Version 2025.9.9.2364
Jump to navigation
Jump to search
Version 2025.9.9.2364
Development phase
Pre-alpha
Release date
September 12th, 2025
2025.9.9.2364 is a KSA build released on September 12th, 2025 which implemented some FTT and lighting changes and included several changes and fixes.
Additions
- Preparation for soft shadows.
- Added a constructor for spotlights for sanitizing values.
- Added first pass on a spot light part. Still need to handle emissives and conditional rendering during the depth pass.
- Animate ocean waves using FFT, extract normals and sideways displacement from FFT data.
- Perform FFT in compute shader groupshared memory using Stockham's algorithm.
- Implement the ocean spectrum from the paper "A unified directional spectrum for long and short wind-driven waves", will add Jonswap spectrum later.
- Use multiple FFT cascades at different scales to minimize tiling and performance cost.
- Add specialization constants info to ComputePipelineWrapper
- Proof of concept RCS light animation based on VolumetricExhaust Emissive Curve
- Re-added missing mesh file declaration for the apollo burn thruster mesh.
- Added very basic emissive map support for our PBR mesh rendering. Currently just applied to the 'no push constants' shader version as we're switching over to the instanced rendering only anyways.
- Show the actual closest approach with the targeted object. Previously we were only showing the first approach.
- Added a light system gpu settings for controlling cluster counts, bias, and clipping planes from cpu.
- Added code to filter out "closest approaches" with very large range rates. These are not actual closest approaches, just minimum distances within the allowed time bounds.
- Added gemini6a as a nearby target for gemini7.
- Added ability to use Transfer Planner to create transfers to other planets whilst in Solar orbit.
- Added separation of Closest Approaches from actual Encounters.
- Restored closest approach range rate filter.
Changes
- Fix for planet shadowing when the camera is at a negative altitudes.
- Allow large gerstner and FFT waves to be combined. Not polished, just for a showcase/example of how to combined different wave data.
- Fix remaining ocean surface jittering
- Optimize atmosphere reflections to only sample LUTS when intersecting the atomsphere. This was a quick optimization that didn't take much effort.
- Converted light system to dynamic instancing, making it easier to create lights on the fly.
- Optimized the light clusters by reducing the clipping planes. Lights past a certain depth are basically not visible so we might as well avoid calculations in the void.
- Fixed inter-planetary Transfer Planner calculations so they are now closer to where they should be. Need some more work to actually iterate on them to hit the target SOI. But we are now going in the correct direction.
- Fixed wrong squash merge conflict resolution for OrbitalTransfers class.
- Fixed manual Burn created on parent orbit move to vehicle orbit.
- Small fix for when lights instances are cleared. It was causing the debug window to show incorrect data.
- Renamed Cluster to LightCluster to be more descriptive
- Fixed a second case where we were only showing the first encounter rather than the closest.
- Adjusted wording and verbosity on patched conic start/end UI markers to improve clarity.
- Improved range rate encounter filter for small closest approach values.
- Fixed another case of incomplete encounter searching.
- Display physics debug overlay info for all vehicles in the nearby vicinity, not just the current vehicle.
- Fix problem with version detection
- Changed patched conics processing to only capture actual encounter information when calculating new flight plans and burn plans.
- Changed Vehicle Set Target processing to now capture closest approach data only to the set target.
- Changed Encounter and Closest approach processing to ensure we only search and show encounters or approaches between the patch start and end times.
- Changed the number of Polynomials to use in CATCH for closest approach processing in the transfer planner. This saves a lot of time and the closest approach value isn't as critical in the transfer planner, it is not used to determine the actual encounters.
- Fixed orbit line drawing culling on hyperbolic trajectories on the parent or grandparent celestial too early. This was causing potential hyperbolic eject trajectories subsequent to escape from a moon to not being drawn until the camera was quite far away.
- Fixed to ensure that flight plan expiry is also set based on closest approach windows searched for target as well. This has changed due to changes to split encounters from closest approaches earlier today.
- Split out per-vehicle data from the overall update loop in VehicleUpdateTask. This is the first step towards solving multiple vehicles in the same loop.
Removals
- Removed override closest approach checkbox and associated code from Transfer Planner as it is no longer relevant.
- Removed vehicle update flags since they are no longer needed.
Changelog JSON
No JSON available for this build.