, giving the impression that all stars are surrounded by equally proportioned dense halos of luminous material. Flares lines, however, are a) clearly non-physical b) familiar to the user c) physiologically realistic. So in my opinion they're both necessary and sufficient for the job required. (An irregular flare made up almost entirely of flare lines, which shifted shape as our viewpoint moved, would be my personal ideal.)

Grant

[DaveMc]

Usually the familiar spikes seen around stars are associated with diffraction effects caused by the secondary mirror suspensions in photographs through telescopes.

Fridger, just to clarify, are the flare spikes really due to diffraction or a reflection phenomenon within the optical system (eye, telescope, etc.)? I don't know, just asking.

Also, this website, http://www.astrocruise.com/articles/decon.htm has some interesting raw and deconvolved images of a star and M63.

Dave

[t00fri]

Usually the familiar spikes seen around stars are associated with diffraction effects caused by the secondary mirror suspensions in photographs through telescopes.

Fridger, just to clarify, are the flare spikes really due to diffraction or a reflection phenomenon within the optical system (eye, telescope, etc.)? I don't know, just asking.

Also, this website, http://www.astrocruise.com/articles/decon.htm has some interesting raw and deconvolved images of a star and M63.

Dave

Dave,

provided we ignore physiological effects in human eyes, then the only possible origin for such spikes is diffraction in connection with secondary mirror suspensions of telescopes. These kinds of spikes around stars would only appear via telescope vision or photography.

Another prominent source of spikes is the human eye itself as Grant detailed above.

A human observer will experience both, depending
on circumstances. In naked eye vision of bright sources of light, the physiologically arising flare spikes are dominant over diffractive effects, notably for extended sources.


Bye Fridger

[bh]

Frank...I really like the new 'flare' image as opposed to the default spikes...It seems the can of worms has been opened...you guys are great!

I still think the way to go on this is some sort of dynamic flare...I realise this is probably impossible within Celestia given the huge number of light sources available at any given viewpoint...interesting stuff all the same!


Regards...bh.

[fsgregs]

bh, grant, fridger, et al:

I can agree that some flare type of effect might actually simulate what we would really see if we were looking at a star from space with our eyes rather than a telescope (ignoring the blinding radiation flux, of course). What I don't think is realistic is the current default flare, with its even lines. It simply looks fake.

I don't think viewers would always get confused into believing a defuse flare with no lens effects means every star has this glow around it. Even if they did, it might not be wrong. Frankly, all stars probably do. Every star has a dim corona around it of one kind or another. If we ignore it because it is too dim to see in reality (unless we block out the central source), then we should have no flare at all, or an irregular one that has dimmer spikes of uneven length. The problem with such a texture is that Celestia will always draw it in the same and all stars will look identical.

Ideally, as the posts above seem to suggest, we can and should create a dimmer, smaller diameter flare.jpg texture that has some hint of a glow and an irregular dim flare with uneven "lines". It should be a LOT smaller than it is now, in both diameter and brightness. I am just not enough of a graphic artist to do such a teture justice. Any help would be appreciated. Feel free to start with the one I edited, dim it down, shrink its diameter another 50%, add a few dim flare lines of uneven length, and we have a really nice new flare that satisfies us all. Any takers?

Frank

[fsgregs]

OK, I took my previous edited flare texture, reduced its diameter more, dimmed it down, added a bit of unevenness resembling a bit of flare, and here is the result.



... and a closer view ...




Here are the same two views with the default flare:



... and the same closer view ...




I guess it is a matter of preference, but I obviously prefer the more realistic flare that I have created. If a lot of you agree, maybe we should consider it for a wider use in Celestia (perhaps as the default). Would this be a topic for a "Poll"?

Anyway, I have put the newest flare texture onto my webpage at:

http://www.fsgregs.org/celestia

Rename or save the default "flare.jpg" texture in the Celestia main folder, and substitute it with the edited one.

Enjoy.

Frank

[norm.shaw]

Are you sure you placed the correct file there? This is what is there now:


http://www.fsgregs.org/celestia/files/flare.jpg