Discussion of possible rendering of Pulsars

[t00fri]

As to Celestia, 99% of the code would remain unaffected. Except in case of VERY LARGE masses being involved on the screen (lensing, neutron stars, black holes,...), we have to fork the render code appropriately.

Right . . . but how is this alternate rendering path going to work? It's a large and difficult project, and likely completely impractical except for a restricted set of objects (such as distant stars, which can be effectively treated as points.)

--Chris

Yes, "Cosmo-Celestia" is a large and difficult project and that's one reason why it's fascinating and why it would also guarantee Celestia's longtime attractiveness.

As I wrote earlier, open any volume of S&T, for example, and have a look how many articles you find about the subjects I am advocating and how many you'll find about suspending spacecraft in the sky... .

Also this kind of task would be largely pursued by an interested group of devs with respective know-how. It would not affect the usual Celestia code in a significant manner.

I don't think it's impractical. We just have to modify Celestia's operation mode suitably, once we are moving on cosmological scales or towards large mass distributions.

Lets take the concrete example of implementing the pulsar catalogue.

To keep the code fast and simple enough, we will have to switch to "strong gravity mode" rather than slowly approaching it. So close access e.g. to pulsars would only be allowed via (a modified) GOTO, while far away you'd use the usual code. At close distance a Celesltia-typical (much improved) version of the above videos could be experienced by the observer. Also leaving such systems would only work via GOTO rather than by moving contiuously.

Conceptionally, it's not much different to displaying close-ups of far-away multiple star systems in Celestia.
There you also fork in render.cpp...Mostly, one also selects the star and goes there by means of GOTO. Of course the render code is different and implements the distortion code of GR that is not too complex, actually.

Bye Fridger

[t00fri]

I don't agree with Fridger about the texture representation. I think it isn't worst to use any reasonable texture than showing a dull gray ball. A gray ball is as much unrealistic as using a light blue texture with some noise.

Cham,

don't misunderstand me or also Selden who made that proposal in a wider sense. A gray texture or any other appropriate substitute is NOT to be seen as any kind of approximation to the true texture, but rather is to signal that our knowledge ends wherever this particular texture becomes apparent. As I proposed, we might also experiment with sprite balls, for example.

Implementing any degree of detail that is not published in some respectable source, contradicts the basic philosophy of Celestia's design.

Bye Fridger

[ElChristou]

...To keep the code fast and simple enough, we will have to switch to "strong gravity mode" rather than slowly approaching it. So close access e.g. to pulsars would only be allowed via (a modified) GOTO, while far away you'd use the usual code...

Hmm... I don't like much the idea, it breack the fluidity of Celestia...

[ElChristou]

...I advocate a GENERIC REPRESENTATION of the pulsars, with its "typical" magnetic field and jets, EVEN if we don't have any real data about them (jets shape and aspect, rotation axis orientation, etc). I'm emphasize the idea : GENERIC REPRESENTATION, not just a dull, boring one ("gray ball")...

Here, I agree with Fridger; let's take the jets; how we will orient them without datas? all in the same orientation? in random orientation? in all case we will publish wrong info...
The Magnetic fields? without an approprieded filter I cannot imagine generalizing a specific representation...

[Cham]

The Magnetic fields? without an approprieded filter I cannot imagine generalizing a specific representation...

What is this sentence supposed to mean ? ElChristou, do you know something about magnetic fields ?

[Malenfant]

Personally, I really don't think this is worth the effort. From the sound of it the rendering code of Celestia would need a major rewrite, and for what? to render some detail of a relative handful of extreme stellar objects that most people aren't going to be interested in (I can hear Fridger exploding from here at that comment). It's just silly IMO - there's so much other stuff that needs to be done - star orbits, multiple star lighting, photometry, atmospheres, etc etc - that IMO is much more relevant and more important for the core Celestia.

I think Chris is right - let's get the pulsar catalogue in there by all means, but we really, REALLY don't need to also rewrite the whole program just to render gravitational distortion to do that.

Regarding with the textures - we don't know what a neutron star looks like close up, so let's just make up something that we think is close enough. And regarding the jet orientation - yes, we don't know the orientation of a lot of these, so I guess we'll have to fudge them too. The point is to show where the pulsars are - that's good enough. We don't need to throw ourselves into a hole by obsessing over details that we have very little data on (if any at all) - we just fudge or randomise and include a caveat somewhere that states the textures and orientations are conjectural. Because we really can't do any better than that.

[chris]

Personally, I really don't think this is worth the effort. From the sound of it the rendering code of Celestia would need a major rewrite, and for what? to render some detail of a relative handful of extreme stellar objects that most people aren't going to be interested in (I can hear Fridger exploding from here at that comment). It's just silly IMO - there's so much other stuff that needs to be done - star orbits, multiple star lighting, photometry, atmospheres, etc etc - that IMO is much more relevant and more important for the core Celestia.

Well, improved photometry, atmospheres, and lighting (especially high dynamic range lighting) are things that I'm more interested in right now. There may be a way to do some limited visualization of gravitational deflection of light without rewriting the renderer, but I tend to favor features that will integrate into Celestia more generally.

I think Chris is right - let's get the pulsar catalogue in there by all means, but we really, REALLY don't need to also rewrite the whole program just to render gravitational distortion to do that.

Regarding with the textures - we don't know what a neutron star looks like close up, so let's just make up something that we think is close enough. And regarding the jet orientation - yes, we don't know the orientation of a lot of these, so I guess we'll have to fudge them too. The point is to show where the pulsars are - that's good enough. We don't need to throw ourselves into a hole by obsessing over details that we have very little data on (if any at all) - we just fudge or randomise and include a caveat somewhere that states the textures and orientations are conjectural. Because we really can't do any better than that.

I agree. The pulsar locations and rotation rates are well known . . . I think we should use whatever theories are available to develop a model of a pulsar that is reasonable; it seems like we could do better than a dull gray sphere. It seems like a featureless glowing sphere (or oblate ellipsoid if it's possible to estimate rotation induced flattening from the rotation rate.) Neutron stars are extremely hot, so I imagine that they should appear rather blue.

Even if we don't know the orientation of the beams, it still seems like it would be useful to show *something*. I don't favor the 'lighthouse beam' appearance, because unless the pulsar is sitting in a dust cloud, there's no scattering medium. I'd rather just depict variable brightness on the surface of the object (provided that the beams do in fact emit in the visible spectrum.)

Writing this, I'm getting more excited about adding the pulsar catalog . . . It seems like there's a lot interesting work to be done.

--Chris

[Cham]

because unless the pulsar is sitting in a dust cloud, there's no scattering medium. I'd rather just depict variable brightness on the surface of the object (provided that the beams do in fact emit in the visible spectrum.)

Chris,

you're forgetting the synchrotron radiation from spiraling electrons in the magnetic field, on the magnetic axis. So even if there isn't any scattering medium, there is radiation flowing out of the beams. However, it's most probably radio waves, so not in the visible spectrum. I even suspect that there isn't any visible beams at all, but just a rotating magnetic field with radio waves emitted from the "sides" of the fake beams. Of course, this is just an hypothesis.

[selden]

Pulses from the Crab pulsar have been observed at all wavelengths: radio through visible to extremely high energy gamma rays.

[Cham]

There is a scattering medium all around, in this case !

[ElChristou]

The Magnetic fields? without an approprieded filter I cannot imagine generalizing a specific representation...

What is this sentence supposed to mean ? ElChristou, do you know something about magnetic fields ?

Almost nothing, but if I recall schooldays, magnetic fields are not visible at naked eyes , so I mean that whithout a filter saying "let's see how look the Magnetic Field" of a body, I cannot imagine generalizing (in the case of 1500 bodies) a specific representation in visible light...

[t00fri]

@Malenfant:
============

I can hear Fridger exploding from here at that comment

I am more upset by your usual technique of injecting deliberately (?) incorrect arguments ...

Here is just a little analysis of your remarkable number of incorrect statements in barely 2 1/2 lines

From the sound of it the rendering code of Celestia would need a major rewrite, and for what?
To render some detail of a relative handful of extreme stellar objects that most people aren't going to be interested in

1) Chris has done at least as big a rewrite/expansion of the rendering code for implementing the multi-star illumination/orbiting. This was done despite the fact that we know less than ~250 specified binary star orbits (cf. my visualbins.stc, spectbins.stc) and less than a handful of orbit data for systems with >=3 suns (nearstars.stc)!

2) The point is NOT to render any details of the pulsars (as you write incorrectly), but to the contrary to render the deflection of light from sources (stars) behind the pulsar's strong gravitational field. This effect would be identical in nature for ALL 1500 (!) pulsars and of course be generically identical to an implementation of gravitational lensing in general. The new code applies equally to implement light deflection around any strongly gravitating object!

3) As I have previously explained, the new (light deflection) code modification ONLY becomes effective in the environment of the strong gravitational sources. Unlike the few binaries, we have 1500 pulsars! Is this a "handful"?? Yet most of Celestia's rendering code would remain unaffected.

4) The public AND educational interest is way larger for such strongly gravitating objects (neutron stars, black holes...) than for normal binary star orbits . Just look into the respective magazins...

@Chris:
=======

I understand that you feel way more familiar with an OpenGL implementation e.g. of multiple star systems than with light deflection for strongly gravitating objects. . In my case it's just opposite ...

After my repeated campaigns for "Cosmo-Celestia" and given my deep conviction for it's "winning importance", the only consequent approach for me would be to fork off and proceed ...I would have done precisely that a long time ago already, if only I had more time available to spend on such an interesting and challenging venture....

++++++++++++++++++++++++++
A "Cosmo-Celestia" realization would of course visualize way more exciting and topical aspects of cosmological research than just the above gravitational light deflection. It would start off with conformal coordinates involving completely different scales of mass, distance and time, implement "wavelength filters" from the start, crucial effects of Universe expansion like galaxy redshifting, large scale structure and much more ...like lots of relevant brand-new data.
++++++++++++++++++++++++++

I am sure some years ago, when deepsky objects where missing in Celestia apart from some dull generic blobs , few people would have anticipated that we'd manage to implement 10000+ galaxies with remarkable rendering detail and high precision of orientation and location...

Bye Fridger

[Malenfant]

@Malenfant:
============

I can hear Fridger exploding from here at that comment

I am more upset by your usual technique of injecting deliberately (?) incorrect arguments ...

Here is just a little analysis of your remarkable number of incorrect statements in barely 2 1/2 lines

And I could argue that you're doing your usual technique of deliberately misinterpreting what I say, because that's exactly what you've done.

1) Chris has done at least as big a rewrite/expansion of the rendering code for implementing the multi-star illumination/orbiting. This was done despite the fact that we know less than ~250 specified binary star orbits (cf. my visualbins.stc, spectbins.stc) and less than a handful of orbit data for systems with >=3 suns (nearstars.stc)!

So? Chris said right here that rendering gravitational distortion would necessitate a major rewrite of the rendering code. I think this is a remarkably bad idea considering that he'd have to do a lot of work just to render a situation that hardly ever crops up.

2) The point is NOT to render any details of the pulsars (as you write incorrectly), but to the contrary to render the deflection of light from sources (stars) behind the pulsar's strong gravitational field. This effect would be identical in nature for ALL 1500 (!) pulsars and of course be generically identical to an implementation of gravitational lensing in general. The new code applies equally to implement light deflection around any strongly gravitating object!

It should be pretty clear that "gravitational deflection" is a "detail of a pulsar" - not uniquely of a pulsar, but it's still a property that would be associated with one. That's what I'm referring to, and I don't think it's worth the effort.

Again Fridger, you need to realise that the Celestia you want to see isn't necessarily the Celestia that others want to see. If you want to go ahead and write your 'Cosmo-Celestia' then go right ahead, but I don't see why it should be imposed on the core program.

3) As I have previously explained, the new (light deflection) code modification ONLY becomes effective in the environment of the strong gravitational sources. Unlike the few binaries, we have 1500 pulsars! Is this a "handful"?? Yet most of Celestia's rendering code would remain unaffected.

That's the whole problem though. Yes, that is a handful, because in the vast majority of situations, NOBODY IS GOING TO SEE THE DEFLECTION because they aren't anywhere near a strong gravitational source. It's ridiculous to spend a lot of effort adding a feature to the program that isn't going to crop up a lot. Things like barycentres, and atmosphere rendering, and photometry, and multiple star lighting - they're actually useful and important because they are encountered in a lot of everyday situations in the program. So I think those are much more useful things for Chris to be spending his time working on.

4) The public AND educational interest is way larger for such strongly gravitating objects (neutron stars, black holes...) than for normal binary star orbits . Just look into the respective magazins...

I'm sure it is. And we can add all the catalogues of these things that we like to the program - but I still just don't see that gravitational distortion has any place in Celestia at all.

Also, I've not seen many threads here of people complaining that they can't see existing neutron stars and black holes. People seem to be getting by fine without them being included in Celestia, and the few that are there don't get any complaints about missing gravitational distortions.

After my repeated campaigns for "Cosmo-Celestia" and given my deep conviction for it's "winning importance", the only consequent approach for me would be to fork off and proceed ...I would have done precisely that a long time ago already, if only I had more time available to spend on such an interesting and challenging venture....

++++++++++++++++++++++++++
A "Cosmo-Celestia" realization would of course visualize way more exciting and topical aspects of cosmological research than just the above gravitational light deflection. It would start off with conformal coordinates involving completely different scales of mass, distance and time, implement "wavelength filters" from the start, crucial effects of Universe expansion like galaxy redshifting, large scale structure and much more ...like lots of relevant brand-new data.
++++++++++++++++++++++++++

It's really clear to me that this is what you want to see in Celestia, but I think this is FAR too technical for general use. This is something that would only appeal to a very, very small minority of users - those who have the technical knowhow to actually even understand what they're looking at. If you want to see that then go ahead and make it (and if you don't have time, then that's unfortunate for you), but I don't think anyone else has an obligation to make your vision a reality.

[ElChristou]

First of all it would be cool if both of you could stop those insignificant and ininteresting personal attacks and try some more valuable teamwork...

...Yes, that is a handful, because in the vast majority of situations, NOBODY IS GOING TO SEE THE DEFLECTION because they aren't anywhere near a strong gravitational source. It's ridiculous to spend a lot of effort adding a feature to the program that isn't going to crop up a lot. Things like barycentres, and atmosphere rendering, and photometry, and multiple star lighting - they're actually useful and important because they are encountered in a lot of everyday situations in the program. So I think those are much more useful things for Chris to be spending his time working on.

Mal, we are all agree the features you are looking for are also important, it's why I was advocating to spend 70% of dev time to dev new horizons and 30% to polish the existing rendering; now that's up to Chris of course.

Also, I've not seen many threads here of people complaining that they can't see existing neutron stars and black holes. People seem to be getting by fine without them being included in Celestia, and the few that are there don't get any complaints about missing gravitational distortions.

Too much people don't even know what is a neutron star, that's perhaps why you cannot find so much complain about them.
I'll go to the oposite direction of yours: we should dev in priority all things that are unknown to public, this way we will really add an educational value to Celestia.
Just imagine the questions a kid/teen will ask when he will find a gravitation lens after hours/days exploring the "boring" solar system...

[Johaen]

Malenfant,

I think that I disagree with you here. I realize that I'm just a single user that has done absolutely nothing to actually contribute to Celestia, but I do have an opinion on this matter. I for one am fully interested in neutron stars and what they look like, and what they do to the environment around them. Plus, all objects bend light, not just neutron stars and black holes. Even regular stars bend light, to a smaller degree. And large galaxies can bend light quite alot. I've seen a few pictures online that have the images of galaxies bended and distorted and even multiple images of the same galaxy, because a foreground galaxy (or maybe a group of them, I apologize for not remembering 100%) is bending the light around it. I think that's pretty neat, and I would love to see something like that implemented in Celestia. I know that this could be difficult to code for, so if it's too difficult, then it doesn't need to be done. But it would be neat to see.

[t00fri]

...

1) Chris has done at least as big a rewrite/expansion of the rendering code for implementing the multi-star illumination/orbiting. This was done despite the fact that we know less than ~250 specified binary star orbits (cf. my visualbins.stc, spectbins.stc) and less than a handful of orbit data for systems with >=3 suns (nearstars.stc)!

So? Chris said right here that rendering gravitational distortion would necessitate a major rewrite of the rendering code.

Sorry, but you apparently only read his first comment, which I argued was not correct. In his later comment he weakened his respective statements considerably.

I think this is a remarkably bad idea considering that he'd have to do a lot of work just to render a situation that hardly ever crops up.

I think I can judge best around here what kind of code the light deflection needs and I was never implying that it will be Chris who is to code that (all by himself). Could it be that Chris never heard a course on general relativity while being a student? At least in Europe, GR is taught AFTER the bachelor...

It should be pretty clear that "gravitational deflection" is a "detail of a pulsar"

Just be told by a theoretical physics Prof. that this is NONSENSE.

It's neither a "detail" nor is it associated specifically to a pulsar. It is associated universally to strong (= very massive) sources of gravitational fields that are deforming the geometry of space-time! After implementing the 1500 pulsars we got plenty of those sources!

That's the whole problem though. Yes, that is a handful, because in the vast majority of situations, NOBODY IS GOING TO SEE THE DEFLECTION because they aren't anywhere near a strong gravitational source. It's ridiculous to spend a lot of effort adding a feature to the program that isn't going to crop up a lot.

Sorry I cannot follow at all your logics.

Chris is planning already some close-up generic features for the 1500 ~10KM sized pulsars for people to watch while you are stating that nobody will ever get close to them??

I think we should use whatever theories are available to develop a model of a pulsar that is reasonable; it seems like we could do better than a dull gray sphere. It seems like a featureless glowing sphere (or oblate ellipsoid if it's possible to estimate rotation induced flattening from the rotation rate.) Neutron stars are extremely hot, so I imagine that they should appear rather blue.

++++++++++++++++
What do you guess how close one has to be to notice a 10KM sized object??? VERY CLOSE, indeed!! Of course, by the time you can see such a tiny object the space time geometry is aleady strongly deformed and that is what should be rendered!
+++++++++++++++++

The way to get sufficiently close (to see pulsars) is identical to the way to see multiple stars orbiting each other: just type GOTO =g!

...multiple star lighting - they're actually useful and important because they are encountered in a lot of everyday situations in the program.

++++++++++++++++++++
Please give me more than say hundred realistic, "everyday situations" of multiple star lighting!
++++++++++++++++++++

[selden]

Fridger,

Don't forget that objects are illuminated by reflected light as well as direct light. I suspect Chris might be thinking about how to implement Saturn-shine on Mimas (just a random example)

Having the code for multiple light sources in place will make this possible, and, I suspect, might have been easier to implement initially for direct illumination.

And, of course, there are exoplanets being found in multi-star systems.

[Malenfant]

I think I can judge best around here what kind of code the light deflection needs and I was never implying that it will be Chris who is to code that (all by himself). Could it be that Chris never heard a course on general relativity while being a student? At least in Europe, GR is taught AFTER the bachelor...

I think Chris is the best judge of this, not you. He's the one that'll have to integrate it into the core program after all.

Just be told by a theoretical physics Prof. that this is NONSENSE.

It isn't. You just don't understand what I'm saying. I'm thinking you don't really understand general english conversation very well.

It's neither a "detail" nor is it associated specifically to a pulsar. It is associated universally to strong (= very massive) sources of gravitational fields that are deforming the geometry of space-time! After implementing the 1500 pulsars we got plenty of those sources!

I've never claimed it was uniquely associated with pulsars, I'm saying that it is nevertheless a property that is associated with pulsars. But even with 1500 pulsars, how many times are people going to be close enough to be able to see any distortions anyway? How many people do you think go visiting every single one of your galaxies, or binary stars? How many do you think will go close enough to a neutron star too? Just to see a slight deflection in light which requires a rewrite of the rendering code to make?

Sorry I cannot follow at all your logics.

Chris is planning already some close-up generic features for the 1500 ~10KM sized pulsars for people to watch while you are stating that nobody will ever get close to them??

He's planning some details that won't require screwing around with the rendering code. The logic is really simple to follow, Fridger - if you can't follow it then it's because you don't want to hear it, but I'll summarise it succinctly enough for you: I think this is too much effort for too little benefit, and there are other extra features that are a higher priority to implement. Clear enough for you?

What do you guess how close one has to be to notice a 10KM sized object??? VERY CLOSE, indeed!! Of course, by the time you can see such a tiny object the space time geometry is aleady strongly deformed and that is what should be rendered!

And I wonder how much the frame rate would drop in the process...

++++++++++++++++++++
Please give me more than say hundred realistic, "everyday situations" of multiple star lighting!
++++++++++++++++++++

Go to any multiple system with a planet.

Clearly you don't think that rendering the non-relativistic universe realistically is important, but I do. I'd rather render the 99% of situations that a viewer is likely to encounter realistically than waste time rendering the 1% of relativisitic situations realistically that they're not likely to encounter. Hence why I think that we should not be wasting time considering this and why we should be focussing on realistic photometry (which is long overdue) and lighting.

Right now (in the official release anyway) Celestia can't even render the Earth's Moon realistically because the photometric function is wrong. And it can't do atmospheric scattering remotely accurately either. And multiple lighting is still flawed. And we don't have reflected lighting either. These are VERY FUNDAMENTAL ASPECTS of visual simulation we're talking about here, that people see all the damn time in Celestia. And correcting these should be a much higher priority than anything else, IMO. As far as I'm concerned, if it can't do those correctly at the moment then arguing about the realism of much more esoteric aspects of the universe is extremely premature and a completely pointless waste of time.

Think about it for a second, and look at the big picture for a change instead of obsessing over details. What's the use of making all these claims about how realistic Celestia would be because it renders gravitational distortions and redshifts when someone just says "yah, but it can't even show the Moon as it really looks". Let's get the essentials in the rest of the program sorted out and then we can worry about luxuries like gravitational distortions, if we are so inclined.

[t00fri]

That was the last time I have started any discussion about dev matters in public!

[Johaen]

How many people do you think go visiting every single one of your galaxies, or binary stars? How many do you think will go close enough to a neutron star too?

I do. Maybe not all of them, but I have visited many of them. I've also turned the brightness of galaxies way up, and the visible magnitude way up, and then zoomed way out to get an idea of our local group of galaxies (which I have been slightly disappointed by the fact that they don't seem to all be there). And then I did the same thing with the Virgo Cluster. It's so amazing to see so many galaxies, and to realize that each on of those has millions of stars, and each one of those stars could have planets around them. It's neat to think about.

How many do you think will go close enough to a neutron star too?

I know I would. I've seen videos of gravitational lensing before, and I'd love to see it in an actual 3-dimensional environment.