Please post here all of your reports about bugs or crashes in SpaceEngine. Before you post any bugs, please follow these steps:
First of all, make sure you updated your video card drivers. This may solve 99% of all isues.
Read the Fixing common issues section below. It is possible that your problem can be solved there.
Read the List of known issues section below and make sure that your bugs are not on in it. You can use your browser's word finder to help search the list.
If the bug is not on the list then please post it in this thread. Attach to your message a screenshot (if possible) and a log file (it's called the se.log and is located in the SpaceEngine/system/ directory). Only the log file will help us to understand your problem and find a solution!
Fixing videocard-specific issues
If you do not know which videocard do you have, open the system/se.log file in the text editor and read the Vendor information in the beginning:
INITIALIZING OPENGL [MT] Vendor: ATI Technologies Inc.
Then look for a solution for you:
Nvidia: Major lag spikes
To prevent major lags with Nvidia, go into Nvidia control panel, and in SpaceEngine's profile set Threaded optimization to off.
Nvidia: Water missing on terras, clouds are below the surface
This is becasue of Nvidia driver update. Revert back to 365.19.
ATI/AMD: Crash on approaching to the black hole, neutron star or white dwarf, on using ship's hyperdrive, or when enabling the Oculus Rift mode or the Fish Eye mode
Open config/user.cfg and change this parameter value to false, so that it looks like this:
EnableMipmapsWarp false // use mipmapping for warp effects rendering
If crashes on enabling the Oculus Rift mode or the Fish Eye mode still remains, change these two parameters to false as well:
EnableMipmapsGUI false // use mipmapping for GUI FBO EnableMipmapsFrame false // use mipmapping for scene FBO
ATI/AMD: Crash on Startup on Windows 10
Open the system folder in your SpaceEngine directory and delete the file atioglxx.dll.
Intel HD: Glitchy landscape and textures on planets
Open config/user.cfg and change this parameter to true:
- "Error loading surface generator shaders. Press YES to run program with procedural planets disabled. Press NO to exit"
To fix that, download and install latest drivers for your video card. If you still get this message, it could be that you have an incompatible video card (see next paragraph).
But you still can try and run SpaceEngine by disabling checking of critical errors at startup. Open main.cfg and change value of this parameter to 'true':
IgnoreCriticalCheck true // ignore checking of OpenGL version and critical extensions supporting
NOTE: I can give no guaranties that SpaceEngine will be stable and work correctly if you do make this change.
If you have an "Error loading surface generator shaders" error and press Yes, SpaceEngine can run, but it will not be able to render procedural planetary landscapes.
After updating or rollback of the drivers, you must delete the cache folder in the SpaceEngine directory.
2) Weak system
SpaceEngine needs a modern "gaming" graphics card to run. Weak cards like Intel Graphics (many laptops have it) or 10 year old devices like GeForce 5700 will not run SpaceEngine. There is no way but to update your hardware. There are no plans to make a second version of the engine for weaker systems with simplified graphics, CPU texture generation, etc. It would be like coding another engine from scratch. Once again, I would remind everyone that the minimum system requirements can be found on the Home page, but I will mention it here as well:
CPU 2.0 GHz RAM 2 GB Video GeForce 8600, Radeon HD 2600 VRAM 512 MB
I stress the special role played by videocard onboard memory (VRAM). 512 MB is the absolute minimum. The engine uses a large amount of data that must be in video memory (textures, meshes etc). The goo performance will be only with a dedicated video card with 1024+ MB of onboard memory. Almost all modern $100+ devices can run SE smoothly. CPU power is not so important for SE, but having many cores could improve performance on fast interstellar and intergalactic flight. CPU does not significantly affect the landscape generation speed because it is done completely on the video card.
3) System with hybrid graphics (NVidia/ATI + Intel HD)
If you're running SE on a laptop with dual graphics, make sure SE runs on NVidia or ATI chip instead of Intel HD. You may see it in the beginning of the "SpaceEngine/system/se.log" file. If it says: Vendor: Intel Renderer: Intel® HD Graphics 4000
then you should open Nvidia Control Panel or ATI Catalyst and force SpaceEngine.exe run on NVidia/ATI graphics card. If you do not find a SpaceEngine profile there, create a new one.
4) Artifacts on procedural planets
If you get missing landscape blocks, blurry or blinking textures, random spikes sticking out of surface, check out these recommendations here:
a) If you have a laptop with hybrid graphics, read item #3.
b) Make sure that you make a "clean" installation of SE (if it is a full version, not a patch!). If you install a new version overwriting the old one, you may get many bugs and glitches.
5) Frequent crashing while generating planetary surface
If you have frequent crashes on planets, or get an "OUT OF MEMORY" message on screen, or "Unexpected deleting of child node" in the log file, try taking these steps first:
- Close any other programs that can consume the video memory (games, video players, graphics editors, etc.). The video memory is the most important resource for SpaceEngine. It may help to disable the Windows Aero theme too, as it consumes a lot of video memory. - Disable "3D water" in Graphics settings (Ctrl-F4) - Reduce "Landscape LOD" to -2 in Graphics settings (Ctrl+F4) - Set up the amount of video memory in the main.cfg file:
VideoMemorySize 2048 // video card onboard memory (VRAM) in megabytes
The value must not be less than 1024, but not more than the total amount of physical video memory plus amount of RAM. For example, if you have a graphics card with 1024 MB of video memory, and more than 2 GB of RAM, you can increase VideoMemorySize to 2048.
6) Spaceship disappears when far from a star
This is not a bug. The current version does not yet have a completed spaceship rendering system - they cannot be rendered if there are no stars or planets in the renderlist. Just imagine that the ship is in hyperspace, so cannot be seen :)
7) Blurry textures on Solar system planets
This is not a bug. Our Solar system planets and its satellites have no procedural textures, they use real ones with limited level-of-detail. You can download addons with high resolution textures to increase the level of detail of real planets. But don't expect 1 cm level of detail; these addons also have limited size, and thus limited resolution (up to 150 meters per pixel for some bodies).
8) Problems with very high resolution displays
If you're using a very high display resolution and have problems selecting things with the cursor, or your screenshots are being output at a lower resolution than you're using, try the following:
1) Open Windows Control Panel, navigate to Display, and set the DPI scaling to "Smaller - 100% (default)" - it should look like this.
2) If you want to adjust the DPI scaling, click on "Set custom text size (DPI)" in the sidebar and change its value to suit your preference.
9) Problems in Windows 10 + AMD/ATI
Delete the file system/atioglxx.dll
List of known issues
Green items have been fixed for the next release
Not real bugs, but effects caused by limitations in the engine:
- Star mode "points" and "sprites" look the same. This is right! This switch changes the rendering technology, not the visual appearance. - Ships do not render in interstellar space - Star catalog has many errors in star classification, so there are many "giant dwarfs" and "dwarf giants" - Many real binary stars rendered as single, and many single rendered as binary - SE simply have incomplete catalog of binary stars - Comet nucleus star-sprite will shine through planets[example]. - Galaxy/nebula sprites will rotate when seen from certain angles - Auroras, comet tails and rings can be rendered in front of ships/moons sometimes - Lens flares are distorted by gravitational lensing (black holes and ship warp drive effect) - Text editor for Wiki descriptions is not working - it is not implemented yet - Solar system moons don't have extremely accurate orbits (requires implementation of custom orbit models, e.g. VSOP-87) - Sunspot distribution is not always realistic - Ocean tag not yet implemented in planet editor - Rounding errors (e.g. 30° in catalog rounds to 29°59'59.99" displayed on HUD) - Bloom effects from the bottom of the screen are replicated at the top of the screen in anaglyph mode - Many HIP and HD stars do not appear in the search menu
Real bugs:
- Cloud cyclones have cut edges sometimes - Lines or junctures are visible on planetary surface - Changing "modes" between simulation and orbital when piloting ships can cause glitches - Few issues with asynchronous loading (stars don't always generate properly, especially when flying to new galaxies) - Flashing landscape LOD on a planet when Wiki for it is displayed - GUI response issues with some users (interface items glitchy or not respond easily) - Flickering flare color in close binary systems [example] - Auto-exposure does not react when a sun is blocked while diffraction spikes are set to normal or super - "Eclipse bug" on distant planets (weird dark shimmering on outer planet caused by inner planet crossing the sun or two suns eclipsing) - Level horizon command (End key) doesn't work properly on oblate objects - Weird behavior of movement keys in tracking mode (T) and after releasing it - If a galaxy model is selected, the camera is not following any object, and the program is in Edit Mode, closing and restarting the program will cause the WASDRF movement keys to behave incorrectly until the camera is set to follow an object - Problems when releasing Left + Right click - Music player will unpause when SE window is minimized and restored, or when entering and exiting the main menu - Switching to fullscreen doesn't work sometimes - Automatic reference object setting with spacecraft doesn't always choose the best object (for example, when near dwarf moons or debris asteroids close to a planet you are trying to orbit) - 3D water fog sometimes starts above the water surface - Visual atmosphere heights are inconsistently and unrealistically generated - Physical atmosphere properties are inconsistently and unrealistically generated, and are not tied to the visual models - Cyclones are not generated in the southern hemispheres of gas giants - Shimmering bloom splotch when very close to the surface of a planet [example] - Lighting errors in system browser when viewing from outside the system - "Sky lighting shadow" in the subsolar region of atmospheres - Problems with the skybox export interface, and with the planetarium after using it, for some users - Camera will be set to 90° FOV when returning from fisheye or cubemap view sometimes - If spacecraft .sss Length is very different from its .cfg length, there may be rendering glitches - VSync doesn't work after usage of Oculus Rift in the Direct mode, sometimes it doesn't work at startup when Oculus Rift is connected.
Jadestar, try moving your SpaceEngine folder to somewhere other than Program Files. You can also try running as administrator.
Thanks for the quick reply.
I just tried both but both failed.
I do have a question for you, Space Enginere or anyone else who knows though: Someone here suggested I try the latest AMD beta drivers (15 something) via the command line (i think?).
However, everything I've ever read on this forum cautioned us AMD users from using the beta drivers and last I checked I think 14.12 or was it 14.4 was the preferred AMD driver for Space Engine.
Also a more general question. What is the prefered way of uninstalling old drivers? I just follow the instructions from AMD but I've seen other people say that's not enough?
I don't know a whole lot when it comes to uninstalling drivers, and with Nvidia just installing the new driver has always been good enough. So someone else will have to come along and answer that question.
All forum users, please read this! My SE mods and addons Phenom II X6 1090T 3.2 GHz, 16 GB DDR3 RAM, GTX 970 3584 MB VRAM
I have an issue with comet tails, the example for which is below. I normally just run SE without comet tails on but I wish I could. I run a laptop with Intel HD graphics, which are known to be pretty shocking with SE I believe.
EDIT:
I resolved the issue with a neat little fix. While having a glance through the log, I noticed there are two sections referring to comets, "shaders/comet_tails.glsl" and "planets/comets.sc". It's the second one that was causing the issue.
In the catalogs folder there are 2 *.pak files, "catalogs0973.pak" and "catalogsSE.pak". I opened both of those, and "planets/comets.sc" was only in "catalogsSE.pak", so I simply copied it into "catalogs0973.pak" and now I can finally see comet tails!
I delved deep through every troubleshooting thread in search of this issue that I had and I found nobody else seems to have ever had this issue, which makes me think I may have had a bad download or something like that, I'm not sure. But there's my fix for my own issue in case I'm not the only one!
EDIT: EDIT:
Turns out it's not exactly a permanent fix, and it often reverts back to how the image looks. Also I feel like the editing of the files is the reason my Sol system has disappeared, leaving me to explore the procedural universe instead.
I have just had Windows 7 Professional 64 bit installed on a Dell Precision T3500 with 21GB of RAM, W3530 Xeon 3GHz quad core processor, with an NVIDIA FX 4800 graphics card. I am using a 1028*768 screen mode.
I have installed the latest drivers from the NVIDIA site, well one of the many drivers possible, and I have just installed space engine 0.9.7.2 in "c:\program files" and it didn't even start and Windows reported that it "had stopped working" with the following error info
Problem signature: Problem Event Name: BEX Application Name: SpaceEngine.exe Application Version: 0.9.7.2 Application Timestamp: 549da900 Fault Module Name: SpaceEngine.exe Fault Module Version: 0.9.7.2 Fault Module Timestamp: 549da900 Exception Offset: 000abcc2 Exception Code: c0000417 Exception Data: 00000000 OS Version: 6.1.7601.2.1.0.256.48 Locale ID: 2057 Additional Information 1: 3cc3 Additional Information 2: 3cc3e5e140f18763c6f3bd3e34639414 Additional Information 3: b7d9 Additional Information 4: b7d9da6ecd26bf9014f7664e01e55a0a
There doesn't seem to be an se.log file in my spaceengine/system directory
I am new to Windows 7 so I may not be familiar with all of the security/admin issues.
Help!
---------------------------------------------
Update
I am pleased to report that running Space engine as administrator fixed the problem.
Jesus guys!!! It is not very professional to fail to start with an exception and system error just because the program is not run by the administrator account. I thought I was running as an administrator just to add to my confusion.
When there's a lot of objects in one screen, space engine will ether crash or my entire computer will CRASH AND RESTART. Can you make graphics settings more advanced?
From data of [Shift+F3] browser, I guess that the stars are first generated by luminosity, then other properties.
The first bug is the wrong mass-luminosity relation of main-sequence stars. In reality, L ~ M^2.3 for M below 0.43 M(sun), L ~ M^4 for M between 0.43 M(sun) and 2 M(sun), L ~ M^3.5 for M between 2 M(sun) and 20 M(sun), and L ~ M for M over 20 M(sun). In SE, I found that L/L(sun) = 0.23 * (M/M(sun))^2.3 for M below 0.43 M(sun), L/L(sun) = (M/M(sun))^4 for M between 0.43 M(sun) and 2 M(sun), and L/L(sun) = (M/M(sun))^3.5 for M between 2 M(sun) and 20 M(sun). The gap on 0.43 M(sun) is a small bug, e.g. RS 0-118-8-9449875-533 is an M0V, with mass 0.43006 M(sun) and luminosity 0.034207 L(sun), while RS 0-118-8-9449653-414 is an M0V, with mass 0.43656 M(sun) and luminosity 0.034185 L(sun) And the gap on 2 M(sun) is a large bug, e.g. RS 0-118-5-18458-77 is an F0V, with mass 2.0002 M(sun) and luminosity 11.317 L(sun), while RS 0-118-5-18458-484 is an F0V, with mass 1.8337 M(sun) and luminosity 11.306 L(sun) The latter one is really a problem!
The second bug is that the temperature-radius-luminosity relationship breaks the Stefan-Boltzmann equation: L/L(sun) = (T/T(sun))^4 * (R/R(sun))^2 (where T(sun)=5778 K, not 5860 K in SE!). And the luminosity displayed is always lower than the calculated value by Stefan-Boltzmann equation. As a result, the sun in SE has radius 1.0659 R(sun) , and a generated star with M = M(sun) and L = L(sun) is a G5V and has radius 1.162 R(sun). What's more, this bug is extremely significant in O-type main-sequence stars - the displayed luminosity is only 0.011 times of the calculated value for an O3V star!
The third bug is the confusing of bolometric luminosity and visual luminosity if we don't see bug 2 as a bug. It seems that the "luminosity" for mass-luminosity relation is bolometric luminosity (total energy emission), but the "luminosity" for temperature, radius and stellar class is visual luminosity (luminosity only on visual light).
The fourth bug is that one type of stellar class has only one temperature. As a result, the sun, a G2V star, has temperature 5778 K, but in SE it has temperature 5860 K, because G2V has only one temperature: 5860 K.
The fifth bug is the wrong generating mode of binary stars. I guess that binary stars generate in this way. First a star generates as usual. Then this star has certain probablity to become a binary. If it becomes a binary, then its companion B "steals" A's luminosity, with the total luminosity stay the same. The problem is: A loses some luminosity but still has the same temperature and stellar class. For star system of over 2 stars, e.g. AA, AB, BA and BB, first the star generates, then B generates and A loses some luminosity. Then AB and BB generates, and both AA and BA lose some luminosity.
EDIT: Bug 3 may lead to those results. a. If we see "luminosity" as bolometric luminosity then it doesn't fit Stefan-Boltzmann equation. b. If we see "luminosity" as visual luminosity, then it doesn't fit mass-luminosity relation. c. Combine a and b we get: the radius, temperature and stellar class don't match the mass of a main-sequence star.
The first bug is the wrong mass-luminosity relation of main-sequence stars. In reality, L ~ M^2.3 for M below 0.43 M(sun), L ~ M^4 for M between 0.43 M(sun) and 2 M(sun), L ~ M^3.5 for M between 2 M(sun) and 20 M(sun), and L ~ M for M over 20 M(sun). In SE, I found that L/L(sun) = 0.23 * (M/M(sun))^2.3 for M below 0.43 M(sun), L/L(sun) = (M/M(sun))^4 for M between 0.43 M(sun) and 2 M(sun), and L/L(sun) = (M/M(sun))^3.5 for M between 2 M(sun) and 20 M(sun).
SE uses exactly this function. It is indeed have a gap at M = 2:
I will change it to smoothed version near M = 2.
Quotehyp_cos ()
The second bug is that the temperature-radius-luminosity relationship breaks the Stefan-Boltzmann equation: L/L(sun) = (T/T(sun))^4 * (R/R(sun))^2 (where T(sun)=5778 K, not 5860 K in SE!).
SE uses the same data tables for stars as Celestia. In this tables, G2V star have effective temperature oа 5860 K. This value also used in SE as solar temperature constant for all equations.
Quotehyp_cos ()
The third bug is the confusing of bolometric luminosity and visual luminosity if we don't see bug 2 as a bug. It seems that the "luminosity" for mass-luminosity relation is bolometric luminosity (total energy emission), but the "luminosity" for temperature, radius and stellar class is visual luminosity (luminosity only on visual light).
Vice versa. Yes, SE don't have bolometric and visual luminosity as separate parameters for stars, and compute bolometric correction instead. There is may be other confusions with it.
Quotehyp_cos ()
The fourth bug is that one type of stellar class has only one temperature. As a result, the sun, a G2V star, has temperature 5778 K, but in SE it has temperature 5860 K, because G2V has only one temperature: 5860 K.
Why you think this is bug? SE uses data tables to derive temperature, rotation period and other values for stars. This is why the same class have the same temeprature.
Quotehyp_cos ()
The fifth bug is the wrong generating mode of binary stars.
Why "wrong mode"? This is method I choose to ganarate multiple systems. In 99% cases secondary star have much less luminosity than primary, so change in primary luminosity is negligable. I use this method to save "average" speactral class of the whole system, so procufural bifurcation of catalog stars will not change their spectral class. Also, computing of the new spectral class of the star with lowered luminosty is a bit tricky and may give very incorrect result.
ENG
New site
Russian Federation 
United States 
Pirate 
Italy 
Australia 
China 
, and a generated star with M = M(sun) and L = L(sun) is a G5V and has radius 1.162 R(sun). What's more, this bug is extremely significant in O-type main-sequence stars - the displayed luminosity is only 0.011 times of the calculated value for an O3V star!
