Cubic Star Clusters

[Old Duck]

Up to this point, most of my exploration has been between the Bubble and the Rim. I'm currently on my way to Colonia, my first trip towards the core, and I'm encountering multiple cubic star clusters. As in, the star clusters are obviously geometric cubes / rectangular instead of roughly spherical or planer. This just "ruins my immersion", and I'm wondering if it's a bug or an unavoidable consequence of Frontier dividing the galaxy into a 3D grid / matrix. I'm on a PS4, though I can't imagine this would have any effect on how the Stellar Forge renders star clusters.

I've not noticed this effect when exploring the outer spiral arms, and that makes me think it the problem may be due to the high star density toward the core. I've done graphics programming in the past, and if you fill a container with a few randomly placed dots, it's very difficult to discern the boundary. Increase the density, and the boundaries become visible. This can be worked around by creating an artificial "noisy" boundary to the random dots (a rough circle inside the rectangular graphical surface, for example), and I will be quite disappointed if Frontier failed to do this in the Stellar Forge...

So, is this a fixable bug? Is it platform-specific, or have any of you noticed it on platforms other than the PS4? If the galactic core is indeed filled with cubic star clusters, I will be limiting all future exploration to the spiral arms and rim

 

[LifesAJourney]

There was a discussion on this somewhere, can't find it now. As far as I remember it's a bug but cannot be fixed without rebuilding the whole galaxy which is not possible now. It brakes an immersion badly, I agree. I've done some screenshots with the sky looking more like chessboard than the sky. And there are lots of affected places on the route to Colonia.

Seems we need to live with that.

 

[Old Duck]

There was a discussion on this somewhere, can't find it now. As far as I remember it's a bug but cannot be fixed without rebuilding the whole galaxy which is not possible now. It brakes an immersion badly, I agree. I've done some screenshots with the sky looking more like chessboard than the sky. And there are lots of affected places on the route to Colonia.

Seems we need to live with that.

Well that's disappointing. Am I correct to assume this is proportional to star density, and thus sticking to the less-dense galactic arms should greatly reduce this effect?

 

[Corbin Moran]

Well that's disappointing. Am I correct to assume this is proportional to star density, and thus sticking to the less-dense galactic arms should greatly reduce this effect?

I think so. At least I have found no examples of this along the galactic rim.

 

[marx]

The ED galaxy is made up of cubes, yes. Spheres would sound good at first, but then you realise you can't "pack" the entire space with them. In case you're curious about how exactly sectors are done and systems are named, check out this thread.
The visual inconsistency itself comes from too large differences in stellar (more specifically, class B) densities between neighbouring sectors. Towards the core, there's a sudden jump that's large enough to "draw" the sector boundaries plain enough to see. A fix for this would likely involve either redoing the sectors, which would mean vast areas of the galaxy, or modifying star visibility on the skybox, which would likely be an ugly hack.

Other than these bordering sectors, you can go either away from, or towards the galactic core, and no longer see such things.

 

[Orvidius]

The trouble comes mostly from the bright stars (mainly O and B stars) that can be seen from thousands of lightyears away. If one sector cube generates a lot of them, and the surrounding sectors don't, then you see clearly defined cubes that are visible from quite a distance. I'm hoping that they can do some trickery with how the skybox is rendered, to reduce the effect. But that may require some dynamic analysis of relative density of bright stars in nearby sectors, and I'm not sure how quickly it do that, unless they cheat and build a pre-calculated table of which sectors to treat differently.

It's not an easy problem to solve. Without regenerating the whole galaxy, the alternative is to fudge the visible distance of some sectors relative to others.

Sadly it happens frequently in the core regions:

 

[Shadowdancer]

http://screenshots.necrobones.com/E...n09/2018-01-06 19-40-28 Pyramoe IE-E c3_2.jpg

That's probably a star catalogue import of some highly localised survey.

 

[Orvidius]

That's probably a star catalogue import of some highly localised survey.

Possible, as this screenshot was taken close to the bubble. But it was notably cube shaped as I travelled by (and it was relatively far), and most of the survey blobs are elongated cylinders or stretched out ellipses or spheroids. I can't rule it out, so it's probably a bad example image.

EDIT: This is a better example of a visible sector boundary, but not at an ideal distance. I don't usually take screenshots of these:

 

[Old Duck]

We are the Borg. Resistance is futile.

 

[LifesAJourney]

[noob]

 

[Sapyx]

Yep. We've all seen it. And no, it's not something fixable without either demolishing and rebuilding the galaxy from scratch, or reprogramming the apparent brightness calculations for stars on the skybox (which would make the skybox of Sol look nothing like the current night sky of Earth).

The problem is not sectors, but subsectors. Specifically, the d-boxels, cubes of stars 128 LYs across. In certain parts of the galaxy, the number of hot bright B class stars generated within these d-boxels varies quite dramatically between neighbouring boxels. Because B-class stars are visible from hundreds of LYs away, these differences can be easily seen on the skybox.

Ideally, the stellar forge should have at least "rounded the corners" in such cases, making them look like globular clusters or "natural" star clusters. But I'm afraid it's too late for that now.

My character is religious; he's from Arexack (don't laugh, he's a bit touchy on the subject). He sees these great patterns in the galaxy and is reaffirmed in his belief that the Galaxy was Designed, not made by chance. He's not quite sure what message the Designer is trying to send, other than "I like cubes".

 

[Old Duck]

Yep. We've all seen it. And no, it's not something fixable without either demolishing and rebuilding the galaxy from scratch, or reprogramming the apparent brightness calculations for stars on the skybox (which would make the skybox of Sol look nothing like the current night sky of Earth).

I went to Colonia as part of the Enigma Expedition. Visiting Colonia was on my bucket list and the expedition is for a very good cause. However, I prefer exploring much closer to the Bubble, as I'm a stargazer in real life who's blessed with a wonderful view of the night sky. I prefer the more "accurate" skybox of Sol and friends, and if anything I want it to be more realistic instead of less (Orion's very dim belt, for example).

I am surprised everyone is saying nothing can be done without demolishing the current galaxy. As a programmer, I would write a "hedge clipper" algorithm that would basically thin out the stars alone the cube's edges in a noisy spherical shape. The only "damage" I see is removing systems that have been discovered by players, and even that should be avoidable in most cases. Hey Frontier, give me access to the Steller Forge database and I'll write that algorithm for you

 

[Ian Doncaster]

I am surprised everyone is saying nothing can be done without demolishing the current galaxy. As a programmer, I would write a "hedge clipper" algorithm that would basically thin out the stars alone the cube's edges in a noisy spherical shape.

The problem with retrospective application, as I understand it from what FDev have said, is that because of how the sectors are procedurally generated, removing (or adding) a star system can affect other star systems - if you remove AB-C d1-100, then it can seriously affect AB-C d1-101 as well.

Avoiding directly touching already explored systems isn't too big a deal - the boxes generally show up in high-density areas where virtually nothing is explored (though there are a couple right next to Colonia that might now be impossible to tweak without affecting someone's data) - but the indirect effects could be substantial.

Sure, they could have fixed it by a variety of means when originally building the galaxy, had they realised it was so visible. Doesn't mean that they can do it now.

 

[Old Duck]

The problem with retrospective application, as I understand it from what FDev have said, is that because of how the sectors are procedurally generated, removing (or adding) a star system can affect other star systems - if you remove AB-C d1-100, then it can seriously affect AB-C d1-101 as well.

Oh of course, I totally forget to take that into account. You are correct, that does change everything, because as I understand it, the galaxy is basically generated "on the fly". I'm used to procedurally generating sets of data that then exist in a database and can be further modified. My bad!

 

[Saool]

It's a similar story with Star Sausages. I was lucky enough to go to Dine with the Devs again this year, and spoke to Dav and DB about them. Same answer, They can't start moving stars, as it effects other stars around them, which effects other stars around them etc. An additional issue is it might seriously screw any Commanders in those systems. The bottom line was, that Galaxy is pretty much fixed.

I would not be surprised however if part of the 2020 work is to look at these issues, with perhaps a new galaxy being generated. Of course how you do that without changing first discoveries, Codex entries, Player created factions etc I do not know, which is a pretty solid counter argument.

 

[iain666]

I would not be surprised however if part of the 2020 work is to look at these issues, with perhaps a new galaxy being generated. Of course how you do that without changing first discoveries, Codex entries, Player created factions etc I do not know, which is a pretty solid counter argument.

I would be very surprised if we saw a new galaxy. Tweaks to the sky box generation code to make these forge artefacts less visibly jarring should be doable though, I'd have thought.

 

[Old Duck]

It used to bother me, but these days I get more enjoyment exploring around the Bubble and towards the rim than I do near the core, so I don't go anywhere that has Stellar Forge weirdness, for the most part.

 

[varonica]

I would be very surprised if we saw a new galaxy. Tweaks to the sky box generation code to make these forge artefacts less visibly jarring should be doable though, I'd have thought.

Yes you could do it by fiddling with the way it's displayed to us, but not by moving them around, but it would still be somewhat a kludge and may actually introduce worse artifacts than the ones they are trying to fix.

 

[iain666]

may actually introduce worse artifacts than the ones they are trying to fix.

So long as it's not signed off by the same person who looked at the obvious cubes and thought 'yeah, that looks grand'...

 

[Old Duck]

So long as it's not signed off by the same person who looked at the obvious cubes and thought 'yeah, that looks grand'...

I think that's the problem with this game - nobody ever "looks", they just write code and publish. The blatantly obvious headlook bug that was introduced on PS4 last year is the most literal example of this.