Thank You Frontier for Fixing Black Hole Lensing, For Science!

[Boomer Kay]

Whatever it looks like I would like more danger accredited to black holes in game. They truly are without teeth. I want a real event horizon that messes up your FSD and spaghettifies your lovely Asp if you get too near. For me that would add excitement and peril that would be hard to resist. I would be out there looking for black holes to play with if I weren't so disappointed with how they are represented in game.

You can blame the car industry of the 2090ies for that. After too many claims of drivers that programmed their autocruise straight into brick walls (using the 1990ies "burning your crotch with too hot coffee" cases as precedence), the feature to "play around dangerous objects like walls and event horizons" was removed from the navigation software.
It's for your own good. Trust us to handle your thruster.

 

[Father Cool]

I can handle my own thruster thankyou sir and if I want it elongating until it is infinitly long and plunged into a black hole I will!

 

[2plex]

While OPs argument is all nice and sound, I think the main problem here is that the 'gravitational lensing' is partly a post-processing effect which simply cannot be aligned to the proposed 'camera view' position, but will always be centered on the middle of the screen, thus it moves with your headlook or ship rotation.
As long as FD doesnt come up with a much more clever idea how to do it, without the post-processing effect, there will be no point in arguing.

 

[CMDR Cosmic Spacehead]

Such science, much wow.
Confused.

Anyway, I just thought I'd pop in to say that, I just visited The Great Annihilator.

It was not great. It did not annihilate me.
Unimpressed!

Good day!

 

[Gorrister]

Always a pleasure reading your posts,Ziljan.
What also bothers me is the fact that the lensing only reacts on the background, but not on stellar objects within the system. A black hole in front of a nearby sun is invisible. There is no lensing at all, nor a disc.

 

[xondk]

.

I have to ask, if you have some actual visuals preferably with movement that you base this on? because from my perspective Elite is doing it the way you describe as true? when you move away/closer things move, when you fly around it the focus point, 'behind' moves?

 

[Faybs]

There has been a lot of justified complaints about the lack of any dangers near black holes, but this thread is different. Instead I am here to talk about the dangerous lack of science. Yes, the gravitational lensing effect is very nice, and as long as you stay perfectly still and don't turn your head or rotate your ship, then it is roughly stable and accurate. However, if you have TrackIR or a VR headset, you will notice that the background stars will actually MOVE position in the sky when you turn your head. The result is a seasick swimming effect that IS ABSOLUTELY WRONG.

In fact, only way turning your head would move the background stars is if:



Possibility A) (FALSE) The black hole focused light to a point, and that point just magically happened to be where ever you currently are! Black holes are amazing and exotic objects, but they are not magical. Yet in ED they use a kind of optical voodoo that is as nauseating as it is inaccurate. As you can see in the image below, the different paths of light for the same star all end up pointing directly to you, no matter where you look. So if you are close enough, you can turn your head, and star will appear to move in the sky. Black holes are not eye glasses. They cannot focus light from a single source onto the same point. They can only focus it onto a rough line. So the image below (while an accurate interpretation of ED gravity) is completely false.

http://i.imgur.com/KElKva7.png




Possibility B) (FALSE) This leaves us with one option to explain ED's "bad astronomy" lensing effect: your ship must be roughly the same size as a black hole!!! Here we see an accurate lensing effect, with the black hole focusing light onto a line. However, in order to match the moving star effect we see in ED, the ship is forced to be so ENORMOUS (or the black hole is so small) that rotating the ship changes the apparent position of the background star! Below you can see a picture of the way accurate lensing would hit a humongous Asp at different points along the hull, so that turning your ship would move the background stars... again, utterly false.

http://i.imgur.com/rT5UAjX.png






THE CORRECT LENSING EFFECT:
(Frontier Devs please look here for how to fix lensing)

Reality C) (TRUE) As stated above, black holes focus light onto a line, not a point. And we all know intuitively that an Asp Explorer is minuscule even compared to a small stellar mass black hole. So below you can see what should happen to a background star when you turn your head or rotate your ship from a given distance. ABSOLUTELY NOTHING. That's right, the background stars will not move at all. Each little pocket of space surrounding your ship would be like a static bubble with only a single radial path for light to hit your ship from the same distant star. The only thing that can alter the apparent position of the background stars would be if you dramatically change your(x,y,z) position in space so that your ship would be hit by a different yellow light beam path. Eg, change your distance from the black hole by several Light Seconds (Ls).

http://i.imgur.com/rXsgruv.png

That's pretty interesting. Have you posted this in the bug reporting forum? since it sounds like they're giving the rendering engine an overhaul the changes you're suggesting might be something they're willing to fix.

 

[Sapient6]

I have to ask, if you have some actual visuals preferably with movement that you base this on? because from my perspective Elite is doing it the way you describe as true? when you move away/closer things move, when you fly around it the focus point, 'behind' moves?

Zero your throttle, activate headlook, and move your view around. The lensing effect should not change, but it does. That is the problem.

Unfortunately, as someone else has pointed out, it's likely that the lensing is being handled in a way that makes a fix more complex than the OP expects.

 

[Ziljan]

I have to ask, if you have some actual visuals preferably with movement that you base this on? because from my perspective Elite is doing it the way you describe as true? when you move away/closer things move, when you fly around it the focus point, 'behind' moves?

As stated in the OP, the static camera view distortion is OK. If you just take a screen shot or use a static camera angle in debug, the black hole looks great. The problem is the additional distortion effects caused by turning your head or your ship. So I am basically asking Frontier to remove the extra calculations done to align the lens to the POV, and instead treat them as coming from the star. In other words, your helmet is not the black hole. The black hole is the black hole.

ED already does this exact calculation when animating shadows cast by the main system lighting source (eg the Sun), so the code/data for the angle between your POV and the light source already exists. Shadows cast by the sun don't move when your turn your right? They are static. Applying this same information to the lensing effect is much simpler than a lighting source because there are no shadow geometries to calculate. All you need is the already generated angle between your POV and the radial line between your ship and the star. Done. The lens effect would be fixed.

That's pretty interesting. Have you posted this in the bug reporting forum? since it sounds like they're giving the rendering engine an overhaul the changes you're suggesting might be something they're willing to fix.

I posted a bug report last summer. Said they'd look into it. Nothing so far.

Always a pleasure reading your posts,Ziljan.
What also bothers me is the fact that the lensing only reacts on the background, but not on stellar objects within the system. A black hole in front of a nearby sun is invisible. There is no lensing at all, nor a disc.

I have never been in a black hole system with a sufficiently close binary star. If you can direct me to one near the bubble, I'd be glad to take a look and run some numbers.

While OPs argument is all nice and sound, I think the main problem here is that the 'gravitational lensing' is partly a post-processing effect which simply cannot be aligned to the proposed 'camera view' position, but will always be centered on the middle of the screen, thus it moves with your headlook or ship rotation.
As long as FD doesnt come up with a much more clever idea how to do it, without the post-processing effect, there will be no point in arguing.

As stated above, the game already calculates the angle between your POV and the main systems light source in order to calculate shadows. So applying this angle to the lensing calculation vector should not reduce performance in any way.

 

[Varrag]

I have never been in a black hole system with a sufficiently close binary star. If you can direct me to one near the bubble, I'd be glad to take a look and run some numbers.

Unfortunately, this one that I've found is not really near (~15K LYs from the bubble), but you can tell by the screenshots (spoiler below - in the second picture black hole is right between my ship and the star) that there is no light bending. As Gorrister said, lensing effect is applied only onto the pre-rendered skybox, and local light sources are not being taken into consideration. Another engine limitation, I suppose.

Btw system info tells that the orbital period for this nice couple is 0.1 days, semi major axis 0.01AU.

Spoiler

 

[Ziljan]

Unfortunately, this one that I've found is not really near (~15K LYs from the bubble), but you can tell by the screenshots (spoiler below - in the second picture black hole is right between my ship and the star) that there is no light bending. As Gorrister said, lensing effect is applied only onto the pre-rendered skybox, and local light sources are not being taken into consideration. Another engine limitation, I suppose.

Btw system info tells that the orbital period for this nice couple is 0.1 days, semi major axis 0.01AU.

Spoiler

http://i.imgur.com/q9lLsfe.jpg
http://i.imgur.com/mMoW26n.jpg

This video below shows very clearly that the radius for lensing does not impact the companion star at all. So yes, I see exactly what you're saying. The back ground star in this video should have shifted into a partial einstein ring. But it stays perfectly flat and circular.

It's an important issue. Though it impacts a small fraction of black holes with nearby companions, whereas the false lensing impacts every black hole. If they could fix both issues, that would be ideal

[video=youtube;OmQlglgb8wI]https://www.youtube.com/watch?v=OmQlglgb8wI[/video]

 

[Brian of Ardagh]

Whatever it looks like I would like more danger accredited to black holes in game. They truly are without teeth. I want a real event horizon that messes up your FSD and spaghettifies your lovely Asp if you get too near. For me that would add excitement and peril that would be hard to resist. I would be out there looking for black holes to play with if I weren't so disappointed with how they are represented in game.

Once you are spaghettified you are truly dead though, aren't you? That means no escape pod, a complete commander wipe. Probably a bit more peril than most would want but it would be much more realistic universe.

 

[Privata]

My guess is blackholes are in a current state of heavy WIP or even a placeholder

 

[Father Cool]

Once you are spaghettified you are truly dead though, aren't you? That means no escape pod, a complete commander wipe. Probably a bit more peril than most would want but it would be much more realistic universe.

Not necessarily true. The escape pod can be launched when the ship passes the event horizon and be fired back over the threshold into normal space just before zero point. That way we would survive and our magic FSD fitted escape pod can whisk us back to the station as normal. In fact another bug bear of mine is that we don't see the escape pod launch as though we are inside it when we lose a ship. It would make a great immersion moment. Imagine doing that and seeing your ship spaghettify beneath you. Of course in reality you would see your ship freeze in time due to the future light cone effect but we can take a fantastical approach here and neglect that point seeing as we neglect the time dialation effects of black hole proximity in any case.

 

[SushiCW]

It's a dead simple fix. All they would have to do is remove the ship orientation from the lensing effect calculation, and anchor the effect to the "camera view" that faces the Black Hole. If anything it might increase FPS slightly.

Given that the black hole effect is done as a screen-space shader (and the background stars are rendered as a static texture for the system you're in), I doubt it's "dead simple" but there's hopefully something they can do.

 

[Ziljan]

Given that the black hole effect is done as a screen-space shader (and the background stars are rendered as a static texture for the system you're in), I doubt it's "dead simple" but there's hopefully something they can do.

The effect already works as a static off-center lens. This is demonstrated when you hold the camera angle still and do a fly by. All they need to do is keep the current effect but make sure the calculation is done as if the head was always facing the black hole, regardless of where you are actually looking.

This calculation is no different than the calculation that is done with light source angles, like stars casting shadows from moon rocks. But much simpler since there is only a single vector angle to calculate: the line between your ship and the black hole.

 

[micky1984]

Always a pleasure reading your posts,Ziljan.
What also bothers me is the fact that the lensing only reacts on the background, but not on stellar objects within the system. A black hole in front of a nearby sun is invisible. There is no lensing at all, nor a disc.

of course its invisible it doesnt emit light hence u can directly see one only its effects

 

[mossfoot]

of course its invisible it doesnt emit light hence u can directly see one only its effects

He means the black hole creates no distortion on the companion star visually.

 

[Ziljan]

of course its invisible it doesnt emit light hence u can directly see one only its effects

Light cannot pass through a black hole's event horizon, so a black hole isn't transparent. If you were close enough to see the black hole in front if a star and not be completely blinded by the companion star, then the star should look like a ring, not a sphere.

 

[xondk]

Zero your throttle, activate headlook, and move your view around. The lensing effect should not change, but it does. That is the problem.

As stated in the OP, the static camera view distortion is OK..

I see, but moving your head wouldn't that make slightly different light that has been bent around the black hole reach your eyes?, maybe the effect currently just needs to be toned down?