Because I just arrived back from a mini explo trip I did some testing on my way to a surface port. I'm uploading some pics to show what EDO does well, and what it maybe does not so well.
grabbing a drink first.
grabbing a drink first.No, we are talking about Abel Laboratory, which is somewhere in between.yea, but we're talking about a 150-200m settlement... or a pyramid fwiw
In this screenshot (the last pics in the post), he's at 423 km altitude.If you want to talk about apples, talk about apples, not oranges.
ISS orbits about at height of 400km. If you want to show what is visible at that distance, show pics taken straight down, not ones towards the horizon.
Issue Tracker: Planetary Tiling
Now that is interesting. Could you give more detailed specs of that laptop? GeForceNow is a game streaming service. It's like using remote desktop to play game on the cloud. So, whatever the configuration, the game should work at its finest with noticeable latency on the controls and...
This fits quite well with the ISS viewpoint.
This is why most of the pictures used have a visible part of the station.
The same way you spotted Etna because you saw the smoke. If i'm looking at google map for Vesuve, i can also spot it on the ISS screenshot. But this is not what is meant by "can be seen with naked eyes".
All right, let’s talk about maths so everyone can pull out their favourite calculator apps and check if I am right. It is not complicated, after all.
If you see something a units wide from the distance of d units, the angular size will be 2*arctg(a/2/d). However, in all the discussed cases d is much larger than a, so we can use just a/d as a good approximation.
Now, that gives you the angular size in radians. To convert to degrees, multiply by 180 and divide by π (in case you forgot and your calculator app does not have a π button, π ≈ 3.14).
For reference, both the Sun and the full Moon have angular diameters of about 0.5°.
Human eye’s best resolution is about 1 arcminute = 1/60°.
If you see something a units wide from the distance of d units, the angular size will be 2*arctg(a/2/d). However, in all the discussed cases d is much larger than a, so we can use just a/d as a good approximation.
Now, that gives you the angular size in radians. To convert to degrees, multiply by 180 and divide by π (in case you forgot and your calculator app does not have a π button, π ≈ 3.14).
For reference, both the Sun and the full Moon have angular diameters of about 0.5°.
Human eye’s best resolution is about 1 arcminute = 1/60°.
Some points before the pics: its hard to compare a 'dot' to the human eye with a pixel. The human eye has the approximate resolution of close to 600 megapixels. Us gamers have to deal with maximum ~8 million pixels. To give ED the best shot I set the game to 4k, ultra rendering, 2x supersampling. You lot owe my GPU an apology btw.
To check whether the scale is off the following general observations seem fair:
1) The scaling of terrain itself is pretty much spot-on, at least on Ultra+.
2) The scale of stations is spot-on at close distances.
3) If the scaling is off (i.e. ports from a distance are too big) it means they will be relatively larger compared to planetary features at distance versus up close.
So lets go with some pics. I shot the pics while approaching in SC at the slowest speed. For people who cant contain their excitement (I know, who could?)
472 km - invisible, not even a single dot. Not there is already a tiny crater visible, or at least a crater of a few pixels, slightly to the north-east of the center of the circle.
442 km - clearly visible, though the main sky scrapers kinda blur into one. It is the same size as the crater immediately to the north-east of it.
220 km - very clearly visible. The outer wall and three main sky scrapers are all resolved. The station is still the same size as the crater to the north east.
126 km - most basic large elements are now distinguishable. The station is still the same size as the crater to the north east.
23 km - Most smaller buildings are now clearly visible. The station is still the same size as the crater to the north east.
12.7 km - all is resolved in great detail. The station is still the same size as the crater to the north east.
Conclusions:
1) From the second the station resolves to some extent, the scale relative to crater is correct, and remains correct all the way to close-up.
2) There is a missing 'pixel LOD', where the station should be visible from a bit further out. The sudden appearance from zero to a handful of pixels is a bit jarring, but the scale itself seems pretty much spot-on. This may be an oversight, a performance related decision or something they simply didn't care about.
Food for thought: this is rendered at the equivalent of 8k. It is perfectly possible that the LOD onset/blending is less accurate in 1080p. The choice is at that point to either start slightly oversized on low resolutions, or have higher resolutions resolve earlier. I didnt check on other resolutions, so it could be they went for parity and as such option 1). It seems stations resolve in the same way as craters do on Medium rather than Ultra+.
To check whether the scale is off the following general observations seem fair:
1) The scaling of terrain itself is pretty much spot-on, at least on Ultra+.
2) The scale of stations is spot-on at close distances.
3) If the scaling is off (i.e. ports from a distance are too big) it means they will be relatively larger compared to planetary features at distance versus up close.
So lets go with some pics. I shot the pics while approaching in SC at the slowest speed. For people who cant contain their excitement (I know, who could?)
The scaling is correct as soon as the station resolves.
472 km - invisible, not even a single dot. Not there is already a tiny crater visible, or at least a crater of a few pixels, slightly to the north-east of the center of the circle.
442 km - clearly visible, though the main sky scrapers kinda blur into one. It is the same size as the crater immediately to the north-east of it.
220 km - very clearly visible. The outer wall and three main sky scrapers are all resolved. The station is still the same size as the crater to the north east.
126 km - most basic large elements are now distinguishable. The station is still the same size as the crater to the north east.
23 km - Most smaller buildings are now clearly visible. The station is still the same size as the crater to the north east.
12.7 km - all is resolved in great detail. The station is still the same size as the crater to the north east.
Conclusions:
1) From the second the station resolves to some extent, the scale relative to crater is correct, and remains correct all the way to close-up.
2) There is a missing 'pixel LOD', where the station should be visible from a bit further out. The sudden appearance from zero to a handful of pixels is a bit jarring, but the scale itself seems pretty much spot-on. This may be an oversight, a performance related decision or something they simply didn't care about.
Food for thought: this is rendered at the equivalent of 8k. It is perfectly possible that the LOD onset/blending is less accurate in 1080p. The choice is at that point to either start slightly oversized on low resolutions, or have higher resolutions resolve earlier. I didnt check on other resolutions, so it could be they went for parity and as such option 1). It seems stations resolve in the same way as craters do on Medium rather than Ultra+.
Last edited:
Erei, it's just a game mate, it's not real
Heh, true. Ironically though as per my post above it seems this is one of the few things the game actually does do correctly. Not that it matters in any way, but still.
Thanks buddy. I agree. I didn't expect a small comment about something else to derail in several pages of threads with IRL pictures.
So... you're compairing the game with itself to conclude that the game is right ?Some points before the pics: its hard to compare a 'dot' to the human eye with a pixel. The human eye has the approximate resolution of close to 600 megapixels. Us gamers have to deal with maximum ~8 million pixels. To give ED the best shot I set the game to 4k, ultra rendering, 2x supersampling. You lot owe my GPU an apology btw.
To check whether the scale is off the following general observations seem fair:
1) The scaling of terrain itself is pretty much spot-on, at least on Ultra+.
2) The scale of stations is spot-on at close distances.
3) If the scaling is off (i.e. ports from a distance are too big) it means they will be relatively larger compared to planetary features at distance versus up close.
So lets go with some pics. I shot the pics while approaching in SC at the slowest speed. For people who cant contain their excitement (I know, who could?)
The scaling is correct as soon as the station resolves.
472 km - invisible, not even a single dot. Not there is already a tiny crater visible, or at least a crater of a few pixels, slightly to the north-east of the center of the circle.
442 km - clearly visible, though the main sky scrapers kinda blur into one. It is the same size as the crater immediately to the north-east of it.
220 km - very clearly visible. The outer wall and three main sky scrapers are all resolved. The station is still the same size as the crater to the north east.
126 km - most basic large elements are now distinguishable. The station is still the same size as the crater to the north east.
23 km - Most smaller buildings are now clearly visible. The station is still the same size as the crater to the north east.
12.7 km - all is resolved in great detail. The station is still the same size as the crater to the north east.
Conclusions:
1) From the second the station resolves to some extent, the scale relative to crater is correct, and remains correct all the way to close-up.
2) There is a missing 'pixel LOD', where the station should be visible from a bit further out. The sudden appearance from zero to a handful of pixels is a bit jarring, but the scale itself seems pretty much spot-on. This may be an oversight, a performance related decision or something they simply didn't care about.
Food for thought: this is rendered at the equivalent of 8k. It is perfectly possible that the LOD onset/blending is less accurate in 1080p. The choice is at that point to either start slightly oversized on low resolutions, or have higher resolutions resolve earlier. I didnt check on other resolutions, so it could be they went for parity and as such option 1). It seems stations resolve in the same way as craters do on Medium rather than Ultra+.
Seems about right, however that space port is how wide? 2km? 3km? more? As in 10-15 times bigger than a pyramid (or a settlement)?
Edit actually, it doesnt seem that right anymore - at 3km it should be noticeable from way higher than 450km
Last edited:
As we don't have in game access to anything we have real world photos of what else can you do experimentally?
Yes, I am comparing one thing that so far everyone agrees is correct with something people think is not. So unless you want to argue the entire planet is incorrectly sized, including everything on it, and that everything shrinks down at the same pace then my method is correct.
At this point it feels like two people just can't admit to being wrong no matter how small the issue is.
Yes.
Yes, that is what I mentioned. There is a missing 'tiny LOD' level from further apart. It goes from nothing to a handful of pictures, rather than one pixel to 4 etc. But as soon as it is visible it is correct.
A pyramid-shaped object would be about a pixel around 400km, this should be a fair bit larger. Of course, in actual game terms this distance is covered in SC in less than two seconds normally so it makes sense not to spend too much time on it. Would still be nice if there were a 'station quality setting' to put at Ultra+. But in the actual game experience it seems to be pretty much spot-on after the somewhat jarring appearance if you focus on it.
Again, at 8k.
Its a fundamentally disingenious comment from MacIntoc anyway. Erei suggested the scale of the station is off, not that the scale of the entire planet is off. And unless the latter is actually what happens, my experiment is perfectly valid. Its just grasping for weird stuff so he doesnt have to concede that the scaling is simply correct. And this is about something so minor I doubt anyone cares anyway.
Why does it take a community uproar for you to actually communicate? When are you going to just come out and tell the community that the game isn't a priority, fdev doesn't want to treat the game like a live service game and as a whole the developers not only don't play the game but have no real passion for it anymore.
It's a shame that there isn't any real competition for this type of game... I'd jump to that game in a heartbeat. (SC isn't even close to being a competition sadly)
I do wonder why on earth no major dev studio looks at this forum and says:"Yes, instead of making Call of Duty 27 and earn piles of money from millions of happy gamers in a year we're gonna spend a ' decade making a vastly complex space game so we can have a bunch of permanently angry boomers yell at us day after day about how lazy we are for only providing a few thousand hours of entertainment for a couple of tenners."
Such a mystery.
Latvian. There is a difference.