Hey, I'll be making an example world to hopefully answer all of these questions and more over the next several days.

In the interim - Yes, your ship will interfere with camera raycasts. If you press "ALT+F11" and check the box "Enable Debug Draw" (only available in SP), you will see all of the raycasts happening (orange lines shooting out from the cameras). They must have an unobstructed view to function properly, as they simply return the first thing the line intersects. So, if you have a 5km scan range, and the camera detects the wing of your ship 20m away, you've just wasted 4980m of scan range. This also applies to a camera attempting to "track" an object.. it can't track what it can't see, so that wing can end up losing you a target lock.

So, you want to place the cameras as far to the "outside" of the grid as feasible. As for why clusters instead of a single camera in each direction... there are some 40000+ scan points that the script will scan in EACH of the 6 directions. The number of cameras in that direction affects the speed with which it can scan all 40000+ points, to a point. The number of scans per tick depend on the number of cameras, as well as the average runtime of the script - it's a balancing act, essentially. But, more cameras is always better, because you must remember there is a 2km per second recharge rate for raycasts. That means, setting your scan range to 2km, you can scan exactly once per second with a single camera, twice per second with two, and so on. Setting your range to 10km, you can scan once per FIVE seconds with a single camera. Increasing the number of cameras increases the number of "Ready-To-Scan" cameras per tick.

For manual targeting, you need to be aware of where the actual "aim point" of the camera is - the crosshair is a projection made directly forward of the ship controller in use. Unless the camera is also placed directly in front of said ship controller, its raycast will not land center between the crosshairs - enable debug draw to see whether it is high / low / left / right of center, and adjust aim accordingly. This is a Keen-ism that I wish they would change, and maybe someday they will!

Example of the two calculations I run:

Calculation 1: Average runtime check - The script will determine how many scans per tick is "safe" based off the current average runtime. With a runtime of 0.1 ms, this number will be around 12-15 scans per tick. Once you enter a gravity well, the average quickly reaches 1.0 ms, which pushes the number way down to 1 scan per 3 ticks, or slower - voxel detection is a sim killer! I highly recommend using a modded sensor or turrets on planets!

Calculation 2: Number of cameras "ready" - The script will determine how many cameras CAN scan per tick based on the total number available (assumes equally distributed number in all directions). With 500 cameras and a 3km scan range, there are 8 cameras assumed to be ready.

To do the math yourself, use the following:

The script will use the lesser value from the above two calculations - so 8 scans would be made per tick in space, whereas 1 per 3 ticks is made in a gravity well. These calculations are updated every 100 ticks.

Hopefully this helped, and hopefully I can get the example world out sooner rather than later.

This is much more comprehensive . Thank you.
Still a couple questions.
Debug draw is useful. I noticed that some clusters don't draw any lines. Does that mean they are obstructed by some blocks to the side of them?
Also about voxels. When building a land based scanner, can the problems be avoided if keeping voxels outside camera's FOV, or land basing is out of option? I already see that i should keep away from asteroid fields.
Is 10k the limit, or it can be overriden, provided that i have a scanner big enough to scan efficiently?

There are 2 possible reasons for cameras not raycasting:

1- I do check that there isn’t an obstruction within 20 meters in the direction of the proposed scan. If there is, then the camera is skipped (it would only return your own ship anyways)

2- A portion of cameras are dedicated to tracking known objects. If there is a known object, tracking cameras will no longer perform regular scans, and will build up charge in case the object moves into their field of view. This prevents cameras from lacking the charge necessary to maintain lock on something due to regular scans happening concurrently.

In space, you are fine moving anywhere with the lidar system. Asteroid fields aren’t an issue for sim speed as there are only a handful of voxels to sift thru. The moment you enter a gravity well, however, the raycast now looks thru a LOT of unnecessary voxel to see if any of it was intersected by the ray. This happens even if you sit at 30k altitude with a 1k scan range.. you can thank Keen. I recommend using a sensor for planet-based detection. Or play in SP.

10km is the limit. Anything more than that isn’t really feasible.

Hi, thank you for your work on this! After watching the videos and reading this I’ve still got a couple of questions.

Someone mentioned putting the camera’s on rotors that are spinning, from what I can tell that doesn’t seem to be supported by the script? Camera’s need to face in a dedicated direction from the control block, right?

Is “boresight mode” possible? In fighter aircraft the radar’s can be slaved to direct all the energy directly front, allowing a pilot to point the nose at a target and the radar will lock it. I’m thinking of building a small ship with say 9 cameras pointed front as a visual range scout. Point the nose at something and the lidar script tell me what it is and how far away it is?

Remote sensing: I noticed in the discription that Lidar will share data via broadcast antennas. If you are a sci-fi fan you might remember a show called Seaquest DSV. Basically about a submarine in the future. It used remote ROV’s called “whiskers” as sensor pods buzzing around the ship. I’m wondering if it would be worth the effort to setup a few of these drones around my base, say 1 km away / would that mean I can get targets at 11km from the base now?


In general I’m looking for a better early warning system for my base on a pvp server. 600-800 meters of vanilla turret range is “danger close” ! Especially with how easy player made missiles are now-days. Any recommendations for the balancing act you speak of? Multiplayer server that allows scripts but you don’t want to kill sim speed / but I’d like to know if someone makes it into 8-10km

Thanks in advance!!

@rottielover

You can place cameras on a spinning rotor, but the script adjusts the raycast angles based on static placement, so just be aware that 1 deg with static placement will be enlarged by whatever amount the camera has spun on the rotor.

You can also place them all facing forward if you only care about scanning forward. You’ll just have to keep facing the object to keep it locked.

As for setting up multiple systems, the max range for the map is 10km so you won’t be able to see passed that on any one screen. If you’re looking to see enemies coming into your territory, I’d recommend setting up a few systems around your base with a scan range of 3-4km. Then you can set your main base up to use no cameras and set to max range (so it only sees relayed grids from your remote systems). Then you can select and launch missiles from your main base.

biggest thing to remember about performance is that the performance hit for raycasts grows exponentially as the length of the raycast grows, so keep the scan range short for best results.

You can just use your turrets as detectors for the main base so you can see anything that slips passed your remote systems.

Don’t forget about the target camera that allows you to manually scan a target and lock onto it :)

Thank you !! Sounds like a plan is forming at this moment ;)

hi, please, which line in your code can i change the max scan range? i am not finding, i want to make tests beyond 10km.

You cannot :)